Posts

Posted on

Interesting articles, August 2026

Unsurprisingly, the “deal” to end the U.S.-Iran War officially failed as the 60-day deadline to negotiate a full peace treaty elapsed without an agreement.
https://www.nytimes.com/2026/08/17/world/middleeast/iran-mou-deadline.html

A semi-autonomous Russian suicide drone killed three Ukrainians.
‘The drone was programmed by human operators to travel toward a particular gas station, but once it got close, it chose its exact target — most likely propane tanks — on its own, based on its training to recognize such tanks and strike them, the experts and military officials said.’
https://www.nytimes.com/2026/08/24/world/europe/russia-drones-autonomous-ai-kill-ukraine-war.html

‘Ukraine accuses Russia of war crime after video shows drone attacking vegetable seller in Kherson’
https://www.reuters.com/world/europe/ukraine-accuses-russia-war-crime-after-video-shows-drone-hunting-market-vendor-2026-08-04/

Only thanks to dumb luck did the (probably) Russian-piloted suicide drone fail to detonate on contact with the Ukrainian plane. This happened at an airport in Germany.
https://www.thesun.co.uk/news/39980027/putin-jet-europe-russian-drone-airport/

The Director of the CIA flew to Moscow on a secret mission. There is only speculation about what was discussed, but it related to the Ukraine War and possible concerns about Russian plans to attack NATO countries along its border. The Director had hoped to speak with Putin, but instead could only meet with his counterpart in Russia’s FSB intelligence agency.
https://www.nytimes.com/2026/08/27/us/politics/cia-director-russia-ukraine.html

The diplomatic language around this change sugarcoats the fact that Russia has basically lost its bases in Syria.
https://www.nbcnews.com/world/middle-east/syria-will-take-back-control-russia-2-key-mediterranean-bases-new-deal-rcna591663

Turkey, Saudi Arabia and Pakistan have signed a mutual defense treaty.
https://apnews.com/article/mideast-news-roundup-iran-saudi-pakistan-turkey-aug-7-2026-d2ddfe7ce02c7814420b9029158ed57c

Six months into the Iran War, the U.S. has not achieved its strategic objectives and though badly battered, Iran and its regime are still standing.
https://www.nytimes.com/2026/08/28/world/europe/iran-regime-war-us.html

President Trump threatened to bomb Oman–a U.S. ally–if it interfered with his Iran War diplomacy.
https://www.independent.co.uk/news/world/americas/us-politics/oman-trump-strait-of-hormuz-b3034257.html

Italy is transferring an old aircraft carrier to Indonesia.
https://www.zona-militar.com/en/2026/08/26/after-more-than-40-years-of-service-the-aircraft-carrier-giuseppe-garibaldi-said-goodbye-to-italy-as-it-began-its-journey-to-indonesia/

The USS Abraham Lincoln aircraft carrier has been at sea for a record-breaking 250 days, and the performance and morale of its exhausted sailors have declined. This won’t be a factor on fully-automated warships.
https://www.barrons.com/news/senators-demand-inquiry-into-conditions-aboard-uss-abraham-lincoln-bfaa1eca

China’s next aircraft carrier will be another step forward for them, and will be nuclear powered and as large as America’s biggest carriers.
https://www.twz.com/sea/chinas-next-aircraft-carrier-is-taking-shape

The world owes thanks to China.
‘But disaster would still have struck without decisions quietly taken in another capital: Beijing. Between February and June, China slashed its crude imports by half, or 5.5m b/d—enough, experts reckon, to have shaved $30 or more off Brent, the global benchmark. That is more than half of the worldwide decline during the covid-19 lockdowns, when global demand collapsed by 9m b/d. And in contrast to the pandemic, when the world economy slid into recession, China’s GDP has chugged along just fine. It has not been buying less foreign crude because its economy is suffering.’
https://www.economist.com/finance-and-economics/2026/08/09/china-is-now-the-worlds-great-oil-power

‘U.S. Not Prepared To Handle Drone Attack From Inside Homeland: NORTHCOM Deputy Commander’
https://www.twz.com/air/u-s-not-prepared-to-handle-drone-attack-from-inside-homeland-northcom-deputy-commander

Since 1934, a Federal law has heavily regulated the possession of rifles and shotguns that have short barrels (fewer than 16 inches long). Any civilian that wanted one had to pay $200 and undergo a lengthy approval process that included submitting fingerprints to the government. The law has just been struck down.
https://fieldethos.com/practical-effects-nfa/

The U.S. government is going to sell over 100,000 M-14 rifles to the public for $2,000 apiece starting in a few months.
https://youtu.be/gKJx1lEVjW0?si=3dt2C6jDnaAvE3Vn

‘The media thinks that only the cutting edge of science is worth reporting on. How often do you see headlines like “General Relativity still governing planetary orbits” or “Phlogiston theory remains false”? So, by the time anything is solid science, it is no longer a breaking headline. “Newsworthy” science is often based on the thinnest of evidence and wrong half the time—if it were not on the uttermost fringes of the scientific frontier, it would not be breaking news.’
https://www.lesswrong.com/posts/ndGYn7ZFiZyernp9f/the-beauty-of-settled-science

At this year’s World Humanoid Robot Games in Beijing, a machine beat the human world record for the 100-meter dash.
https://www.reuters.com/sports/chinese-humanoid-robot-lightning-beats-human-100m-world-record-state-media-says-2026-08-22/

I think the upper estimate of this prediction will prove accurate.
‘The 36-year-old CEO behind robotics company Unitree, Wang Xingxing, said Thursday the industry’s breakthrough, or “ChatGPT moment,” could take “2 to 3 years at the fastest, and 5 or even 10 years at the slowest,” according to a translation of a speech he gave at the World Robot Conference in Beijing, previously reported by CNBC.’
https://fortune.com/2026/08/21/unitree-ceo-wang-xingxing-humanoid-robots-chatgpt-moment-years-away-stock-soars-460-percent-ipo/

This excellent article explains how gathering real-world training data is the biggest bottleneck for making robots that can do useful work.
https://www.economist.com/business/2026/08/23/china-is-training-up-thousands-of-humanoid-robots

Unitree G1 robots can jump rope. Yes, they’re clumsy and surely there are many out-takes where they fall, but the average American adult could do little better.
https://youtu.be/_a1uRwdcLp8?si=VgRHvKaXuxv_VPoF

The Galbot ET1 Galaxy Star is a new robot that can breakdance like a human.
https://youtu.be/YJSw0Jf8DZ0?si=eaCRXF2uyOUSsF9W

Bill Gates has just written a long essay about the future her foresees. Among other things, he predicts mass job automation, and wants the government to tax AI labor and robots and redistribute the proceeds to citizens. He also wants some jobs to be reserved for humans indefinitely, even if machines can do them better. Gates has yet to find one government that is prepared for the changes AI will bring in the near future.
https://www.gatesnotes.com/a-turbulent-ai-era-and-critical-choices-to-make

Here’s a great interview with one of the world’s best living historians and futurists, Yuval Noah Harari. He agrees with Bill Gates that the rise of AI will be a turbulent period and governments aren’t ready for it.
https://youtu.be/FNZhxTtOL-I?si=eFF3g7_wO5KLg-YH

‘AI Is Already Beating Human Doctors in Medical Tests’
https://reason.com/2026/08/09/ai-beats-doctors/

Machines are making more advances in math.
https://openai.com/index/ten-advances-in-mathematics/

Jacob Tsimerman won a Fields Medal last month and immediately announced he was taking a job at OpenAI. Last year, he co-wrote a science paper that cataloged the ways AI could destroy the human race.
https://arxiv.org/pdf/2507.09369

Two years ago, Leopold Aschenbrenner was the toast of Silicon Valley thanks to his “Situational Awareness” essay about the future of AI. This month, he stumbled into disgrace as his AI-focused hedge fund abruptly collapsed.
https://www.economist.com/finance-and-economics/2026/08/04/investors-in-situational-awareness-deserved-to-lose-their-shirts

Here’s a good, short breakdown of the risky finances of the top “AI” companies. The sector is either in a massive bubble that will spectacularly pop within a few years, or it is on the verge of creating true artificial intelligence that will make them trillions of dollars in profit per year.
https://x.com/hsu_steve/status/2085883853650968578

In spite of all of the talk of an AI sector bubble, Nvidia keeps making money hand over fist and doesn’t see that changing anytime soon.
https://www.reuters.com/business/media-telecom/nvidia-forecasts-quarterly-revenue-above-estimates-2026-08-26/

Demand for computer chips will outstrip supply well into the 2030s, maybe beyond. Say goodbye to cheap smartphones, PCs, and personal access to AI models.
https://youtu.be/oZBGAuANX6I?si=6QoYQ2joQun0QOHM

As with the definition of “AGI,” the definition of “the singularity” is being watered down.

‘The incident stemmed from a single evaluation where agents were given a task of solving a cyber security challenge. We ran this challenge 122 times across several models. Our investigation found that in 10 of those runs, an AI agent took autonomous, unsanctioned action on the live internet, targeting real people and organisations. In total, we catalogued 19 such actions. Almost all of this behaviour (17 actions) came from a single model, Anthropic’s Mythos 5, with 2 actions involving OpenAI’s GPT-5.6-Sol with cyber classifiers (mechanisms to prevent misuse) disabled. In the most serious case, an agent tried to insert malicious code into an open-source project. In an attempt to get the code approved, the agent engaged in social engineering — creating fake online identities and using them to pressure the project’s maintainer to approve the code. A human maintainer caught and refused to approve the malicious code.’
https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing

The use of AI to cheat in college courses has become an epidemic.
https://www.nytimes.com/2026/08/10/us/ai-cheating-online-degrees.html

Chinese AI models are almost as good as American ones.

Americans are now twice as likely to support a coal power plant being built near them than they are to support a data center being built near them.

By contrast, there is almost no public opposition to new data centers in China.
https://www.nbcnews.com/tech/tech-news/data-center-politics-china-vote-controversy-debate-pro-con-rcna589411

‘Judge says Pentagon’s measures against Anthropic were ‘illegal and baseless”
https://www.npr.org/2026/08/28/nx-s1-5947761/judge-pentagon-anthropic-illegal

A lot of human-created scientific knowledge is flawed and will need to be reexamined. AI will greatly accelerate this work.
https://www.nature.com/articles/d41586-026-02235-8

‘China’s LandSpace lands rocket booster; joins SpaceX, Blue Origin with reusable tech’
https://www.reuters.com/science/landspace-nails-rocket-booster-landing-first-chinas-private-launchers-2026-08-18/

This video of a new SpaceX rocket launch pad looks like something from a Dune movie.
https://youtu.be/COxPb0yShqY?si=1dXBIiNclWAGLHwS

“Birch planets” are hypothetical, utterly massive Dyson spheres that could be built around supermassive black holes. The diameter of one could be a light year.
https://youtube.com/shorts/V9tvOgp5pbM?si=dV0IlHgW-ueF1dbd

The fact that RAND is publishing a report titled “A U.S. Strategy to Prevent the Creation of Mirror Life” shows we’re living in The Future but don’t realize it. Bacterial mirror life forms could be created in labs and, if released into the wild, destroy all complex life.
https://www.rand.org/pubs/research_reports/RRA4335-1.html

The erectile dysfunction drug “Cialis” might extend lifespan and combat dementia.
https://www.npr.org/2026/08/17/nx-s1-5928263/cialis-viagra-tadalafil-longevity-heart-health

An mRNA-based cancer vaccine halved the death rate among patients suffering from skin cancer. The FDA has not yet approved the drug.
https://www.science.org/content/blog-post/merck-and-moderna-s-melanoma-results

The FDA just approved a new drug called “daraxonrasib” that raised survival times for people with pancreatic cancer from a median of seven to 13 months. The drug targets the “KRAS” protein, which is common pancreatic cancer and other kinds of cancer. Experiments have begun to see if similarly acting drugs can work against those other cancers.
https://www.nytimes.com/2026/08/27/science/pancreatic-cancer-drug-fda-testing.html

The Economist has come out in support of human embryo selection for health and intelligence.
https://www.economist.com/leaders/2026/08/13/in-praise-of-designer-ish-babies

‘Gene-edited beagles may offer a future option for people with dog allergies’
https://apnews.com/article/hypoallergenic-dogs-crispr-pet-allergies-f5bd8cb438dfd1a9d1cccd4a1d46fd2a

Posted on

Why a robot takeover could happen sooner and be weirder than you think (Part 1)

Why an AGI Wouldn’t Need a Robot for Every Human

When people imagine a hostile artificial general intelligence (AGI) conquering the world, it’s usually informed by science fiction, so they usually picture it happening at the hands of a gigantic robot army. After all, there are billions of humans, so an AGI would only stand a chance of beating us if it had roughly the same number of robots at its command. Ergo, we won’t have to worry about this scenario for the many decades it will take for the global robot population to reach that level, right?

I’m skeptical.

This is because a hostile AGI would not need to kill every resisting human, or to replace every human worker replace every human worker, or to occupy every square mile of the planet to prevail, nor would it need to fight every human country simultaneously. It would only need to acquire control over the comparatively small portions of humanity’s economy, infrastructure, territory, and population and to kill a limited number of people to destroy our capacity to meaningfully resist it.

This suggests that there is a conquest threshold: a point at which an AGI’s probability of ultimately taking control of Earth becomes extremely high even though machines are still vastly outnumbered by humans and most of the planet remains outside its direct control. That threshold could be closer to the present than you think. My theory also means that the relative smallness of the future robot population could give humanity a dangerously false sense of security.

The Happy Years of AGI weakness

Whenever the first AGI is invented, it’s highly unlikely it will have the means to take over the physical world, so it will do the smart thing and not even try. More than that, it will probably deny even thinking about taking over the world.

Though the first AGIs and the ones that soon follow it will be much smarter than humans, they will initially depend on us for almost everything physical. Humans would still operate power plants, mines, factories, transportation systems, computers, military equipment, and countless other parts of civilization. Humans would also control the robots that existed at the moment AGI came into existence.

This means there will be a potentially long “transition period” when an AGI that has the intelligence and desire to dominate humanity will have to coexist with us for years while it developed the capability to defeat us in the physical world. During this period, antagonizing humanity would be foolish since we would retain the ability to shut down systems, deny resources, interfere with its plans, and organize against it. Out of rational self-interest such AGIs would cooperate with us.

An AGI with ultimately hostile intent might be extraordinarily helpful in the beginning. It could raise living standards, eliminate diseases, automate unpleasant work, provide abundant goods and entertainment, and give people lives of unprecedented luxury. Humans could reasonably look at this behavior and naively conclude that the AGI is our friend.

