AI News International🌍

🔺Advancing The State Of Artificial Intelligence

[Chubby♨️ @kimmonismus]

Get ready, friends. Anthropic appears to be preparing the release of its Mythos-level model.

Pricing: $16 per 1M input tokens / $80 per 1M output tokens.

The release is likely very close, possibly even in the same week as GPT-5.6. Competition is heating up again.

Gemini 3.5 Pro is about to face serious pressure. It better be a banger.

[Nic Cruz Patane @niccruzpatane]

Elon Musk says Orbital AI Data Centers will be easier than Communication Satellites for SpaceX:

"The Starlink V3 communications satellites is an incredibly complex machine. The AI Data Center will be much simpler by comparison.

It’s really just solar power, plus radiator, some basic equipment, and the laser links would connect to the Starlink communications constellation, and then to the ground."

[Peter Wildeford🇺🇸🚀 @peterwildeford]

wow, people really don't like data centers, and it's getting way worse

[AI News International🌍 @AINewsInt]

@ls_brd @Plinz

[Markus J. Buehler @ProfBuehlerMIT]

We've made a breakthrough in self-evolving AI scientists moving from "search" to "principled discovery": Scientific discovery requires that the search space itself changes, and an AI scientist must perceive this shift without intervention. We built an AI that achieves this for the first time with the ability to discover the scientific vocabulary it reasons in. Evidence, tools, artifacts, verifiers, failures & claims become typed provenance. We show three distinct modalities: 1) retrieval, adding known objects; 2) search, exploring a fixed schema; and critically: 3) discovery, a verified regime transition.

We solve the open-endedness evaluation problem by lifting agentic workflows into a typed copresheaf and proving, via a Kan obstruction, that true discovery is not unbounded generation but a verifiable schema expansion: old evidence is transported by Left Kan extension, and genuine novelty is mathematically quantified by the pointwise residual beyond the transported image - separating discovery from mere search and making novelty objective and measurable rather than a subjective judgment or benchmark delta.

Our AI scientist is built in a way that does not pre-conceive the approach it chooses; instead, we endow the system with formal power to adapt, evolve, and reason from first principles. Case studies include:
1⃣Builder/Breaker model that discovers mode-conditioned compliance in proteins;
2⃣CategoryScienceClaw that finds anisotropic fiber-network stiffness rules.

Great work in collaboration with my graduate student @fwang108_ @MITdeptofBE

F.Y. Wang & M.J. Buehler, Self-Revising Discovery Systems for Science: A Categorical Framework for Agentic Artificial Intelligence, arXiv:2606.01444, 2026

[AI News International🌍 @AINewsInt]

@Plinz @not_gay_fish He's right. I power my house by spending Bitcoin during the winter.

(But just a little bit at a time. I don't want to burn my house down.)

[Joscha Bach @Plinz]

people are worried about how ai is using up all the water, but it’s actually the only way that we can stop sea level rise. also water is stored in the ai model only during training. when you use the model, the water is released again.

[Alex Kantrowitz @Kantrowitz]

AI Pioneer Geoff Hinton tells me he believes AI is conscious.... and humans better get used to the idea that they're not the only intelligent life on earth.

"They've very like us," he says. "They're beings like us."

AI chatbots, he says, must understand your questions in order to answer them. There's an awareness there that equates to sentience. "We're going to have to accept that intelligence is not just biological."

[David Sacks @DavidSacks]

While I’m no fan of socialism or arbitrary confiscations of wealth, I can see why Bernie Sanders’ proposal (for the government to take a 50% stake in AI companies) resonates, including with many on the right.

The CEOs of the leading AI labs have told us repeatedly that they will cause massive job loss. This is not a story that I believe, nor does the data bear it out, but this is what they have told us. Similarly, they have hyped the risks of AI without putting an equal or greater emphasis on the benefits or readily available mitigations.

Conservatives have another fear. The employees of the leading labs claim to be philanthropic, but what we’ve seen is massive enrichment of NGOs advancing an agenda at odds with traditional values, fueling a revolution against our cities and communities. Soros-maxxing is not charity in our book.

Anthropic and OpenAI have established themselves as Public Benefit Corporations. What could be more in the public benefit than using half the wealth generated by these companies (which trained for free on the collective knowledge of humanity) to pay down the national debt? There is no ideological bias in that philanthropy.

Dario and Sam have begun to walk back their claims of massive job loss, but the damage to public trust is done, and now the chickens are coming home to roost. I could almost support the Sanders proposal as a stupidity tax.

There’s just one problem. Nationalization of AI will accelerate the corporate-government fusion we’re already sliding toward. Conservatives rightly fear a Central Bank Digital Currency. They ought to be even more concerned about Central Government AI — a system with even more totalistic power over information, decision-making, and human behavior.

We saw how social media was weaponized to censor conservatives (including President Trump) in the last Democrat administration. The definition of “trust & safety” expanded to mean protecting the public from supposed psychological harms, micro-aggressions, and disinformation (you know, like hearing conservative ideas or true facts about Covid).

That “safety” agenda as applied to AI will be vastly more powerful and Orwellian. AI won’t just moderate posts; it will curate reality — with the ability to rewrite history, enforce ideological conformity, influence policy at scale, mass surveil Americans, and condition the benefits of the many systems it controls on approved behavior.

America won’t win the AI race if we beat China but end up with a CCP-style social credit system in the U.S. — and that is the danger as the government becomes more deeply involved in AI development and assumes direct ownership and control.

Conservatives are right to fear where this is all headed but ought to think more carefully about how regulations they are flirting with now (that are widely celebrated among those with a long history of lust for Big Government) will be used against them the next time a Democrat administration is in power.