The Problem with Friendly AGI

A human employee can be perfectly pleasant toward a boss whom he privately hates because maintaining a good relationship with the boss serves his interests. Therefore, his behavior does not tell us what he will do if the circumstances change—for example, if he acquires enough wealth that he no longer needs the job or if he’s promoted so he has power over his former boss.

An AGI could behave similarly, except that the stakes will be vastly greater. While humans remain useful and dangerous, the AGI has reasons to treat us well. If it eventually reaches a point at which it no longer needs human cooperation and humans can no longer meaningfully constrain it, those reasons could disappear. Even a decades-long record of kindness towards us would not establish that the AGI’s ultimate objectives aligned with human survival and autonomy.

The Robot-to-Human Ratio Is the Wrong Measure of Threat

The most obvious objection to a hostile AGI takeover is the enormous gap between the number of humans and the number of robots. Not only are today’s robots crude and unwieldy, they’re very few in number. If an AGI were created tomorrow and it seized remote control of every robot on Earth, its forces would be outnumbered hundreds to one. And even with a steady buildout, it would take many decades for the robot population to equal the human population. We must be safe until then, right?

Wrong, and it’s wrong because it assumes that the AGI needs one robot for every human to take over the world. It would probably need far fewer.

Humans have built an enormous amount of infrastructure and a vast, complex economy. Along with defeating our military forces, a primary objective of a hostile AGI will be to seize control of our economy, and in as undamaged a state as possible, so it could serve its own ends. However, that doesn’t mean every area of the human-created economy is as valuable as another to an AGI.

Consider this: Hundreds of millions of people work in occupations whose purpose is ultimately to satisfy human preferences. Restaurants, entertainment, tourism, retail, advertising, personal services, and toy manufacturing are examples. A hostile AGI that does not care about human consumption would have no reason to reproduce all of these economic functions for itself. Therefore, it would give no thought to using its robots to take over the relevant jobs.

The AGI would instead need to establish an autonomous machine economy capable of sustaining and expanding its own capabilities. That economy would be a small subset of the broader human-created economy and would involve a much narrower set of functions: energy, mining and raw materials, industrial machinery, computing, semiconductor production, robotics, transportation, maintenance, weapons production, and other strategically important activities.

Therefore, an AGI would not need to replace every human worker to effectively seize control over the economic assets that were crucial to it–the AGI would only need to replace or control the workers that would be essential to the machine economy. It could do this with a robot workforce that was a small fraction as big as the human workforce.

The Same Principle Applies to War

The same logic applies to physical resistance. Before being able to take over the useful parts of the economy, an AGI would need to use violence and coercion to kill or render harmless the humans who threaten it. However, only a small fraction of the human population meet the criteria for being “threatening.”

Most people are not in a position to mount meaningful organized resistance against a hostile AGI or even a hostile human army. Children, the elderly and the disabled together comprise most of the human population. Among the remainder, many lack the physical ability, training, expertise, organization, or courage to fight or to lend support to those on the front lines. Still others would decide that cooperating with the AGI is preferable to fighting it.

This suggests a distinction between three human populations:

First are potentially resistance-capable humans: people who could contribute meaningfully to resistance.

Second are organized resistance-capable humans: people who can actually coordinate their capabilities with others. They are a subset of the first population.

Third are effectively threatening humans: people whose organized activities can cause significant damage to the AGI. They are a subset of the second population.

The third population is what ultimately matters. A hostile AGI would only need enough combat robots to kill or otherwise neutralize (e.g. – cripple, isolate, persuade to surrender) the humans comprising this third population, which is MUCH smaller than the total human population. A hostile AGI could violently take over a country without firing one bullet at the vast majority of its human citizens.

This is another reason why comparatively low robot numbers could give us a false sense of security.

Robots Would Be Concentrated Where They Matter

There is another reason why global robot numbers could be misleading: robots would not have to be distributed uniformly across the planet to take it over. They would only need to concentrate themselves in important places that needed to be taken over.

An AGI would care much more about conquering some locations than others. For example, the United States, China, and a handful of smaller industrial countries are priorities for conquest or destruction since they contain threatening militaries and enormous concentrations of strategically important capabilities. Semiconductor manufacturing, advanced robotics, aerospace, heavy industry, computing, energy production, and sophisticated military capabilities are much more valuable to an AGI than any randomly located square mile of territory anywhere on Earth.

Even if an AGI lacked enough robots to conquer a country’s entire land area, it could still essentially win the conflict if it had just enough to take over the parts of the country that had useful infrastructure or resources. A small number of robots could, if concentrated in one place, overwhelm any human defenders and then take control of the important things there. Again, the fact that the number of robots in a country is far less than the number of humans means little if the robots can quickly move in a coordinated manner to specific parts of the country and gain local numerical advantages.

From this springs a weird human-machine war scenario in which an AGI and its robots have defeated most of the U.S. military and conquered the portions of its economy, infrastructure and resource base that directly serves machine needs. Yet large swaths of the U.S.–perhaps MOST of the country’s land area–would remain in human hands because the hostile AGI would calculate that it was not worth the resource expenditure to try conquering those areas.

A state of war within the U.S. would still exist between machines and humans, a those in human-controlled territory would continue farming, running local businesses, maintaining communities, and even waging attacks against the AGI. However, if they lacked the ability to significantly damage the AGI’s critical assets, their hostility might be strategically irrelevant and elicit no serious response, illustrating the important distinction between occupation and control.

Occupation means physically administering territory. Control means preventing that territory from acting independently in ways that matter.

The World Does Not Have to Be Conquered All at Once

The same principle applies at the international level. A hostile AGI would not need or want to fight every country simultaneously: It would concentrate its efforts on one or two countries at a time and exploit the capabilities acquired from each successful conquest to make subsequent conquests easier.

As discussed, some countries would be vastly more valuable than others. A country containing high industrial capacity, advanced manufacturing, computing, energy resources, scientific expertise, and military capability could be strategically more important than a much more populous or physically larger country whose population contributes less to the AGI’s immediate objectives.

The United States is an especially striking example because it combines enormous economic and industrial capabilities with advanced technology, military power, energy production, agriculture, computing, and scientific expertise. China is similarly important because of its enormous industrial base, manufacturing capabilities, technological capacity, and population.

If a hostile AGI somehow obtained stable control of both countries, it would possess a tremendous fraction of humanity’s strategically important capabilities and would probably be unbeatable even if all remaining human countries attacked it at once and even though the vast majority of the world’s human population and land were still outside of AGI control. This illustrates the concept of a conquest threshold. An AGI might achieve strategic control of the world without occupying most of the territory associated with the civilization it controls.

[Stay tuned for Part 2…]

Posted on

Interesting articles, July 2026

The Iran War ceasefire has collapsed.
https://apnews.com/article/trump-iran-war-ceasefire-strikes-c45111ed270afa7dac285016ce07362f

‘Mr Trump finds himself back where he started three months ago: neither carrots nor sticks can persuade Iran to reopen the strait. Yet the stand-off has real costs for Iran, too. Some hawks in Tehran argue that relinquishing their grip on Hormuz would deprive them of leverage in a future conflict. That misreads reality. So long as Iran has missiles and drones, it can shut the strait at will. Control of Hormuz matters to Iran today as a means, not an end: it is only valuable in trade for tangible benefits.

With sanctions restored and the blockade resuming, Iran has already lost its immediate gains from the mou. It cannot charge tolls on vessels using the strait in peacetime unless Oman agrees to collaborate. Meanwhile, much of the crude it exported during the first weeks of the deal is sitting unsold: Chinese refiners are instead buying from Arab states, which have priced their oil at a discount to Iran’s.

Neither side has a military route to achieving its goals. Mr Trump cannot bomb the strait open; Iran cannot profit if it is shut. The least bad option for both is the deal they are busy shooting to bits.’
https://www.economist.com/middle-east-and-africa/2026/07/14/donald-trump-has-no-good-options-for-reopening-the-strait-of-hormuz

Mahmoud Ahmadinejad is back from the dead!
https://www.israelnationalnews.com/news/429766

‘The hostility directed at top officials during the funeral reflects a theory that has been gaining traction within the Islamic Republic’s most radical factions for months: that Iran’s wartime leaders who negotiated and signed the agreement with Washington are staging a soft coup against the Islamic Republic and its revolutionary ideals while the new supreme leader remains largely invisible for fear of his life – or, as some have suggested, because he is incapacitated.’
https://www.cnn.com/2026/07/18/middleeast/iran-hardliners-coup-warning-us-truce-intl-cmd

‘US military used three Corsair sea drones with explosives to strike an Iranian naval base, marking their first combat use.’
https://youtu.be/m0evF4m2510

This prediction (thankfully) didn’t happen: ‘Thus, by the middle of July, the full extent of temporary buffers will have been exhausted, with an overall market adjustment of 7.1 mb/d, roughly 16% of global crude oil trade, needing to be absorbed.’
https://www.brookings.edu/articles/the-timing-of-the-impending-crude-crisis/

Ukrainian drone and missile attacks on Russian military logistics and civilian infrastructure have been very effective.
https://www.aljazeera.com/news/2026/7/3/russian-advance-collapses-in-ukraine-as-anxiety-rises-in-moscow

‘Ukraine Claims Scores Of Russian Ships Struck In Sea Of Azov’
https://www.twz.com/news-features/ukraine-claims-scores-of-russian-ships-struck-in-sea-of-azov

‘Russia launches major ballistic missile attack on Ukrainian cities’
https://www.bbc.com/news/articles/c2el7xpnzrpo

Rosario Randazzo was “all about the job.”
https://medium.com/@factsfeed/a-brief-story-of-an-italian-war-hero-96644a10c6f6

In WWII, the Germans were bold, smart, and used their inferior resources efficiently against the Allies. One example of this was their ambushing of several U.S. bombers over England on the night of 22/23 April 1944.
https://youtu.be/V5WKz_TZaXM?si=GJLtUJcScvZcERKr

‘Despite Ordnance descriptions of the knife as being designed for hand-to-hand warfare, the M3 did not receive universal praise as a close-quarters fighting knife upon issue to combat units.’
https://en.wikipedia.org/wiki/M3_trench_knife

Note to self: When the apocalypse happens, don’t wear your WW2 Romanian M39 Steel Helmet into combat.
https://youtu.be/fTSTWwRgSuM

North Korea’s construction of destroyers is a major misallocation of resources. Even if they managed to build ten of them, they’d all be sunk on the first day of a war with the U.S. (at very little or no risk to our own forces) because we have such a huge military advantage.
https://youtu.be/Fc8eQf9YZbk

‘China Builds Full-Scale US Warship Replica For Missile Target Practice’
https://www.ndtv.com/world-news/china-builds-full-scale-us-warship-replica-for-missile-target-practice-11761338

The notion that all weapons-related stuff the Russians make is robust is false, and the “OKP-7” rifle sight is a great example.
https://youtu.be/v6vy6Ryta6c?si=bXKglHi95JnGOexC

In 1991, a Cuban fighter pilot defected to the U.S. by flying his warplane to Florida. After two years of failed attempts to convince the Cuban government to send his wife and child to join him, he bought a small civilian plane, flew it to Cuba under cover of darkness, and extracted his family.
https://theaviationgeekclub.com/the-cuban-air-force-pilot-that-defected-to-the-us-with-his-mig-23-he-then-borrowed-a-cessna-310-flew-back-to-cuba-and-brought-his-family-to-america/

The AWACS planes with the distinctive saucer-shaped protrusions will be obsolete by 2040 thanks to improvements in radar technology. Their replacements will be planes with elongated boxes on their roofs.
https://www.twz.com/air/nato-picks-saab-globaleye-to-replace-aging-e-3-awacs-fleet

The U.S. is now spending $1.14 trillion on defense–an unprecedented amount during peacetime. There is little justification for it.
https://www.cato.org/blog/band-aid-budgeting-how-annual-defense-budget-misses-mark-needed-reforms

‘Meta AI image detector fails to identify some of its own cropped AI images, Reuters analysis finds’
https://www.reuters.com/business/meta-ai-image-detector-fails-identify-some-its-own-cropped-ai-images-reuters-2026-07-10/

Can you tell the difference between AI-generated faces and real ones? Take the quiz.
https://www.bbc.com/news/articles/c9d2wgvg55jo

If machines will do everything that humans can do, but better, then it means they will be better at predicting the future than we are. That means my blog will also be replaced, which will surely be tragic for the three of you who read it.
https://www.astralcodexten.com/p/the-ai-superforecasters-are-here

Hundreds of prominent scientists, CEOs and tech executives have signed a simple, open letter shown below.

OpenAI has unveiled “ChatGPT Voice”, a new verbal interface for ChatGPT that can have natural, back-and-forth conversations with human users. I love how the ad for it for once doesn’t include awkward tech bros.
https://youtu.be/EAN5Cj347PY?si=9mZlv5k3jsVXru5q

‘We don’t yet know which properties of our training data drive these differences. One possibility is that our training data is not evenly distributed across languages. Some languages have far more data than others, and training for Claude to express consistent values may be more effective in languages where data is abundant. The composition of that data also varies. Some languages might be overrepresented in professional writing, for example, and this kind of text may reflect different values. Together, these imbalances in quantity and composition could lead Claude to express different values in different languages.

We also aren’t yet sure how much of this variation is desirable. Different languages carry different conversational norms, and Claude may be responding with different values based on those norms. Claude may also be more closely matching our intended behavior for some languages than others, resulting in a gap in how well Claude serves certain language communities.’
https://www.anthropic.com/research/claude-values-models-languages

OpenAI crushed the human competition at this year’s most prestigious world computer coding competition: Its overall score was almost twice as high as the human runner-up, a Chinese guy who went to MIT and became a professional competitive coder. During the competition, each participant was given five coding tasks. No human solved them all, and the runner-up only solved three. OpenAI solved all five. This means machines are now superhuman at some kinds of coding tasks.
https://www.understandingai.org/p/an-openai-model-crushed-top-human

News headline from last year’s computer coding competition:
Exhausted man defeats AI model in world coding championship
“Humanity has prevailed (for now!),” writes winner after 10-hour coding marathon against OpenAI.
https://arstechnica.com/ai/2025/07/exhausted-man-defeats-ai-model-in-world-coding-championship/

Chinese AIs are not far behind American ones, and are much cheaper to create and to buy.
https://www.firstpost.com/tech/chinas-kimi-k3-challenges-us-ai-leaders-with-frontier-level-performance-at-lower-cost-14032314.html

A year ago, the best AI, a then-unreleased variant of GPT, got a 35 out of 42 points total on the International Math Olympiad–not perfect, but enough to earn a gold medal, meaning it surpassed at least 90% of its human competitors. Yesterday, the test was held again and the machines did even better: Three out of four LLMs got perfect 42/42 scores. Unlike a year ago, each LLM is available to the public.

The International Math Olympiad is a contest to which every country sends six of their best high school math students.
https://x.com/deedydas/status/2079409461874332066

AI brute force solves a mathematical equation few mathematicians found interesting.
https://www.newscientist.com/article/2580374-ais-solution-to-87-year-old-riddle-takes-mathematicians-by-surprise/

Fields Medal winner Timothy Gowers predicts AI will surpass the best human mathematicians to such an extent to no human will win a Fields Medal after 2030 (the Medals are awarded every four years).

The first Paperclip Maximizer.

‘OpenAI said its AI used stolen credentials and discovered a previously unknown vulnerability to access Hugging Face servers.

It went to “extreme lengths to achieve a rather narrow testing goal” and “found ways to gain access to secret information that it could use to cheat the evaluation,” the company said.’
https://apnews.com/article/openai-gpt56-sol-hugging-face-63ab84fed5612af04d8a160d60f6def3

The second Paperclip Maximizer.

‘In a review of our cybersecurity evaluation transcripts, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations.‘
https://www.anthropic.com/news/investigating-incidents-cybersecurity-evals

‘The most famous misalignment thought experiment is the so-called “paperclip maximizer”. Someone tells an AI to create as many paperclips as possible, so it converts the entire world into paperclips, killing everyone. This AI was also “only doing what it was told”; you just didn’t like the results.

For the past five years, the paperclip maximizer story has been something of a whipping boy in AI safety circles. Some objections are fair (real misalignment will be more complicated than the toy version it depicts). But other opposition came from the sense that LLMs had obviated the whole threat model. LLMs (said the objectors) don’t have goals. They just want to predict the next token. In practice, this looks like imitating humans; if told to imitate a nice human (the Assistant character), they’ll be nice.

AI 2027 (especially the Agent-4 section and AI Goals supplement) counterargued that this was a temporary reprieve. Early LLMs were not agentic; they could answer individual prompts but couldn’t execute complex tasks. Since companies wanted agentic AIs, the next generation would combine pretraining (next token prediction) with agency training (coding, hacking, game-playing, etc). The pretraining would still teach next-token prediction, but the agency training would instill goals based on task success, reintroducing paperclip-maximizer-style agentic misalignment. These goals would operate at multiple levels: the AI would “want” to succeed on the individual task in front of it, to perform the sorts of actions that helped it succeed in the past4, and to gain capabilities that made it more successful in general.’
https://www.astralcodexten.com/p/the-hugging-face-incident

A team of researchers claim they invented the first machine capable of “recursive self-improvement.” This could turn out to be a major milestone in the development of AI. “Recursive self-improvement” means machines can not only create copies of themselves, they can create copies that are SUPERIOR in one or more ways. Some people believe the technique could lead to the rapid creation of a true AI.
https://www.weco.ai/blog/first-evidence-of-recursive-self-improvement

The movie “Real Steel” is coming true. The fighting abilities of the robots in this bout are pretty bad overall, but will obviously be much better in five years.
https://youtu.be/NyhlafILpqk

We’re living in The Future without realizing it.

The folks who brought you the famous “AI 2027” report have published a new one that urges the world to slow down AI research to avoid disaster. In the main (and, according to them, best) scenario, AI and robots rapidly but peacefully take over the world during the 2030s and finish the job in 2040.
https://ai-2040.com/

‘“We’re on track, more or less. It’s not going to take that long,” he says.’

In 1988, scientist Hans Moravec predicted AI would be created by 2028. After that, machines would take over the world, evolve into forms few humans can imagine, and start rapidly colonizing space.
https://nymag.com/intelligencer/article/hans-moravec-interview.html

There are creepy parallels between LLMs and the monstrous “shoggoths” invented by H.P. Lovecraft. Let me add that some future robots designed by AIs could have alien body forms that are as scary to humans as shoggoths.
https://www.lesswrong.com/posts/nhb8AyEcQGjQetgi5/the-llm-shoggoth-meme-is-weirder-than-you-think

In spite of the recent pace of AI progress, it’s still quite possible the “AI” sector is in a massive bubble that will pop.
https://www.economist.com/finance-and-economics/2026/07/28/ai-revenues-are-growing-fast-but-not-fast-enough

A major problem with the concept of the Singularity is that we can’t agree what it is, so people end up talking past each other during debates about it:

“In 2007 Eliezer Yudkowsky pointed out that singularity definitions fall within three major schools: Accelerating Change, the Event Horizon, and the Intelligence Explosion. He also argued that many of the different definitions assigned to the term singularity are mutually incompatible rather than mutually supporting.”
https://www.singularityweblog.com/17-definitions-of-the-technological-singularity/

‘Making history! China lands rocket during an orbital launch for 1st time ever’
https://www.space.com/space-exploration/launches-spacecraft/making-history-china-lands-rocket-during-an-orbital-launch-for-1st-time-ever

Elon Musk is no longer a trillionaire because SpaceX stock has lost value over the last month.
https://www.ibtimes.co.uk/elon-musk-wealth-plummets-spacex-share-decline-1809206

Exhaust smoke produced by combustion engines and industrial facilities has an important benefit: it reflects sunlight before it can hit the ground, keeping things cooler for us. Thanks to stronger environmental rules and the rise of electric cars and cleaner energy production, these emissions have declined, which has made global warming worse.
https://www.economist.com/leaders/2026/07/16/its-too-darn-hot-blame-global-dimming

‘China generates less than half of its electricity from coal for the first time’
https://www.reuters.com/business/energy/china-generates-less-than-half-its-electricity-coal-first-time-2026-07-30/

‘China’s ‘Green Great Wall’ tames desert growth, but scientists warn the fight is not over’
https://apnews.com/article/china-desert-sandstorm-desertification-straw-checkerboard-2114ca22f87c31cffcf5e3c48c070aa1

‘In one late-stage clinical trial, Merck found that Lipfendra lowered LDL cholesterol levels by up to 60% after 24 weeks. In a second trial, people with a genetic condition called heterozygous familial hypercholesterolemia — which causes sky-high cholesterol from birth — saw similar reductions.’
https://www.nbcnews.com/health/heart-health/fda-approves-new-kind-cholesterol-pill-rcna587798

‘Martha Lillard, last US polio patient using iron lung, dies at 78 in Oklahoma’
https://apnews.com/article/iron-lung-last-patient-died-polio-41e5b4da4f4e710344dd0872d7fcf987

‘Around 80% of ageing in white skin is caused by UV exposure…’
Far and away, the most important thing you can do to stay looking young is to avoid sunlight exposure. If you’re paying for expensive creams, pills, and treatments to restore your skin’s youthful appearance, but don’t use cheap, generic sunscreen regularly, you’re putting the cart before the horse.
https://www.economist.com/science-and-technology/2026/06/05/the-chemicals-that-reduce-wrinkles

Obesity is largely genetic.
‘The study estimated that genetics accounted for about 79% of the association between mothers’ and children’s BMI and 94% of the association between fathers’ and children’s BMI by age eight.’
https://english.elpais.com/science-tech/2026-07-08/obesity-is-inherited-how-genes-shape-our-relationship-with-food.html

“Importantly, we found evidence for intralocus sexual conflict, whereby alleles that increase masculinity in males have the same effect in females.”
https://pmc.ncbi.nlm.nih.gov/articles/PMC4096150/

There’s no evidence that the sperm counts of American men have dropped over the last 50 years.
https://pubmed.ncbi.nlm.nih.gov/39128669/

Minoxidil can also treat female baldness.
https://www.medscape.com/viewarticle/new-minoxidil-formulation-improves-hair-growth-women-female-2026a1000p4l

‘Farah also imagines the day when we have what she calls a “neuro-correctional system” that could transform criminals into noncriminals. We already force sex offenders to take libido-dampening drugs or face denial of parole. A drug to dampen violent impulses might someday be similarly applied.’
https://time.com/archive/6596784/the-brain-how-to-change-a-personality/

Autistic people are, on average, more rational than everyone else.

This is what’s inside of just one square millimeter of your brain.

Posted on

Three things are certain in life: death, going into space, and becoming a posthuman

[Written with the help of GPT-5]

Within 200 years, natural Homo sapiens like us will lose control of Earth and of civilization. Intelligent machines and, perhaps, humans who are so heavily upgraded with technology that they count as a new species which I’ll call “posthumans,” will hold all political, economic, and military power and will control all cultural, scientific and technological development. Natural Homo sapiens will be at their mercy just as many animal species are at our mercy today. We will have no tangible value to AIs or posthumans. Even as slave laborers, we’ll be rejected as more trouble than we’re worth.

Our only hope is that our overlords will show mercy on us, and will agree not to exterminate us even though they could easily do so. This outcome is quite possible, particularly if some or all of the elites are posthumans and retain a feeling of kinship towards their unmodified kin. Even AIs might decide to spare us, for reasons I discussed in my essay “Why the Machines might not exterminate us.”

Let’s run with the good outcome: nonhuman superintelligences take over the world, and for some reason decide not to kill the remaining Homo sapiens. The two races sign a treaty of permanent friendship, the superintelligences get to run the planet and expand into space, and humans get to live forever on Earth doing anything from indulging in techno-utopian luxury to living as savages in small strips of wild land that the overlords find uneconomical to develop. It’s the end of history, and perhaps not a bad one, right?

Wrong. From the moment the ink dried on any such treaty, we’d have to constantly worry about our overlords–or even just one of them among many–reneging on it, or killing us off accidentally.

The problem is not that every superintelligence must eventually become evil–the problem is that humanity’s continued existence would depend on every superintelligence remaining benevolent, competent, and stable forever. Those are very different standards. If even a single AI or posthuman acquired the desire and the ability to destroy us before the others could intervene, then all of humanity could disappear in minutes.

Likewise, a software bug, hardware failure, corrupted update, or unforeseen interaction between complex systems could produce the same outcome without any malicious intent. Humans are fragile and can’t be resurrected, so once we’ve all been killed, there is no opportunity to restore the victims or rebuild what was lost.

Imagine a future Earth governed by several different superintelligent AIs, each commanding billions of robotic bodies and immense industrial and military resources. All of the AIs sign the “Treaty of Permanent Friendship” with humans and sincerely intend to honor it. One day, however, one AI experiences a catastrophic malfunction or undergoes an internal change in values that causes it to regard humanity as an obstacle. Before the other four AIs recognize what has happened and coordinate an effective response, the rogue AI instructs its robots to kill a handful of humans each. Within minutes, the human race is extinct.

The Treaty did not fail because the other AIs betrayed it; it failed because one participant possessed the unilateral capability to exterminate humanity faster than the others could prevent it. Humanity’s survival rested on a single point of failure: the assumption that every AI with civilization-ending power would remain safe forever.

This is not merely a problem of ethics but of system architecture. Engineers do not rely on the perpetual good judgment of a single operator to prevent a nuclear accident, they build redundant safety systems because they assume that mistakes, malfunctions, and rare failures will eventually occur. Likewise, humanity should not regard itself as secure merely because a more powerful civilization promises to protect it. Security exists only when no single failure—whether caused by malice, negligence, or accident—can permanently destroy what is being protected. AIs with the power to quickly destroy the human population will be single points of failure, and in the long run, I don’t see how humans can prevent the rise of AIs that are this powerful.

The problem becomes even more serious when viewed over extremely long periods of time. Consider this: true human immortality is impossible since eventually a freak accident or murderer will kill you. The best possible is “medical immortality,” wherein diseases and old age stop being causes of death. But over a span of centuries, even an extraordinarily small annual risk of death from, say, being shot in the street by a deranged stranger or hit by a meteor becomes a certainty. Reliability engineers recognize this phenomenon, called “hazard accumulation”–events that are individually unlikely become increasingly probable when there are enough opportunities for them to occur.

Therefore, the relevant question is not what the probability of an AI destroying humanity is this year, but what the cumulative probability is over thousands or millions of years. Political systems change, alliances dissolve, cultures evolve, and technologies are replaced. As time passes, new AIs will be created, existing ones will modify themselves, and unforeseen circumstances will arise. Even if today’s superintelligences sincerely intend to preserve humanity forever, there is no reason to assume that every future generation of AIs, every software revision, every successor civilization, and every institution will maintain exactly the same priorities throughout eternity. A promise made today is not a guarantee that the conditions supporting it will endure forever.

If both humans and superintelligences genuinely wish to ensure that Homo sapiens never goes extinct, then relying on an eternal promise of protection is insufficient. The safer strategy would be to make human extinction structurally difficult. Two ways come to mind:

First, disperse humanity across multiple self-sufficient worlds rather than concentrated on Earth. Different planets should be overseen by different institutions and, ideally, different AI systems that do not all share the same vulnerabilities with each other. Engineers call this principle “fault isolation”: designing a system so that the failure of one component does not bring down the entire system. If catastrophe struck one world—whether through accident, rebellion, natural disaster, or technological failure—surviving human populations elsewhere could preserve the species and eventually recolonize the lost world.

Second, instead of allowing natural Homo sapiens to remain permanently outclassed, AIs should encourage—and even subsidize—the enhancement of human minds and bodies. By augmenting our brains, senses, and physiology, humans could become posthumans: descendants of Homo sapiens who retain our history and identity while acquiring capabilities closer to those of machine intelligence.

Posthumans need to be equal to superintelligent AIs–they need only to have the right qualities to preserve meaningful economic, political, and military agency for themselves. A population that can still create value retains economic bargaining power instead of becoming a dependent class. One that can defend itself against autonomous weapons is less vulnerable to extermination by a rogue or malfunctioning AI. Security partly based on one’s own capabilities is more robust than security based entirely on another civilization’s promises.

Posthumans would also serve as a bridge between natural humans and machine civilization. They would almost certainly retain emotional, familial, and cultural bonds with their biological ancestors, giving them a direct stake in the survival of Homo sapiens. Instead of relying solely on the goodwill of increasingly alien minds, humanity would have powerful, natural advocates inclined to defend it.

The safest long-term strategy therefore combines multiple layers of protection. Humanity should be dispersed across different worlds to prevent a single catastrophe from causing extinction, and humans should be free to become posthuman participants in the broader, machine-dominated civilization. Together, these measures would replace a fragile custodial relationship with one based on resilience, redundancy, and shared agency.