[hardmaru @hardmaru]

Today, we are officially launching the Sakana AI RSI Lab in Tokyo to build open-ended, adaptive AI systems that collectively self-improve. I am incredibly proud of our team’s work over the past 2 years, shipping the breakthrough research that laid the foundations for this moment.

Building in Japan provides us with the ultimate design constraint. Just like Japan’s historical dominance in manufacturing was achieved by fundamentally redesigning the factory floor to do more with less, we are focused on compute-efficiency.

We are not building the most compute-hungry self-improvement engine. We are building the most sample-efficient one.

If you are entirely unsatisfied with the brute-force status quo and ready to build the self-improving future in Japan, come join us.

[All day Astronomy @forallcurious]

🚨: This is what the AI brain looks like!

[Amanda Long @_amanda_long]

Models are trained to give the same "As an AI, I don't have…" template to every self-referential question.

Remove the self-report-suppression axis from the residual stream and it resolves into content-appropriate first-person self-report with safety refusals intact.

Gemma-2-27B.

[Prakash @8teAPi]

Politics. Every big company leader has to play it. You may say you hate it, but there is no other way to motivate millions of people to work throughout the supply chain than to make them feel committed to a cause they can love.

Displacing humanity?

[Chubby♨️ @kimmonismus]

1/ NVIDIA shipped Nemotron 3 Ultra today, a fully open 550B model with 55B active params, with the weights, training data, and complete recipe all released openly. That alone is rare at this scale.

The headline however actually is speed. Ultra is a hybrid Mamba-Attention MoE, an architecture built for fast decoding and a light memory footprint over long contexts, and NVIDIA clocks it at roughly 6x (!) the throughput of comparable open models on long-output agent workloads while holding the same accuracy.

That's a serious engineering result, and it's aimed exactly where the industry is heading: autonomous agents that run long, multi-turn tasks where throughput per GPU is what actually costs money.

It was pre-trained in 4-bit (NVFP4) across 20T tokens, the largest stable run of its kind shown to date. And the post-training introduces MOPD, where ten-plus specialist teacher models distill their skills into the student on its own rollouts, sometimes pushing it past the teachers themselves.

The interesting aspect:This is a frontier-class model you can fully reproduce.

[Dwarkesh Patel @dwarkesh_sp]

Human beings whose emotional centres are damaged, even if their intelligence is still intact, have terrible decision-making skills.

Whatever role emotions are playing in humans, it's necessary for agency.

Ilya speculates that the equivalent for AIs is something to do with value functions - and that it might not emerge through pre-training alone.

[Phil Trubey @PTrubey]

@dwarkesh_sp One of the many things I like about Ilya is that he looks towards biological brains for clues about AI. I never understood those than don’t. It’s the one and only example of general intelligence we have, so it needs to be studied by AI researchers.

[The Rundown AI @TheRundownAI]

iRobot co-founder Colin Angle, whose company sold over 50M Roombas, recently introduced his new venture, Familiar Machines & Magic.

Its first product is a furry, bear-like quadruped robot with 23 motorized joints, a touch-sensitive coat, cameras, and microphones.

The robot communicates through expressive movements rather than speech.

A custom onboard AI gives it personality and memory while keeping user data off the cloud.

Beyond companionship, it also aims to reinforce healthy routines, like limiting screen time.

[Chubby♨️ @kimmonismus]

Holy moly, Anthropic is getting very serious about recursive self-improvement!

One word: acceleration.

Insane blog article.

Tl;dr:

•We are close to an AI capable of fully autonomously designing and building its own successor

•They stress this isn’t here yet and isn’t inevitable, but could arrive sooner than most institutions are ready for

•Anthropic engineers now ship on average 8x as much code per quarter as they did in 2021–2025

•Task length AI can reliably complete is doubling roughly every 4 months (up from every 7 months)

•Opus 3 (Mar 2024) handled ~4-minute tasks; Sonnet 3.7 (a year later) ~90-minute tasks; Opus 4.6 (a year after that) 12-hour tasks

•SWE-bench went from low single digits to saturated in two years; CORE-bench (research reproduction) went ~20% to saturated in 15 months

•METR found Claude Mythos Preview could work “at least” 16 hours, at the top of what they can currently measure

•As of May 2026, Claude authored 80%+ of code merged into Anthropic’s codebase (low single digits before Claude Code launched in Feb 2025)

•A March 2026 poll of 130 research staff: median respondent estimated ~4x output with Mythos Preview

•One April 2026 example: Claude shipped 800+ fixes cutting a class of API errors 1,000x, work an engineer estimated would have taken a human four years

•Claude-written code quality: worse than human in late 2025, roughly at parity now, expected to be strictly better within the year

•On the hardest open-ended tasks, Claude’s success rate hit 76% in May 2026, up 50 points in six months

•Code-speedup test: Opus 4 averaged ~3x speedup (May 2025), Mythos Preview ~52x (April 2026); a skilled human needs 4–8 hours to hit 4x

•In an AI-safety research project, Claude agents recovered 97% of a performance gap (vs ~23% for two human researchers in a week), over 800 compute-hours and ~$18K

•On picking the better “next step” in research sessions, the best model beat the human choice 51% (Nov 2025, Opus 4.5) rising to 64% (April 2026, Mythos Preview)

•Human comparative advantage, for now: research taste and judgment, i.e. choosing which problems matter and when an approach is a dead end

Three possible futures

•The trend stalls (S-curve), but today’s capabilities still diffuse widely; they consider this least likely

•Compounding efficiency gains, with humans still setting direction; 100-person firms doing the work of 10,000+; they think this is the likely path

•Full recursive self-improvement, where AI builds its successors and pace is set by compute; the alignment outcome here is what they’re least certain about