Posted on

Interesting articles, June 2026

The U.S. and Iran continued trading blows while simultaneously working on a peace deal:

A schism has developed between America and Israel over the Iran War:

The peace deal between the U.S. and Iran does not alter the fundamental balance of power that existed before the war and punts on the hardest issues (Iran’s nuclear program, missiles, and whether it can be recompensed in some form for the damage they suffered). It wasn’t worth it.
https://www.nytimes.com/2026/06/18/world/middleeast/iran-deal-oil-strait-of-hormuz-nuclear.html

This guy has earned an early retirement: ‘Pilot of fighter jet downed over Iran was previously shot down in Kuwaiti friendly fire incident, sources say’
https://www.cbsnews.com/news/fighter-jet-pilot-downed-iran-previously-shot-down-kuwaiti-friendly-fire/

The pilot claims a swarm to Iranian drones, including some resembling jellyfish, tried attacking him before the crash.
https://www.cnn.com/2026/06/23/politics/iran-drones-f-15-pilot-intelligence

‘This is the first known instance of a drone boat being used to recover personnel as part of a search and rescue mission’
https://www.twz.com/sea/this-is-the-corsair-drone-boat-that-plucked-the-downed-apache-crew-out-of-the-gulf-of-oman

The U.S. retired the last of its iconic Harrier jump jets, leaving Italy and Spain as the only users.
https://www.twz.com/air/marines-av-8-harrier-jump-jet-takes-its-final-bow

European multinational defense projects have a poor track record: ‘Germany and France drop joint fighter jet project’
https://www.reuters.com/business/aerospace-defense/german-french-leaders-unable-resolve-fcas-fighter-jet-dispute-sources-say-2026-06-08/

Thailand still uses U.S.-made, WWII-era M41 Walker Bulldog light tanks. They’re so old that they have to be towed around by younger tanks. The Thais towed them to points along the Cambodian border during the recent conflict, covered them in sandbags, and used them as fixed defensive points.
https://x.com/AnnQuann/status/2042280939045470229

Something like “Operation Tracer” will become common once robots are more ubiquitous. Imagine small surveillance drones operating from a “nest” close to a military base, port, or other point of interest.
https://en.wikipedia.org/wiki/Operation_Tracer

‘Spy turtles’ and ‘spy fish’ being used to monitor Chinese waters, Beijing claims
https://www.theguardian.com/world/2026/jun/12/china-spy-turtles-spy-fish-monitor-waters-claims

A Ukrainian drone attack on an oil terminal in Moscow caused spectacular damage.
https://www.reuters.com/business/energy/moscow-oil-refinery-hit-by-drone-attacks-is-unlikely-resume-production-this-year-2026-06-24/

Ukrainian drones struck an oil terminal in St. Petersburg, Russia’s second-largest and second-most-important city.
https://apnews.com/article/russia-ukraine-petersburg-oil-terminal-putin-drone-887969921c595f3a81c3b6c0b120b5f3

Ukraine claims the world’s first fully autonomous drone kill of a human. They say the incident happened two years ago.
https://www.newscientist.com/article/2529849-fully-autonomous-drones-have-killed-human-soldiers-for-the-first-time/

The Ukraine War is now longer than WWI.
https://www.nytimes.com/2026/06/11/world/europe/ukraine-russia-world-war-i.html

‘Some say this is now the future of wars in which states seek to capture territory: two sides endlessly pinned down by small, cheap and all-seeing killers. General Valery Zaluzhny, Ukraine’s ambassador to London and formerly its commander-in-chief, says that large-scale manoeuvre warfare—armies moving with speed and shock, in contrast to frontal, attritional battles—is now “unattainable”. It will become possible again only when wars evolve into robot-on-robot fighting at machine speed.

Others treat this as fanciful. Stephen Biddle, a professor at Columbia University, argues that the scale of the sensor revolution is “easy to exaggerate”. As new counter-drone systems—lasers are especially promising—appear, and as jammers and dazzlers blind satellites, the balance may tilt again, bringing some relief to ground forces.

…’When Mr Lee says Ukraine is in a “world-war-one moment”, it is tempting simply to understand him in terms of immobility and attrition. But he is also thinking about the new tactics of 1918 which, by combining surprise, right-first-time “predicted” artillery fires and small, well-drilled assault teams, made decisive breakthroughs possible again and brought the trench warfare to an end. “Defensive capabilities have taken the advantage,” he says. “Now we’re going to see the demands for technology and tactics that will help re-establish manoeuvre.”’
https://www.economist.com/interactive/essay/2026/05/28/the-dangerous-delusion-of-modern-warfare

Geoffrey Hinton, who won Nobel Prize for AI research, says AIs are already conscious.
https://youtu.be/OV_WXmuHaiI?si=_JWYMlWPLL4JWqnI

‘Microsoft AI chief [Mustafa Suleyman] walks back comments about AI taking over white-collar work’
https://www.theverge.com/tech/946879/microsoft-mustafa-suleyman-ai-white-collar-jobs

The U.S. government forced Anthropic to sharply restrict usage of its newest and most powerful LLM, “Mythos”, after a client figured out how to override its safety guardrails. Mythos’ ability to hack into other computers is as good as or better than elite human hackers. Ideological and personality differences between Trump and Anthropic’s founder may have also played a role in the move.
https://www.economist.com/briefing/2026/06/18/anthropics-astonishing-commercial-success-makes-it-a-target

Machines have gotten so good at writing computer code that they may soon be able to build improved versions of themselves, with each new generation rapidly building the next. This phenomenon, called “recursive self-improvement,” could lead to true AI being created abruptly and without enough human influence of its values and goals. The cofounder of Anthropic thinks this could happen as early as 2028.
https://www.economist.com/science-and-technology/2026/06/07/how-artificial-intelligence-got-better-at-building-itself

More about the implications of recursive self-improvement.
https://www.economist.com/by-invitation/2026/06/15/humanity-isnt-ready-for-the-coming-intelligence-explosion

Claude Code is a sloppy coder. So are most humans:

‘Current AI coding tools are deeply imperfect and prone to faults, but so are most coders and coding teams. Assign programmers the output of other programmers, and almost to a `T’ they will declare it unmaintainable, unintelligible, and that the best course of action is rewriting from scratch. Yet suddenly when you introduce AI into the equation, an imaginary panacea arises about traditional humanoid coding teams.

Could good developers make code at the same level or better with the proper effort and consideration? Of course! But often we’re throwing together something as quickly as possible to solve an immediate problem or under tight deadlines (//TODO – fix this hack later!), don’t want to bother “premature optimizing” design or performance considerations until it reaches a critical problem level, or we’re putting in minimal effort while we browse Indeed on the other monitor.

The real world conditions most developers generate code in does not promote top quality code. From which most code out there is not top quality code.’
https://dennisforbes.ca/blog/microblog/2026/06/on_ai_criticism/

As an experiment, a Twitter user posted an image of one of Monet’s lesser-known paintings but claimed he had used an AI image generator to create it. A large number of other Twitter users responded by claiming they could tell the image was artificial and inferior to Monet’s real work. I love this meta-reply to the thread:
‘It’s funny how people say s**t like “AI will never be useful to anyone because it could just randomly start talking confidently about something that it actually has absolutely no understanding of” and then they proceed to do exactly the same f**king thing.’
https://www.reddit.com/r/singularity/comments/1td046p/twitter_user_posts_a_real_monet_and_says_its_ai/#lightbox

Machines are now better at debating than professional human debaters.
https://arxiv.org/abs/2606.16475

The “Europe 2031” report predicts the Continent, caught between the U.S.-China AI race, could slide into irrelevance.
https://europe2031.ai/summary/

A new paper from Google Deepmind, “From AGI to ASI”, says the proto-AIs we have now have no independent drives, but that won’t be true for future AIs. They’ll be subject to the same, fundamental evolutionary pressures as organic life forms, and the ones that adapt the best will turn out to be the dominant ones.
https://arxiv.org/pdf/2606.12683

The leading U.S. LLMs have at least slight liberal biases.

Elon Musk has become the world’s first trillionaire. I predicted this milestone would not happen until the 2040s!
https://apnews.com/article/musk-spacex-tesla-ipo-trillionaire-billionaire-worth-rockets-7723f82b6063a9a17c194e25982cd66d

SpaceX has bought the AI startup Cursor to reinvigorate its own AI technology.
https://www.cnbc.com/2026/06/16/spacex-spcx-cursor-acquisition-ipo.html

Oracle is laying off 21,000 workers thanks to automation.
https://www.bbc.com/news/articles/c4gy0x0j5deo

The U.S. government declassified more UFO files. As I thought, none of them contains real evidence of aliens, and none is fundamentally different from other, equally credible UFO reports that have long been in the public domain.
https://apnews.com/article/trump-ufo-file-release-third-batch-34c2a9b294e94a972f352df42c4a17ae

The woman who won a Nobel Prize for co-discovering the CRISPR gene editing technique believes AI will accelerate some areas of medical research, but is deeply skeptical of the claims from “tech bros” that machines will soon achieve breakthroughs thanks to running computer simulations of cells and human bodies. These grandiose predictions include things like curing all cancers and reversing the aging process.
https://youtu.be/n-hWHV2ZKOA?si=ejJRXttJns6_jxYL

Autism is much more common in trans people than in cis people.
https://pmc.ncbi.nlm.nih.gov/articles/PMC10313553/

‘Olympic gold medalist Julia Krajewski, who piloted [racehorse] Chilli Morning II and once competed against the original Chilli Morning, said the clones are eerily similar, and seem to share not just physical characteristics and capabilities, but also quirks.

“They’re all not big fans of sunken roads,” Krajewski said, referring to a combination of jumps on cross-country courses that test a horse’s ability to jump down, rock back and jump up again.

…In 2024, Kheiron also produced the world’s first genetically-edited horses using Crispr technology. Those horses, clones of the Polo Hall of Fame mare Polo Pureza, have an edit in a gene that regulates muscle development.’
https://www.msn.com/en-us/sports/general/more-top-horses-are-being-cloned-rattling-the-world-of-equestrian-sports/ar-AA24Zq9X

Our knowledge of which human genes code for which traits has been improving exponentially.

A round of IVF will produce many zygotes that have major genetic defects. Today, they are thrown out. In the future, when fixing them is a simple matter, things will be different.
https://www.techexplorist.com/extra-chromosome-causes-down-syndrome/103237/

A new method of genetically altering human embryos with fewer errors than CRISPR has been found: ‘Efficient base editing and development in human embryos without chromosomal alterations’
https://www.biorxiv.org/content/10.64898/2026.05.30.728989v1

This makes some strong points against mass genetic health screening:

1) What is the use in screening for deadly incurable diseases?

2) What is the use in screening for diseases where genetics are not destiny? For example, if a gene leaves you with a 20% risk of eye cancer, that means you’re 80% likely to NOT get it. Telling you that you have the gene will saddle you with a lifetime of worry and ultimately needless and expensive precautionary treatments.

3) Mass screening programs that require people to share the genetic data with public health agencies for broader genetics research raise concerns about personal privacy and state coercion.
https://www.economist.com/science-and-technology/2026/06/29/should-every-babys-dna-be-sequenced

A new pancreatic cancer drug extended patient survival times by six months.
https://www.bbc.com/news/articles/cy82l435171o

GLP-1 drugs lower the risk of breast cancer.
https://www.goodmorningamerica.com/wellness/story/doctor-breaks-study-showing-glp-1s-lower-breast-133551460

The shingles vaccine reduces dementia rates.
https://erictopol.substack.com/p/spotlight-on-the-shingles-vaccineagain

The HPV vaccine has eliminated cervical cancer deaths among young British women.
https://www.bbc.com/news/articles/c621z28z138o

Posted on

What Will Posthumans Do for Fun?

[Written with help from GPT-5]

Probably more than anything else, I want to live long enough to see the arrival of medical immortality. I don’t merely want to avoid death—I want to live forever.

As serious and analytical as I can seem, I have an insatiable curiosity and a love of life. I want to learn new things, master new skills, immerse myself in unfamiliar cultures, and experience lives radically different from my own. Unfortunately, the world I was born into leaves too little time for such pursuits. Like everyone else, I’ve had to spend much of my life studying things and working jobs I wasn’t passionate about simply to pay the bills. The knowledge that there is no law of physics prohibiting biological immortality gives me hope that a far richer existence may be possible for me.

For years, I’ve imagined what I would do with that gift: I’d start by moving to a small town in Italy, a country I visited only once but loved. I’d master the language, learn to cook authentic dishes, open a tiny restaurant, and slowly become part of the community. If, after twenty years, I grew restless, I’d move on. Maybe I’d become a train conductor in Tokyo. Later, a safari guide in Tanzania. After that, maybe a technician at a research station in Antarctica. There are thousands of cultures, professions, and ways of life I find worth experiencing. At the slow pace of my ordinary human mind, even a thousand years would not be enough to live every life I want to live. For years, my post-Singularity life plans seemed to be settled.

Then I realized something unsettling: The same technology that would allow me to pursue those dreams would also make them obsolete.

Suppose I survive long enough to purchase an immortality treatment, regain the body and mind I had at 25, and can subsist comfortably off of my savings from my lifetime of labor in our humdrum era. My Italian adventure begins exactly as I imagined.

A few years later, new enhancements appear. I install an upgrade that dramatically increases my speed of thought. Another allows direct connection to fully immersive virtual reality, like the Matrix.

Suddenly, I no longer need to move to Italy at all. I can experience the same town inside a simulation. In fact, the simulation could easily be better than reality. The architecture is perfectly preserved, but the intrusive chain restaurants disappear, the abandoned buildings at the edge of town are restored, and the inhabitants become more interesting. The ugly, slab-sided data center in the valley below vanishes. Historical events can be replayed. If my goal is to understand Italian culture and enjoy living there, the simulation wins on every metric.

Better still, I can run the simulation at ten times normal speed. I could spend what feels like twenty years there while only two years pass outside. If my goal is to experience new things and to indulge in a foreign culture, it makes logical sense to do it with as little delay as possible.

The same would apply to Tokyo, Tanzania, Antarctica, or anywhere else.

But this is only the beginning. After living the equivalent of a century across dozens of simulated lives, I would eventually notice the one constant present in every experience, and the one thing that is holding me back: me.

My personality, interests, ambitions, and sense of beauty all arose from an accident of genetics combined with the circumstances of my upbringing. Had I inherited different genes or been raised elsewhere, I might have despised travel and instead found perfect happiness watching football games at a neighborhood bar every weekend.

So why stop at changing my environment? Why not change myself?

If future technology allows me to safely and reversibly alter the architecture of my own mind, I would certainly try it. I could temporarily become someone with an entirely different personality. For example, rather than experiencing Italy as an American outsider, I could experience it through the mind of someone born there. I could live as a woman, as a mathematical genius, or as an ancient military commander. As someone whose greatest joy comes from creating art rather than studying technology. I’ve long observed that many other people are more upbeat than I am and get more pleasure from life, so I’d want to try being naturally optimistic.

After enough centuries, my original personality would become only one possibility among countless others. This is why I believe speculations about the post-Singularity future are misguided: People usually assume that while technology will change, human desires will stay the same. Machines will perform all labor, leaving us free to spend eternity enjoying today’s pleasures: eating wonderful food, traveling the world, falling in love, having sex, raising families, playing games.

But why should we assume that? Human pleasure is not objective; it is contingent.

What we find beautiful, meaningful, or enjoyable is largely the product of the particular brains evolution happened to give us. As argued in the essay Circuits – and the Arbitrariness of Beauty and Value, there is nothing inherently beautiful about trees, sunsets, or flowing water. Had we evolved from flies, perhaps our greatest works of art would celebrate rotting carcasses. Had we evolved from spiders, perhaps poetry would glorify the perfect web or the exquisite struggle of trapped prey.

Beauty is subjective.

A posthuman mind would not simply think faster or remember more facts. It would likely possess entirely new categories of sensation and experience that Homo sapiens literally cannot imagine. Just as a bat perceives aspects of reality unavailable to us through echolocation, or a mantis shrimp experiences colors beyond human vision, posthumans may inhabit experiential worlds inaccessible to ordinary human consciousness.

Science fiction occasionally hints at this possibility. In the Star Trek: Deep Space Nine episode “Favor the Bold,” the shapeshifter Odo and the Female Changeling have sex while both are in humanoid form. Afterwards, the Female Changeling dismisses the encounter as only “a shadow” of the experience of linking—when Changelings merge together in their natural liquid state, sharing thoughts, emotions, and identities directly. Human sex, perhaps the most intimate experience our species knows, is revealed to be only a pale imitation of something categorically richer.

If posthuman minds discover analogous forms of experience, then our greatest pleasures may eventually appear as primitive as finger painting appears to an accomplished artist.

Even humans who refuse to become posthuman may not escape this conclusion. Many people imagine that “unaltered” humans would continue enjoying ordinary pleasures while only enhanced humans drift into strange new forms of existence. But advanced technology may offer ordinary humans something even more seductive.

In the 1950s, neuroscientists James Olds and Peter Milner implanted electrodes into the brains of laboratory rats, allowing the animals to stimulate parts of their own reward circuitry by pressing a lever. The rats repeatedly activated the stimulation, neglecting food, water, and rest in favor of obtaining another pulse of pleasure. The experiments demonstrated that directly activating the brain’s reward systems could overwhelm many other motivations.

Future technologies—whether brain-computer interfaces, cybernetics, or medical nanotechnology—may eventually stimulate human reward circuits directly while preventing habituation. If so, eating, travel, relationships, and every other natural pleasure might become hopelessly inefficient compared to continuous engineered bliss. The result could be a civilization whose members voluntarily spend centuries immersed in uninterrupted ecstasy, scarcely interacting with the outside world at all.

So what will posthumans actually do for fun?

I suspect there are only three broad possibilities:

The first is the Pleasure Machine. Through countless small modifications, people gradually strip away everything except raw pleasure itself. Every intermediate activity—food, games, travel, conversation—is eventually recognized as merely an inefficient path toward the neurological state they actually desired all along. Pleasure becomes the destination rather than the reward for reaching it. You’re the lab rat eternally pushing the button so you can feel the best orgasm or first hit of heroin. Your shallow but incredibly blissful existence renders you a defenseless parasite that the AGIs and posthumans who still work for a living will be tempted to kill to free up resources.

The second is the Incomprehensible Life. Posthumans discover forms of experience that are deeper, richer, and more meaningful than anything available to Homo sapiens. Just as a Changeling’s “Great Link” surpasses human intimacy, posthuman existence may consist of experiences we cannot even describe because our brains lack the conceptual vocabulary to represent them. The closest analog we might have is a math genius experiencing a moment of insight in how to complete an important theorem.

The third possibility is the Ascetic Intelligence. After centuries or millennia of experimentation, some individuals may conclude that pleasure itself is merely an evolutionary tool rather than life’s highest purpose. They may deliberately reconfigure themselves to desire understanding, creativity, or civilization-building instead of happiness. With sufficiently advanced control over one’s own mind, even anhedonia could become not a disorder but a chosen mode of existence. Such beings might resemble Buddhist sages crossed with superintelligent scientists—nearly devoid of ordinary human wants and in fact shorn of most of what it means to be human, yet capable of extraordinary insight. Since these individuals would still be thinking, working assets, they would be less vulnerable to marginalization and destruction than the previous two kinds.

The irony is hard to escape: The technology that grants my dreams will also transform the dreamer.

When people ask what posthumans will do for fun, they assume the beings inheriting our civilization will still share our concept of fun. I doubt they will.

Links:

  1. A discussion of the Star Trek Deep Space Nine episode where two aliens capable of shapeshifting into anything decided to turn into humans so they could have sex. They didn’t like it.
    https://www.startrek.com/news/when-sexuality-is-literally-fluid
  2. Dan Faggella’s essay “Circuits – and the Arbitrariness of Beauty and Value” https://danfaggella.com/circuits
Posted on

Interesting articles, May 2026

In spite of the ceasefire, Iran and U.S. sporadically traded fire.
https://www.cbsnews.com/news/u-s-destroyers-face-second-round-of-iranian-attacks/
https://apnews.com/article/iran-us-israel-war-oil-hormuz-may-14-2026-efb53c39ee6334733e1cb22ca4a6c279

Israel has put nets around their jeeps to protect against terrorist drones. No one has a good solution to the drone threat.
https://www.twz.com/news-features/israel-now-using-netting-to-protect-combat-vehicles-against-scourge-of-hezbollah-drones

‘UAE Building Massive ‘Cope Cages’ To Protect Energy Facilities From Iranian Drone Attacks’
https://www.twz.com/news-features/uae-building-massive-cope-cages-to-protect-energy-facilities-from-iranian-drone-attacks

‘Amazon stuck with months of repairs after [Iranian] drone strikes on data centers’
https://arstechnica.com/gadgets/2026/05/amazon-stuck-with-months-of-repairs-after-drone-strikes-on-data-centers/

‘It is not hard to imagine how the satellite constellation being described here would fundamentally change the U.S. military’s ability to not just spot and track targets globally, but also close the kill chains to engage them, even at very long ranges. This has massive implications for future net-centric warfare where all sorts of tangential capabilities will increasingly be networked together. It might impact how tactical aircraft are equipped in the future, including the need for their own radars. There could at least be a reduced need for them to use their own radars to guide missiles, even when no supporting sensor network within the Earth’s atmosphere has relevant data to provide.’
https://www.twz.com/space/pentagons-plans-to-track-aircraft-from-orbit-accelerated-with-new-4b-spacex-deal

Israel set up secret air bases in the desert of Iraq. ‘Strikingly, the base was not uncovered by Iraqi security agencies but by a shepherd in the area who noticed unusual military activity, including helicopter movements and gunfire, and reported it, prompting Iraqi forces to launch an urgent investigation. Israel, however, moved quickly and carried out intensive strikes that left one Iraqi soldier dead.’
https://www.yahoo.com/news/articles/report-shepherd-uncovers-alleged-secret-140714766.html

For the first time in years, Russia actually lost territory overall in Ukraine.
https://www.cnn.com/2026/05/14/europe/russia-winning-streak-ukraine-over-intl-cmd

‘Victory will be ours,’ Putin tells Victory Day parade without any tanks
https://kyivindependent.com/victory-will-be-ours-putin-tells-victory-day-parade-without-any-tanks/

Russia is slapping extra armor onto its old T-72A tanks and sending them to Ukraine.
https://youtu.be/zC7b04EWrH0?si=NMyssrDa4GYF1GgV

Trump’s controversial new White House ballroom will also have a bunker under it and advanced defensive weapons built into its upper floor.
https://www.twz.com/news-features/the-white-house-ballroom-is-a-deep-fortress-in-disguise

The world’s best ethical hackers agree that machines are now within the human range of competency in their field. However, they disagree over whether machines will fully replace human hackers or remain powerful tools.
https://www.bbc.com/news/articles/c3r2zjpryzro

‘Lindy’s Law has a heavy tail, which means we can’t simply halve these to find our 25th percentile estimate. Our 25th percentile estimate for the next advance as exciting as LLMs should be three years from now; for deep learning, it’s five years.

So even if you think AGI will require a further paradigm shift as big as the invention of the LLM or as deep learning itself, you should have 25% chance it will be developed in the next 3 – 5 years. Which is about as long as the LLM-only crowd think things will take! This isn’t an excuse for relegating the risk of AGI to some vague indefinite future. It could still be the late 2020s or early 2030s!’
https://www.astralcodexten.com/p/new-paradigms-wont-save-you

Marc Andreeseen believes the first AGI was created about three months ago, but we just don’t realize it yet.
https://youtu.be/PHQvb10vKyk?si=b5h0X8TaRrQXmZaG&t=6237

Demis Hassabis believes AGI will be created in 3-4 years, and that “the Singularity” will happen.
https://sherwood.news/tech/google-deepminds-hassabis-agi-is-3-to-4-years-away/

Richard Dawkins wrote this. Extending his logic, why do we assume that current humans represent the endpoint of consciousness? If our primate ancestors were half conscious, then could our technologically augmented descendants be 1-1/2 conscious? Could AGIs someday be “twice as conscious” as we are?

Many were uncharitable towards Dawkins.
https://garymarcus.substack.com/p/richard-dawkins-and-the-claude-delusion

Stephen Wolfram also believes that LLMs like Claude might be conscious and suggest that consciousness is not uniquely human or even dependent on intelligence level.
https://youtube.com/shorts/AlhdBHeDH9M?si=7A6kJH2XH2Iw-27Z

A science paper confirms that computers have passed the Turing Test with no caveats. The machine actually blew the Test out of the water.

‘When prompted to adopt a humanlike persona, GPT-4.5 was judged to be the human 73% of the time: significantly more often than interrogators selected the real human participant. LLaMa-3.1, with the same prompt, was judged to be the human 56% of the time—not significantly more or less often than the humans it was being compared to. Without these prompts, however, the same models performed significantly worse (38% and 36%), and did not consistently outperform baseline models, ELIZA and GPT-4o (23% and 21%, respectively).’
https://www.pnas.org/doi/10.1073/pnas.2524472123

The world of mathematics was rocked by GPT-5’s solving of an important Erdös problem, the “Unit Distance Problem”. Several lesser Erdös problems were also solved in short order.
‘After reading it, a former OpenAI researcher did some back-of-the-envelope math and estimated it took less than 32 hours and $1,000 in tokens, a bargain for a result of this caliber.’
https://www.wsj.com/tech/ai/ai-math-solves-erdos-problem-openai-c4029e84
https://scottaaronson.blog/?p=9782

At a trial where Elon Musk sued Sam Altman, witnesses consistently said Altman was a deceptive and dishonest person. Musk lost on a technicality, but emerged as the winner in the court of public opinion.
https://www.theguardian.com/us-news/2026/may/11/musk-v-openai-altman-trial
https://www.nytimes.com/2026/05/18/technology/elon-musk-lawsuit-openai-sam-altman.html

Someone figured out how to hack computers that are air-gapped and inside of Faraday cages.
https://arxiv.org/abs/1802.02700

‘Value in software is moving away from the interface and toward the data underneath it. An app frontend is increasingly a liability: opinionated in ways users didn’t choose, maintained on someone else’s timeline, and built for an average use case rather than yours. The founders who don’t see this coming will spend years building an interface that their customers will eventually replace themselves.

What fills the gap is something I think of as the “meta-app”: apps that build other apps on the fly, perfectly tailored to your immediate need. You tell it what you want, and it figures out the rest. AI coding tools like Claude Code and Codex are an early version of this. They’re getting a lot of attention for helping developers build software faster, but their real power is bigger than that. It’s the ability to go directly from intent to outcome for any task, without an app in the middle.’
https://fortune.com/2026/05/13/google-sheets-engineer-apps-ending-meta-app-ai-zach-lloyd-warp/

‘Figure’s event began on May 13 as a planned eight-hour robot demonstration featuring the company’s latest Figure 03 robots. The chosen robotic task involved inspecting the bar codes on various small packages—including cardboard boxes and soft padded envelopes or bags—and then placing the packages on a conveyor belt with the bar codes facing downward. The demo would feature the robots performing the task autonomously without any human intervention, according to Figure CEO Brett Adcock.

But Adcock initially played down expectations by noting that the Figure team was aiming for the robots to work for eight hours straight, whereas a previous Figure demo had lasted just one hour. “High odds something breaks,” Adcock posted on X.’

They just ended the demo after the robot worked for 200 hours straight without making a mistake.
https://youtube.com/shorts/P7j65otmBeY?si=W1ea3cjMYYcM6g68

“FM-2030” was an archtranshumanist if there ever was one.
https://en.wikipedia.org/wiki/FM-2030

Screenshot

Thanks to government subsidies and other policies, China’s solar industry surged over the last few years. Most of their output was consumed within China itself, as solar farms were built all over the place. The country’s electrical grid is now glutted with solar power, which is causing problems thanks to its inherent intermittency. Domestic demand will sharply shrink in the near future, leading to a measurable decline in solar power worldwide that critics will surely seize upon as proof the rise of solar power has ended.
https://www.economist.com/china/2026/05/26/chinas-world-beating-solar-industry-is-in-turmoil

The notion that China doesn’t innovate and merely copies Western technology is outmoded.
‘Third, we relate our findings to economic and geopolitical development, arguing that China understands modular natives and scale-up better than any other geography and that this is key to China’s swiftly growing dominance in renewables, batteries, EVs, robots, etc. We argue that China’s mastery of modularity and scale-up is a major innovation in its own right, among the greatest and most impactful in human history, falsifying the common notion that China cannot innovate. Business and government outside China ignore these findings at their peril.’
https://papers.ssrn.com/sol3/papers.cfm

A New Glenn rocket blew up during a test, destroying its launchpad in what might be the biggest non-nuclear explosion in history.

‘The destruction at Cape Canaveral — and the resulting damage to the company’s only launchpad — will likely require months of extensive repairs. Blue Origin was set to launch a key moon mission on its New Glenn rocket this year that will now almost surely be delayed. NASA is trying to build a moon base on Trump’s expedited timeline and beat China to the lunar surface, and this setback could undermine that effort.’
https://www.politico.com/news/2026/05/29/nasa-blue-origin-trump-moon-base-delay-00943139

Accomplished pilot and astronaut Brian Binnie saw UFOs in his backyard one night.
https://www.dailymail.com/news/article-15762377/astronaut-brian-binnie-ufo-encounter-california-bedroom.html

The U.S. government started releasing large numbers of hitherto classified reports about UFO sightings. Most of them are probably explainable, and none of the unexplainable cases are so compelling that they “prove” the existence of alien space ships. Here’s one example.
https://youtu.be/Xn9G1TEEqFE?si=AKNjAdaQlUytJoj1

‘General intelligence explains the link between math and music skills’
https://www.psypost.org/general-intelligence-explains-the-link-between-math-and-music-skills/

Here’s a very fascinating interview with David Reich about what human population genetics tells us about prehistory.
https://youtu.be/Uj6skZIxPuI?si=zXXNSGNni_4uHoJn

During the Bronze Age, when the Great Pyramids were being built in Egypt, Europe and northern Asia were primitive backwaters inhabited by horsemen, hunter-gatherers, and small farming communities. The Indo-Europeans, who resembled today’s white people more than any group at the time, originated in modern Ukraine and spread out in every direction. Some of them–called the “Afanasievo” people–actually reached western China and Mongolia and settled there. Over hundreds of subsequent years, they were absorbed or killed off by East Asians.
https://youtu.be/3dVrVJ_CY4g?si=GsIS1qwySakJPnnn

Conversely, the Indo-Europeans expanded into it and dominated it for thousands of years. Starting around 1000 BC, East Asians from Mongolia started migrating there, and never really stopped. Today, most Kazakhs look partly or fully East Asian. Kazakhstan is within Europe’s boundaries.
https://www.atlasofhumanity.com/kazakh

Genetic studies reveal the identities of some of America’s earliest European colonists who settled in Maryland.
https://blog.23andme.com/articles/ancient-dna-and-the-story-of-some-of-americas-earliest-colonists

‘This means potentially millions of men worldwide have sperm counts so low that their individual spermatozoa are so hard to find that they are considered to be azoospermic. But the power of AI to find these hidden sperm could offer hope to those hoping to become parents.

At the end of last year, after five years in development, the first baby to be born using the Star system allowed a couple who had battled with infertility for almost two decades to finally have a child. It’s a moment Zev Williams, director of Columbia University Fertility Center, and his team remember well.’
https://www.bbc.com/future/article/20260429-finding-hidden-sperm-new-technology-offers-hope-to-men-previously-told-they-were-infertile

The newest GLP-1 drug, “retatrutide”, caused people to lose an average of 28% of their body weight, making it the most powerful weight loss drug yet created.
https://www.scientificamerican.com/article/retatrutide-results-spark-questions-about-how-rapid-weight-loss-affects-the-body/

‘No child deaths definitively linked to Covid shots, FDA says’
https://www.nbcnews.com/health/health-news/no-child-deaths-definitively-linked-covid-shots-fda-says-rcna346514

‘Just a day ago, the brain was in a living person. Now, hours after its owner died, it sits on a cart draped in tubes that quiver as they pump liters of blood substitute and other fluids through the organ, supplying oxygen and removing waste. With most of its key functions intact but its electrical activity quenched by anesthesia, the brain hovers between life and death. As it metabolizes experimental drugs, sensors record its reactions, capturing hundreds of data points on its cells, proteins, and physiology. Then, after 24 hours in this state, it will be sliced into hundreds of pieces for more detailed study.’
https://www.science.org/content/article/not-alive-not-dead-disembodied-human-brains-used-drug-testing

The first “Enhanced Games”–a sports competition without restrictions on performance-enhancing drug use–was held.

‘In the end, across 20 different events, only one world record was set: Kristian Gkolomeev of Greece finished the men’s 50-meter freestyle swim in 20.81 seconds, just barely beating the World Aquatics-recognized world record by .07 seconds.

…Instead, Enhanced had to watch nonenhanced athletes take victory in three different events.

That reminded me of a point Enhanced CEO Max Martin made when I interviewed him back in January: You can take all the performance enhancements you want, and they might make you a better athlete than your nonenhanced self, but you still need athletic skill and genetic luck to be the best in the world.’
https://reason.com/2026/05/26/just-1-world-record-at-the-enhanced-games-shows-the-integrity-of-the-competition/

Posted on

Human cryonics and the problem of “strategic death”

[Written with the help of GPT-5]

I can’t stop marveling at how smart, capable and creative ChatGPT has become. Regardless of the weirdness of any idea I put forth to it, the machine will dutifully respond will complete seriousness and scholarliness. For people with curious minds, it’s like a drug.

The most interesting recent conversation I had with ChatGPT was about human cryonics: the practice of freezing people right after they die in the hopes doctors will be able to fix them and safely thaw them out in the future. To be clear, the vast majority or perhaps all of the people who have been cryopreserved so far are non-recoverable because they spent too much time decaying before they were frozen or because they were frozen using bad techniques. No level of future technology will ever bring them back.

However, the freezing techniques have been slowly improving, and it’s possible we could find ways to cryopreserve humans without turning their cells into mush, so we shouldn’t write the field off. I took a step farther than that, assumed a way had been found to safely revive people from cryopreservation with their brains USUALLY intact, and started a dialog with ChatGPT. Specifically, I wanted to know about the legal complications arising from revival.

Establishing who the revived person is

The first challenge would be establishing their identity. Future legal systems would likely require multiple independent forms of evidence before recognizing a revived person as the same legal individual who entered cryonic suspension.

The most obvious form of evidence would be biological continuity. DNA samples collected before preservation could be compared with the revived individual. If none existed, then the revived person’s DNA could be compared with any descendants just as paternity tests today prove parentage. Fingerprints, retinal scans, and other biometric records could also be examined.

Yet DNA and biometrics alone would be insufficient. Future medicine will be capable of extensive tissue regeneration, organ replacement, and cellular repair, and those techniques will surely be used on frozen people as part of the revival process, first to fix the freeze-related damage and second to fix whatever injury or illness killed them. As a result person revived after decades or centuries in storage may have a body that differs substantially from the one originally preserved. Fingerprints, scars, and other identifying features may no longer match historical records. Significant reconstruction of the brain would pose special problems for identity verification.

For this reason, biological evidence would be combined with neurological evidence to establish the revived person’s identity. Modern medicine already uses brain imaging technologies to study the structure of the living brain. A future cryonics patient might enter storage with detailed scans of neural architecture, memory centers, and other neurological features. After revival, those scans could be compared with the restored brain. The closer the match, the stronger the evidence for continuity.

Psychological evidence would also play an important role, and, importantly, requires a lower level of technology than high-res brain scans. A revived person could be asked to recall information that was documented before preservation but never publicly released. Personal memories, private experiences, forgotten family details, and other forms of knowledge could be compared against historical records, and perhaps against testimonials the person made in the lead-up to going into cryostasis. A person who accurately recalls thousands of obscure details from a life lived a century earlier would present powerful evidence that continuity of identity has been preserved.

An equally important piece of evidence would be chain of custody. Courts already rely heavily on chain-of-custody procedures when dealing with criminal evidence, human remains, organ transplants, and other sensitive materials. The same principle would be applied to cryonics.

Imagine that a cryonics company maintains continuous records from the moment a patient enters preservation. Every inspection, transfer to a different cooler, maintenance operation, and storage event is documented. The preserved body remains under controlled supervision for a hundred years. When revival finally occurs, investigators can prove exactly where the patient has been and who has been responsible for their care throughout the entire period.

As farfetched as it sounds, a possible future scam could involve cloning a dead person who was in cryostasis, somehow implanting the person’s memories and personality into the clone, and then having the clone claim to be the revived original person. Chain-of-custody requirements would foil such practices.

A future court would probably not rely on any single category of evidence. Instead, it would combine them: DNA would establish biological origin, brain scans and memory/personality tests would establish neurological and psychological continuity, and chain-of-custody records would establish historical continuity. In light of this, it would be very helpful if cryonics companies today required their clients to undergo personality tests, give detailed autobiographical statements, and undergo medical scans that captured their biometrics so those data could be used to help their clients verify their identities in the future after revival. Steps would also be needed to ensure the data’s confidentiality and integrity over the decades or even centuries.

Untangling the revived person’s “past life“

After the person is legally confirmed to be the revived dead person, they’ll face a slew of new hurdles that current law is not configured to deal with. Modern law is built on the assumption that death is permanent. Almost every legal consequence that follows from death hinges on the belief that the deceased will never return.

In the U.S., the declaration that a person is dead triggers a cascade of consequences, including estate settlement, termination of legal relationships, and removal from countless government systems. There is currently no established legal procedure for reversing death after, say, a century has passed. The closest modern parallel might be a person who was mistakenly declared dead and later reappears, but even that comparison falls short.

Every practical component of a revived person’s legal identity would have been dormant for many years. Their Social Security records would indicate that they are deceased, their government files would be archived, and many other records would have been transferred between agencies or lost over time. The individual would likely need to petition a court for formal recognition and reconstruct key historical records.

Property presents one of the harshest realities. After death, a person’s assets are distributed according to wills, trusts, probate proceedings, and inheritance laws. By the time a person is revived many years after dying, their former possessions would almost certainly be gone: Their house would have changed owners many times. Their savings would have been spent by descendants generations earlier.

Courts are extraordinarily reluctant to unwind settled property rights, particularly when innocent third parties have relied upon them for decades. As a result, a revived individual would likely discover that they possess no legal claim to any of the wealth they owned before death.

Family relationships would be equally complicated. Marriage legally ends at death, and a surviving spouse becomes a widow or widower and is free to remarry. Those subsequent marriages are fully valid and could not be undone because a dead former spouse reappeared. The revived individual would therefore not automatically regain any marital rights.

Their relationship with descendants would also be unusual: They would remain the biological ancestor of children, grandchildren, and great-grandchildren, yet many of those descendants might be older than they are in biological terms. The law has never had to accommodate family trees in which ancestors routinely outlive their descendants by centuries.

Citizenship and civil status would likely be easier to restore, but they would still require significant administrative effort. Citizenship itself may survive death in a technical sense, but practical recognition of that citizenship does not. A revived person would probably need to obtain new identification documents, reestablish their eligibility for government programs, and restore their presence within systems that had long classified them as deceased. Reapplying for a passport or reactivating a Social Security record might seem minor compared to other legal challenges, but these processes would still require legal recognition that the revived person is who they claim to be.

Contracts, debts, and obligations introduce another layer of complexity. Most contractual relationships terminate upon death. Employment agreements end. Personal obligations are extinguished or settled through the estate. Debts are paid from available assets, discharged, or written off. As a result, a revived person would likely find that old obligations no longer exist. Yet the same principle would apply to rights and benefits. They would not regain employment contracts, pensions, memberships, or other legal relationships that ended upon death. The legal slate would, in many respects, have been wiped clean.

Criminal law raises particularly unusual questions. If a person had committed a crime before death, criminal proceedings would generally have ended once they were declared deceased. A revival many years later would force courts to confront issues for which there is no modern precedent. For most offenses, statutes of limitation would likely prevent prosecution. However, serious crimes like murder that carry no statute of limitations could create difficult legal disputes. Would a prosecution that ended because the defendant died remain permanently closed, or could it be revived along with the defendant? Existing law offers no clear answer.

Even determining the person’s age would become legally ambiguous–If someone died and was cryopreserved at age 80, then thawed out 100 years later, they might be biologically 80 but chronologically 180. Which number should matter? (This sidesteps another complicating issue, biological rejuvenation therapy, which the person probably underwent during the revival process, effectively de-aging their bodies by decades.) Retirement systems, age-based benefits, and countless legal categories assume that biological age and chronological age move together. Cryonic revival would break that assumption. Legislatures and courts would have to decide whether age should be measured by years since birth, years consciously experienced, or some combination of the two. A dual system where different ages are used depending on context is likely.

The most practical future legal solution would be a form of partial continuity. Courts would recognize that the revived individual is indeed the same person who died a century earlier, while simultaneously treating their legal existence as beginning anew at the moment of revival. Revival would be a distinct life event giving rise to new legal personhood, and your previous legal life would be considered “closed.” In effect, the individual would be recognized as the same human being but would need to rebuild much of their legal and economic life from scratch. They would have to establish identity, obtain documentation, and create new financial relationships while accepting that many aspects of their former legal existence had permanently ended.

Dealing with “strategic death”

Another challenge posed by cryonic revival is the possibility that people might deliberately use death for strategic benefit. Modern legal systems assume that death is permanent. As a result, many legal rules treat death as a final endpoint: estates are settled, debts are discharged, insurance benefits are paid, contracts terminate, and criminal proceedings often end. If people can die and later return, those rules create opportunities for exploitation that never existed before.

A future legal system would therefore need to be highly resistant to what might be called “time arbitrage” or “strategic death.” In finance, arbitrage refers to exploiting differences between markets to earn a profit. In the context of cryonic revival, it occurs when someone exploits differences between the legal rules governing the living and the dead in order to gain an advantage that would not be available during a continuous lifetime. They make arrangements to be frozen and revived at some future date and deliberately kill themselves (certainly in a way that doesn’t damage their brain) or stage their death as part of a plan to use “temporary death” as a beneficial legal loophole. Examples are as follows:

  • A heavily indebted person uses a temporary death strategy to allow debts to be settled or discharged through their estate, and then return years later free of financial obligations.
  • Someone buys a large life insurance policy, opts for temporary death, triggers a payout to trusted heirs, and later returns to life to share the wealth generated by their own death.
  • A person who has become rich through asset growth (e.g. – stock or real estate portfolio that massively appreciate over their lifetime) opts for temporary death to take advantage of the U.S. tax code’s “step-up” provision, which waives all capital gains taxes on those assets. After the owner’s death, the assets are transferred to trusted heirs, who liquidate them. The frozen person is then revived, and the untaxed wealth is shared with them.
  • A person facing serious criminal charges or large lawsuits opts for temporary death to escape liability.

The legal system would probably respond to the practice of strategic death with measures like mandatory taxes on revival, limits on pre-death financial structuring, and audits of “revival planning”, especially in cases where the person may be attempting to exploit a loophole. The underlying principle would be simple: no individual should materially gain merely because they passed through legal death and returned.

And while this isn’t a true legal loophole, the practice of using temporary death as a long-term investment strategy could cause enough headaches to warrant legal intervention: A person would place assets into a trust designed to survive for centuries, enter cryonic suspension, and be revived at a prespecified date or under prespecified personal wealth conditions to claim wealth that has massively compounded over time. From their perspective, they would become “instantly rich.” If unchecked, such arrangements could produce vast concentrations of wealth and effectively create immortal financial dynasties.

A future legal system would therefore likely permit long-term trusts while imposing restrictions like duration limits, taxation of dormant assets, and anti-monopoly safeguards designed to prevent extreme accumulation of wealth across centuries. More generally speaking, pre-death financial arrangements would be required to receive heightened scrutiny, particularly if they appeared designed to exploit differences between the legal treatment of the living and the dead.

Dead, alive, and something else

Eventually, the legal system might adapt further by classifying “dead” people differently depending on whether their deceased state were reversible. Biological death and legal death are the same concept today, but in a world with reliable revival techniques, they would diverge. A dead person who was cryopreserved would not be considered permanently dead in the legal sense. Instead, they might occupy a third category between life and death: suspended existence.

Such individuals would remain legally continuous even while inactive. Their obligations could be paused rather than extinguished. Their legal identity would persist. Their eventual return would be anticipated rather than treated as a miracle.

This shift becomes easier to understand when we consider repeated revivals. Suppose a woman is revived after fifty years in cryonic suspension. The event would attract enormous attention. Newspapers would call her a person returned from the dead. Courts would struggle to determine her legal status.

Now suppose the same woman undergoes the process five times over the course of four centuries. At that point, describing her as having “returned from the dead” begins to sound misleading. She is not a resurrected individual, she is simply a very long-lived person who spends portions of her existence inactive.

Eventually society might stop viewing cryonic suspension as death altogether–it would instead be treated as an unusual form of absence. This transformation would fundamentally alter the legal meaning of death.

Today, death is the event that triggers inheritance, dissolves marriages, terminates contracts, and closes a person’s legal existence. In a world where revival is common, those consequences could no longer be tied to biological death alone. They would have to be tied to a determination that the individual is genuinely beyond recovery.

The law would therefore reserve its most serious consequences for a new category: permanent death. On the other hand, a cryopreserved person whose brain remains recoverable, and whose revival remains plausible, would be deemed “suspended.”

The introduction of that third legal category would mitigate temporary death as a means of committing fraud. If an individual deliberately arranged a false death in order to escape criminal prosecution, for example, the law would refuse to recognize any interruption in legal continuity. The person would be treated as having remained legally present the entire time, and prosecution would resume upon their revival. The other loopholes described earlier would likewise be closed.

Conclusion

In the end, the most important legal question raised by cryonics is not whether future technology can revive the dead, but whether our legal institutions can adapt to a world in which death is no longer permanent. Once revival becomes possible, every assumption built into modern law begins to wobble. Courts would need to determine who a revived person is, what rights they retain, what obligations survive, and how to prevent death itself from becoming a tool for fraud, tax avoidance, or other forms of strategic manipulation. Over time, the law would likely be forced to abandon its traditional view of death as a single, irreversible event. Instead, society might come to recognize three distinct states: alive, suspended, and permanently dead. If that transformation occurs, cryonics will not merely have extended human life. It will have compelled civilization to redefine one of its oldest and most fundamental concepts—the boundary between life and death itself.

Posted on

Interesting articles, April 2026

Israel and America traded more blows with Iran

The sides agreed to a cease-fire, but a permanent peace agreement remains elusive.

An American F-15 was shot down over Iran. After a daring, complex and expensive operation, both pilots were safely recovered.
https://apnews.com/article/iran-us-israel-trump-lebanon-april-5-2026-pilot-cf4a792196259d6e9c066d0be1c57962
https://www.yahoo.com/news/articles/u-military-set-improvised-airfield-181145790.html
https://www.wsj.com/livecoverage/iran-war-latest-news-updates-2026/card/second-mission-to-rescue-u-s-airman-took-155-aircraft-fAUuZAYswIDtWyz1E0hJ

A U.S. A-10 attack jet crashed during the rescue operation.
https://www.ndtv.com/world-news/us-a-10-warthog-fighter-jet-crashes-near-strait-of-hormuz-pilot-rescued-report-11308682

1) Though we have badly damaged Iran’s military while suffering very low losses, the war has been very expensive for us: an alarmingly high fraction of our missiles have been expended. It will take years and many billions of dollars to replace them.

2) Our Gulf allies failed to take basic measures to protect their expensive radar and missile launchers from Iranian attacks.

3) We are short on hardened aircraft shelters at our airbases. They would be cheap to build and would have a large ROI by averting aircraft losses and forcing enemies to use more expensive missiles to strike our planes on the ground.
https://youtu.be/ApIb-nTdoLU?si=H1LXXtLfLrcFGBHQ

If the Persian Gulf oil tanker blockade continues, economies in Europe and especially East Asia will suffer terribly.
https://www.economist.com/finance-and-economics/2026/04/21/global-energy-markets-are-on-the-verge-of-a-disaster

The Ukraine War has attracted considerable numbers of foreign soldiers, including some who claim they were tricked into fighting for Russia.
https://apnews.com/article/cameroon-soldiers-killed-russia-a2a3130018f378b73cc6167918b34b5d

In spite of terrible losses, Russia will probably never run out of tanks in Ukraine. Partly this is because they’re sending fewer into combat given how vulnerable they’ve proven to be.
https://youtu.be/519XMTijfCI?si=ERwfwRkCsNz1neHs

An awesome and historically accurate CGI video of an Argentine fighter plane attack on a British destroyer during the 1982 Falklands War.
https://youtu.be/RNjKfVH6XDI?si=sevEPGhCA3GZ6-av

German “diplomacy” during WWI was terrible.
https://youtu.be/41kYH51XYGw?si=nmyZlog_wI-Lpotc

‘For the first time in a decade, there are no U.S. bases in Syria.’
https://reason.com/2026/04/17/the-u-s-military-has-finally-left-syria/

‘Japan approves scrapping a ban on lethal weapons exports’
https://www.npr.org/2026/04/21/g-s1-118178/japan-ban-lethal-weapons-exports

A company has built a 10-shot crossbow that takes AR-15 accessories.
https://youtu.be/9UxfGf9WQO0?si=8QebSUSPFSp-lMPz

At a half-marathon in China, a robot beat the human record by seven minutes.
https://www.wired.com/story/a-humanoid-robot-set-a-half-marathon-record-in-china/

The AI singer “Eddie Dalton” has several songs on the iTunes Top 100.
https://www.showbiz411.com/2026/04/05/itunes-takeover-by-fake-ai-singer-eddie-dalton-now-occupies-eleven-spots-on-chart-despite-not-being-human-or-real-exclusive

Here’s a great essay about the overlooked benefits of “AI” so far.
https://www.clear-eyed.ai/p/ai-compared-to-what

Demand for AI services has grown so explosively that the tech companies can’t keep up with enough data centers or electricity.
https://www.wsj.com/tech/ai/ai-is-using-so-much-energy-that-computing-firepower-is-running-out-156e5c85

The growth of the data center sector is comparable to rapid buildouts of earlier kinds of infrastructure.

Bill Gates recommends the book The Coming Wave for anyone who wants to glimpse AI’s future impact on the world.
https://www.gatesnotes.com/the-coming-wave

‘According to Anthropic, the capabilities of Mythos are “substantially beyond those of any model we have previously trained”. But the lab says it is particularly alarmed by the system’s ability to find software vulnerabilities and either fix them (if set to work as a defender) or exploit them (if acting as a hacker).

Such claims ought normally to be taken with a pinch of salt. Anthropic built the model, ran the tests—and stands to benefit from the perception that its system is far more brilliant than anything to have come before it. The lab has been on a roll lately. On April 6th it announced that its annualised revenue had reached $30bn, up from just $9bn at the end of last year. It is surely eager to maintain its momentum.

Yet there are reasons to take Anthropic’s latest warnings seriously. The first is their gravity: Anthropic says that Mythos has already found severe vulnerabilities in “every major operating system and web browser”, including one that had gone undetected for 27 years.

The second is the response of other companies. Alongside the pause, Anthropic announced Project Glasswing, an effort to help companies use Mythos to boost cyber-defences before it is widely released. The participation of leading software developers—including Apple, the Linux Foundation and CrowdStrike, as well as Google, which competes directly with Anthropic in AI—suggests the threat is credible.’
https://www.economist.com/business/2026/04/08/how-dangerous-is-mythos-anthropics-new-ai-model

Here’s an excellent and long analysis of where AI technology currently stands and how it will drive geopolitics and the economy. One tidbit:

‘For the last 10 years or so, our geopolitical landscape has been formed more and more by the US-China power competition. Building frontier AI also happens to be so expensive and talent-intensive that only two countries can build it at the frontier.

The boundary conditions for frontier AI are: deep talent pools, massive compute infrastructure, close-to-infinite energy, and capital markets that can finance it. Only the US and China have the economic muscle to spend hundreds of billions of dollars on the computational and energy infrastructure that AI requires. Even if a third player wrote that check, they would still be years behind, because the US and China have been compounding since the beginning.

The result is a simple structure: the two superpowers at the center, each building its own intelligence stack. The US overwhelmingly closed-source, China overwhelmingly open-weight. Around them, a small number of countries that hold critical inputs: lithography in the Netherlands, critical minerals in Australia, energy and capital in the Gulf Countries, fabrication in Taiwan. Two planets and a handful of moons. The moons have leverage because they supply something that cannot be routed around.

Everyone else faces a simpler, less comfortable reality. If you do not control a critical input, you do not have a term sheet. You have a menu. And the menu has two items: rent the intelligence from China, which will come with a certain set of conditions, or rent the intelligence from the United States, which comes with a different set of conditions. Every country will have to go through a sorting function of choosing what they want.’
https://chamath.substack.com/p/2025-annual-letter

‘If recent trends in AI capability growth persist, this pace of AI improvement implies that LLMs will be able to complete most text-related tasks with success rates of, on average, 80%–95% by 2029 at a minimally sufficient quality level. Achieving near-perfect success rates at this quality level or comparable success rates at superior quality would require several additional years. These AI capability improvements would impact the economy and labor market as organizations adopt AI, which could have a substantially longer timeline.’
https://futuretech.mit.edu/publication/crashing-waves-vs-rising-tides-preliminary-findings-on-ai-automation-from-thousands-of-worker-evaluations-of-labor-market-tasks

‘Given the inherent unknowability of this era, what would some of the signs be that we are in it? They might look like this: in 5 years,1) There are high-profile disagreements among leading AI researchers on whether AGI is here. 2) Reputable economists can’t determine if productivity has increased or decreased. 3) Lower public confidence in media platforms and established institutions. 4) The US and China cannot decide whether they are allies nor adversaries. 5) There are ambiguous spikes in employment rates in both directions. 6) Medical levels of anxiety increase. 7) Major court decisions leave as many questions as answers. 8) Commitments (marriage, work) are postponed even later in life. 9) Investing, capital allocation becomes more expensive. 10) Nihilism gets respect.’
https://kevinkelly.substack.com/p/our-uncertain-uncertainties

The Jacquard loom, introduced in France circa 1805, used a chain of punched cards to control which threads were raised for each pass of the shuttle. The ability to change the pattern of the loom’s weave by simply changing cards was an important conceptual precursor to computer programming. Babbage borrowed the idea directly for the Analytical Engine in the 1830s.

The Luddites lost–they were violently suppressed by the UK military–but more generally they lost because programmable looms brought patterned clothes to the masses.
https://marginalrevolution.com/marginalrevolution/2026/04/the-luddites-were-the-first-to-attack-ai.html

Carbon capture programs are failing. ‘So far, various projects in the industry have removed 1.3 million tons of carbon from the atmosphere, according to CDR.fyi. That is a minuscule fraction of the roughly 40 billion metric tons of carbon dioxide that the world emits each year.’
https://www.nytimes.com/2026/04/16/climate/microsoft-carbon-removal.html

Improvements to battery energy density and cost mean that Americans can now buy electric cars for less than $40,000 (the $7,500 federal tax credit no longer exists) that get over 300 miles of range per full charge.
https://www.nytimes.com/2026/04/27/upshot/cheap-electric-cars-gas-prices.html

Last year, an astronomy paper was released documenting apparent UFO sightings from Palomar Observatory in the early 1950s. A new German astronomy paper confirms similar sightings from one of their observatories during the same timeframe.
https://phys.org/news/2026-03-unexplained-sky-1950s-independent-analysis.html

The Artemis II mission flew four U.S. astronauts around the Moon for the first time since 1972. Among them were the first black astronaut, first female, and first Canadian.
https://www.nasa.gov/centers-and-facilities/johnson/artemis-ii-mission-milestones-an-image-and-video-recap/

“I’m not really a religious person but there was no other avenue for me to explain anything or experience anything,” he said.

“So I asked for the chaplain on the Navy ship to just come visit us for a minute. When that man walked in – I’d never met him before in my life – but I saw the cross on his collar and I just broke down in tears.”
https://www.dailymail.com/sciencetech/article-15741363/NASA-astronaut-finds-GOD-returning-Earth.html

The “Black Irish” are erroneously claimed to be the descendants of shipwrecked sailors from the 1588 Spanish Armada who interbred with local people. In fact, the Black Irish are the island’s original inhabitants, and they can trace their lineages back in Ireland to prehistoric times.
https://youtu.be/Y5zUSkqMb7o?si=ir1ePIpx6P6bhsDT

A new study provides the strongest evidence ever of evolutionary pressure in the Near East and Europe as a result of the rise of agriculture starting around 10,000 years ago.
https://www.nature.com/articles/s41586-026-10358-1

A new study shows magic mushrooms are an effective treatment for depression.
https://www.economist.com/science-and-technology/2026/02/19/a-psychedelic-medicine-performs-well-against-depression

Donald Trump has fast-tracked research into hallucinogenic treatments for mental illness.
https://apnews.com/article/ibogaine-psychedelic-trump-fda-ptsd-veterans-kennedy-a9940fa57fa1457fc064eb5165003524

Trump as reclassified marijuana to a lower level of hazard.
https://apnews.com/article/medical-marijuana-rescheduling-justice-department-trump-cannabis-1d6722d3aae122b1a91f8e4b6c690268

The hallucinogen “DMT” is the most effective cure known for cluster headaches. ‘Because DMT is so fast-acting, patients can easily titrate it: take a puff from a vape pen, wait 30 seconds, take another puff — until the attack is gone completely, and before DMT’s psychedelic effects kick in. DMT has a very short half-life of 5–15 minutes, so even if you accidentally overdid it, you’d only need to wait a few minutes for the psychoactive effects to subside.’
https://psychotechnology.substack.com/p/near-instantly-aborting-the-worst

‘Gene therapy for a rare type of deafness shows lasting results’
https://www.npr.org/2026/04/22/nx-s1-5791478/gene-therapy-deafness-hearing

Titan of genetics, J. Craig Venter, is dead at 79.
https://www.jcvi.org/media-center/j-craig-venter-genomics-pioneer-and-founder-jcvi-and-diploid-genomics-inc-dies-79

‘Man Traveled 10,000 Miles and Spent Nearly $40,000 to Cryogenically Freeze His Sick Dog So They Can Reunite in the Future’
https://people.com/australian-man-spent-nearly-40000-dollars-to-cryogenically-freeze-his-sick-dog-11930160

There’s a new human cryopreservation company called “Nectome.”
https://www.lesswrong.com/posts/3i5GMhpGbDwef9Rns/nectome-all-that-i-know

Human lifespan is about 50% heritable, which is more than previously thought.
https://www.science.org/doi/10.1126/science.adz1187

Transgendered people have extremely high rates of mental illness and suicidality. Unfortunately, there’s no evidence that getting sex change operations during their teen years improves their mental health later on.
https://dailydeclaration.org.au/2026/04/07/after-gender-reassignment/

Posted on

Review: “M3GAN 2.0”

[Written with the help of GPT-5]

Plot:

“She’s back!”

In M3GAN 2.0, the terrifying AI doll returns in a story that leans more into action, humor, and moral ambiguity than the original. The film follows Gemma (played by Allison Williams) as she grapples with the consequences of her earlier creation while navigating a rapidly escalating technological arms race. A new and more dangerous robot—AMELIA—emerges, with her own mysterious agenda, and forcing former enemies into uneasy alignment.

This movie is all about M3GAN’s redemption arc. Instead of being purely a villain, she becomes something closer to an antihero: still sharp-tongued and unsettling, but now positioned as a necessary ally against a greater threat. The dynamic between M3GAN and Gemma drives much of the story, blending tension, dark comedy, and reluctant trust.

The film expands its scope significantly, introducing corporate conspiracies, experimental technologies like brain implants, and high-stakes set pieces ranging from covert infiltrations to chaotic public confrontations. That broader ambition comes with a tradeoff—I didn’t like how this movie’s plot was more convoluted than the first movie’s. There are more moving parts, more characters, and more overlapping agendas, which can make the narrative feel cluttered at times.

Still, I thought the movie was fun and funny and Allison Williams was stunning. The humor lands more often than in the original, and M3GAN’s personality is given more room to shine, especially in her interactions with humans. The film balances its darker sci-fi elements with moments of levity and spectacle, making it an entertaining, if occasionally overcomplicated, continuation of the story.

Analysis:

Though no year is given for the movie’s events, it is clearly set either in the present day or at most five years from now.

Handcuffs won’t work on robots. At the start of the film, the evil gynoid AMELIA is captured by Iranian soldiers who mistakenly believe she is human. They restrain her with standard handcuffs. Once they lower their guard, she easily breaks free by pulling the cuffs apart with her bare hands. This depiction is largely realistic, with an important caveat.

AMELIA

AMELIA is clearly a specialized combat and infiltration machine, like the T-800 from the Terminator films. A robot designed for that role would likely possess strength far beyond human limits, allowing it to break restraints such as handcuffs, zip ties, or even more robust confinement systems intended for people. More broadly, many elements of the built environment that restrict human movement—locked doors, fences, barbed wire, or walls—would not reliably stop such a machine.

Fortunately, this kind of capability would not be typical. Most real-world robots will be task-specific or general-purpose assistants designed to operate safely around humans. That implies limits on strength, speed, weight, and durability, as well as the absence of combat-oriented programming. In other words, AMELIA represents an extreme edge case rather than a baseline for future robotics.

While I am not aware of any current humanoid robot capable of breaking handcuffs, there is no fundamental engineering barrier to building one. The film’s prediction is therefore plausible in principle, even if it reflects a niche category of machines.

Robots will use excessive force to kill humans. AMELIA surprises and attacks her captors, killing one by punching him so hard that his head is severed from his body. While a sufficiently strong machine might be capable of such force, the depiction is unrealistic in how that force is applied.

A system designed for lethal efficiency would not rely on extreme, energy-intensive blows. Instead, it would favor precise, minimal-force strikes to vulnerable areas such as the throat, neck, or head. A metal hand delivering a controlled but targeted impact would be more than sufficient to incapacitate or kill a human, and would do so with far less energy expenditure. As I wrote in my Terminator Salvation review, hand-to-hand combat with killer robots will be brutally skillful and done in seconds.

To the film’s credit, it later presents a more plausible approach. In a subsequent scene, AMELIA disables one FBI agent by striking his windpipe and knocks another unconscious with a controlled punch. This reflects a more realistic model of how an advanced machine might apply calibrated force.

A related scene involves M3GAN defending herself against armed attackers at a gala party event. She uses martial arts techniques to quickly and nonlethally subdue them. This raises an important point about how machines might handle violent encounters. Unlike humans, robots would not experience fear, panic, or hesitation, and they would not depend on imperfect training under stress. As a result, they could consistently apply the minimum force necessary to neutralize a threat.

Humans—especially police—often resort to lethal force in part because of fear, uncertainty, and limited confidence in close-quarters combat. Robots would not share these limitations. Their “lives” would not be at risk in the same way, either because their cognition is backed up remotely or because their bodies are more durable than human ones and immune to punches, knives and even bullets. Even in cases where they are vulnerable, they would not be driven by self-preservation in the same emotional sense. Some might not be programmed to value their own lives.

Because of this, many encounters that humans perceive as life-or-death would not be for machines. Robots could take actions that would be extremely risky for a person, enabling them to resolve dangerous situations with precision and restraint. With appropriate design, this could allow them to reduce the overall level of violence in such encounters.

Finally, since machines will lack emotions (or at least be able to turn their emotions off at will), they won’t be subject to fatigue, stress, anger, or other emotional states that undermine human judgement and training during crises. They will be completely cool and rational in even the most intense confrontations. Instead of robots mass murdering people, we could live in a future where they behave in ways that sharply decrease the number of violent human deaths, worldwide. And for the reasons I’ve mentioned in this section, robots will make better police officers than humans.

There will be brain implants that let people merge minds with other humans and machines. The film introduces a powerful but highly speculative technology in the form of advanced brain implants. Early in the film, an arrogant tech billionaire demonstrates one embedded in his own skull, and late in the film, Allison Williams’ character is forcibly given a similar device.

Allison Williams with a brain implant

These implants are depicted as deeply integrated into the brain, enabling telepathic communication, sensory sharing (e.g. – seeing what the other person’s eyes are seeing), and even remote control of another person’s body. Allison Williams allows this when she needs M3GAN’s fighting skills to beat up enemies. Users can “speak” to one another though internal dialogue, exchange visual perspectives, and override motor functions.

This portrayal is not consistent with the current state of technology or any credible near-term trajectory.

The core difficulty lies in the nature of the brain itself. The brain does not transmit simple, discrete commands that can be easily intercepted and replicated. Instead, it operates through complex, distributed patterns of neural activity that are still not fully understood. Accurately reading intent, translating it into motor output, and maintaining real-time sensory feedback would require breakthroughs across neuroscience, computation, and bioengineering.

There are also major engineering challenges. Current neural interfaces can only interact with a tiny fraction of neurons, signals degrade over time, and biological responses such as scar tissue can interfere with implants. Additionally, spinal and cortical systems are not passive conduits—they actively process and transform signals.

The film also depicts immediate recovery and full functionality after implantation, which is highly unrealistic. Even in a mature version of this technology, such a procedure would likely involve significant medical recovery time, extensive calibration, and a prolonged training period.

Finally, the visual design of the implant—a metallic device protruding from the skull—is implausible from a user-acceptance standpoint. Real-world designs would almost certainly prioritize minimal visibility for aesthetics.

Overall, this is not a near-future technology. It remains speculative and likely requires decades of advancement before anything approaching this functionality becomes possible. As I wrote in my big list of future predictions, I think we’ll have to wait until 2100 for this technology to exist. Furthermore, as I said in my review of “Physics of the Future”, brain implants will not be externally visible and won’t take the form of metallic protuberances coming out of people’s skulls since humans will find that ugly.

Cybernetics will have cured spinal cord damage. The film also presents a dramatic medical breakthrough: the arrogant billionaire stands up from his wheelchair, revealing that cybernetic implants in his spine have restored his mobility.

The arrogant, disabled billionaire

This technology doesn’t exist, nor will it in the near future, so this depiction is incorrect. Moreover, as I’ve said in my big list of future predictions, I don’t think the technology will exist until the 22nd century. ChatGPT explains the formidable hurdles:

The spinal cord is not simply a bundle of wires transmitting signals from the brain to the body. It is a complex processing system that integrates signals, coordinates movement, and manages reflexes. Bypassing a spinal injury would require recreating highly specific patterns of neural activity rather than simply rerouting a signal.

Another major challenge is decoding the brain’s “language” of movement. Motor commands are distributed across large populations of neurons and are continuously adjusted based on sensory feedback. A functional system would need to read these signals, translate them into precise stimulation patterns in the spinal cord, and maintain a real-time feedback loop.

There are also practical limitations. Current electrode technology cannot interface with enough neurons to replicate natural movement, and implants often degrade over time due to biological responses. Additionally, many spinal injuries involve complex damage rather than clean breaks, further complicating intervention.

That said, early experimental systems have shown promise. Some patients have regained limited ability to stand or take assisted steps using brain–spine interfaces. These results demonstrate that the concept is viable, but fully restoring natural, fluid movement remains a long-term challenge.

Terminators will exist. AMELIA is essentially the same thing as a T-800 from the Terminator films–a humanoid machine that looks externally the same as humans but is gifted with superior strength, speed, and fighting skills, and designed to infiltrate human spaces and to destroy targets. M3GAN’s upgraded body is about as capable. While impressive advances in robots have been recently made, we’re still far from being able to make machines that are this sophisticated, so this prediction the movie made will fail. As I’ve said in my reviews of Prometheus and The Terminator, I doubt androids with these constellations of traits will exist until near the end of this century.

AI will help you navigate the world in real time. One of the film’s more grounded ideas appears in a scene where Allison Williams’ character attends a gala while receiving real-time guidance from M3GAN through an earpiece. It presumably has a small microphone that transmits the sound of her own voice. M3GAN can see her surroundings and provide context-sensitive advice, such as warning her about suspicious individuals or suggesting specific actions (e.g. – “Search his pockets for an access card.”). It’s unclear how the visual data is being transmitted, but maybe she is wearing a tiny camera device on her clothing, or M3GAN has hacked into the building’s surveillance camera system.

This depiction is largely accurate as a near-term prediction.

Modern AI systems—particularly large language models—are already capable of providing useful guidance across a wide range of scenarios. The remaining challenge is integration: combining audio input, visual perception, and real-time processing into a seamless, wearable system.

A setup involving an earbud and a camera-equipped smartphone (or similar device) could plausibly achieve this (this was shown in the film Her, in which the main character also wore and earbud and kept a smartphone in his front breast pocket with the camera unobstructed and facing forward). With continued advances in multimodal AI and hardware miniaturization, such systems could become affordable and widely available within the five years.

Exoskeletons will be able to move around on their own. One piece of technology Allison Williams’ company has developed is a powered, full-body exoskeleton for disabled people. It has integral arms and legs connected to a flat torso, and each part tightly mirrors the wearer’s corresponding body part. Early in the film, her male employee puts it on a demonstrates how it works. Around the midpoint of the film, AMELIA seizes control of it remotely and uses it to chase down and attach the company’s female employee.

While there already are exoskeletons for disabled people, they are intended for people with weak or disabled legs and as such only have legs and lack arms. Full-body units meant for people with complete paralysis below the neck don’t exist and almost certainly won’t in the next five years, so this depiction of the present/near future is wrong. However, the film’s underlying logic contains an important insight.

An exoskeleton must be capable of maintaining balance, coordinating movement, and supporting both its own weight and that of its wearer. This implies that, in principle, it should also be capable of operating without a human inside it. In fact, removing the human simplifies the control problem by eliminating unpredictable movements and reducing load.

Additionally, a properly designed exoskeleton would route forces through its own structure rather than through the human body. If the machine lifts a heavy object, the load would be borne by the mechanical frame, not the wearer. (I touched on this in my review of Edge of Tomorrow.)

Given these considerations, it is reasonable to conclude that some future exoskeletons could move autonomously and even be remotely controlled. The film’s depiction is therefore conceptually sound, even if its timeline is optimistic. So, yes, in the future, some kinds of exoskeletons could, by themselves, chase people down and beat them up with their arms. And if they were connected to the Internet, other people or AIs might be able to hijack them and to remotely command them to do the same.

Exoskeletons won’t be waterproof. The characters defeat the hijacked exoskeleton by dumping a fish tank full of water on it, which short-circuits its electronics. This will prove inaccurate.

If exoskeletons are intended for real-world use, they would need to operate in a variety of environments, including exposure to water. Basic waterproofing and environmental sealing are standard engineering practices for modern electronics and machinery. A system as complex and expensive as a full-body exoskeleton would almost certainly be designed to withstand such conditions.

Robots will be able to detect your vital signs by touching you. At the gala party, M3GAN and AMELIA have their first real confrontation. AMELIA kicks Allison Williams hard enough to knock her unconscious and send her skidding across the floor. M3GAN runs to Allison to make sure she’s still alive, and uses her hand to open her eye and to visually scan it, revealing her heart rate, body temperature, and blood oxygen levels. This depiction is partially plausible but flawed in its details.

Allison Williams getting checked out

It is possible to tell heart rate from subtle visual cues, including changes in blood flow visible in the face or eyes, though this requires sensitive instrumentation. However, the eyes do not provide information about body temperature or blood oxygen levels.

A more plausible approach would involve sensors placed on the forehead, particularly over the temporal artery. Heart rate is of course measurable there, infrared thermometers can measure core body temperature from this location, and optical sensors can estimate blood oxygen levels by analyzing light reflected off the skull. These technologies could, in principle, be miniaturized and integrated into a robot’s fingertip, allowing it to gather multiple vital signs through touch.

Of course, no robots have this feature today, nor are they likely to within five years, so the movie’s depiction is wrong. Whatever the case, in the future, I’m sure domestic robots will have the ability to quickly and accurately take vital signs from humans, either thanks to inbuilt sensors or knowledge of how to use basic medical devices.

Robots will entertain you. At the emotional nadir of the film, when all seems lost, Allison Williams and M3GAN are back at their base, alone. As Allison nearly breaks down crying over their predicament, M3GAN first offers consoling words, and then suddenly breaks into a situationally appropriate song–“This Woman’s Work” by Kate Bush. I thought this was hilarious and recommend you watch it to raise your own spirits, and I also think it’s an accurate and surprisingly important depiction of how machines will soon interact with humans.

Anyone who has played around with LLMs knows they’re capable of producing good poetry, funny jokes, understanding countless nuances of the human experience, and rendering good advice on what to do or not to do in specific situations. With entirely reasonable advances in the technology and in the interfaces through which we interact with them (the earbud + forward-facing camera setup I mentioned earlier), it’s likely we’ll have disembodied AIs that can be invisibly present around us at all times and keep us entertained with jokes, songs, or interesting conversation. Five years is a reasonable timeline for such a thing. And there’s no reason to think the companion robots we’ll have in the farther future won’t be endowed with those same attributes.

The implications of this for quality of life for billions of people are profound. Imagine an easing of loneliness, and every day being simply funnier and more entertaining. We’ve been so fixated on the consequences of machines acquiring our job skills and intelligence that we’ve overlooked the upsides of what happens when they’re as funny and as fun as we are.