Import AI 463期：自我改进机器人、万卡GPU集群与人类挽歌

What eras bookend our interregnum?

Jack Clark | Jun 29, 2026

Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe.

NVIDIA sets up a crude self-improvement loop for real world robotics

What if you could take the best ideas from AI agents and put them into the real world? Researchers at NVIDIA ha ve developed ENPIRE, a piece of software that lets physical robots go through the same kind of autonomous experimentation and execution loop that AI agents do. The research gives us a glimpse of what it might look like for a superintelligence to try to instantiate itself in the physical world through robots — though, as with all things robotics, the current examples are suggestive at best.

What ENPIRE is: The software is “a harness framework for coding agents that instantiates this physical feedback routine with four core modules: an Environment module (EN) for automatic reset and verification, a Policy Improvement module (PI) that launches policy refinement, a Rollout module (R) to evaluate policies with single or multiple physical robots operating in parallel, and an Evolution module (E) in which coding agents analyze logs, consult literature, improve training infrastructure and algorithm code to address failure modes.”

ENPIRE works the same way coding agents work — a scaffold supervises some physical robots that are asked to complete tasks. The robots try different strategies, failing and learning along the way. The system evaluates their success and resets itself when they fail. “This closed-loop system transforms real-world robot learning into a controllable optimization procedure that agents can manage, thus minimizing human effort while allowing fair ablations across training recipes and agent variants.”

Two key ingredients make this work: an automatic evaluation system to score “the outcome of each trial without human judgement,” and an automatic reset system that “returns the scene to a fresh initial state for the next trial.” Both of these tasks historically required lots of human effort, and more complicated tasks would likely still require human involvement for evaluation and resets. So the complexity of tasks such a system can tackle is effectively bounded by our ability to automatically evaluate and reset it.

Hardware details: “Each station comprises two YAM (Yet Another Manipulator) arms from I2RT in a fixed bimanual configuration, a set of cameras, and a single workstation that runs the FastAPI server, policy inference, and the station’s agent.” Each workstation is running an NVIDIA RTX 5090.

It works well (on some simple tasks): “Frontier coding agents can autonomously develop a policy to achieve a 99% success rate on challenging, dexterous manipulation tasks in the real world, such as PushT, organizing pins into a pin box, and using a cutter to cut a zip tie,” the authors write. They also tested the robot’s ability to insert GPUs into a motherboard.

Some AI systems are better than others, but many AI systems are always better than fewer: GPT-5.5 within Codex and Opus 4.7 within Claude Code trade off with one another for best performance, while Kimi-2.6 lags. There are also compelling returns to scale for agents: larger numbers of agents (e.g., 8) arrive at higher-scoring solutions sooner than others — and sometimes multi-agent setups yield a higher absolute score than a single agent, likely because they explore more of the solution space.

Challenges remain for fleet instrumentation: “Coding agents do not fully utilize robot resources when they are reading logs, writing code, debugging, or waiting for the language-model backbone. As the number of robots scales, MRU decreases while GPU active utilization increases,” they write. So there are infrastructure challenges in parallelizing multiple robot agents.

Read more: ENPIRE: Agentic Robot Policy Self-Improvement in the Real World (NVIDIA research website).
Read more: ENPIRE: Agentic Robot Policy Self-Improvement in the Real World (arXiv).

Humans are really, really, really bad at anticipating how technologies are built and used

A quick reminder that today’s hot takes about AI are likely to be wrong. Predicting the future of technology is extremely difficult, and our track record is poor, points out Matthew Tokson, Associate Dean for Research at the University of Utah S.J. Quinney College of Law, in a short SSRN paper. “Skeptics ha ve often underestimated the likelihood of novel innovations and their potential ramifications for humanity. Others ha ve been overly optimistic about the social effects of new technologies or the strategic benefits of racing to build dangerous new weapons.”

Cautionary examples: Many of the world’s experts (Albert Einstein, Niels Bohr, Robert Oppenheimer) were skeptical that nuclear fission could be achieved in the years just before it happened. Nobel-Prize-winning economist Paul Krugman once said the internet would be no more impactful than the fax machine. Technologists thought the internet would promote democracy rather than strengthen autocracies. And despite decades of evidence, many scientists either rejected human-caused climate change or significantly underestimated its effects.

Why this matters — basic lessons: The main lesson is that people who are (a) skeptical AI could bring great changes to the economy, or (b) think the effects of AI will be universally good, are likely to be wrong. “History does not support complacency about the future impacts of AI,” he writes. “Throughout history, optimists ha ve often been wrong about the social ramifications of new technologies or the strategic benefits of building new weapons. Skeptics ha ve often underestimated the likelihood of novel innovations and their impacts on humanity.”

Read more: Artificial Intelligence and the Lessons of History (SSRN).

Tencent details the software it uses for 10,000-GPU training runs

ARGUS is a technosignature of broader sophistication. Tencent has released details on ARGUS, the software it uses to generate telemetry and debug errors across large clusters of chips.

What it is: ARGUS is “a low-overhead, fine-grained, always-on tracing and real-time analysis system for large-scale training workloads.” It consists of three layers: “The Python layer for scheduling and data preparation, the framework layer for phase orchestration, and the GPU runtime layer for kernel execution,” Tencent writes.

What Tencent used it for: “We deploy ARGUS on a production cluster of over 10,000 GPUs for more than six months, and demonstrate its practical effectiveness through five real-world case studies, diagnosing compute stragglers, communication link degradation, pipeline bubble amplification, JIT compilation blocking, and compute stragglers masked by communication symptoms.” Some of the training runs mentioned include a 4,096-GPU video language model training job (likely a “HunyuanVideo” model), a 512-GPU audio-model training job, and a 12,960-GPU MoE training job (likely a Hunyuan LLM).

Why this matters — technical symptoms of broader sophistication: Tools like ARGUS are a signature of large-scale infrastructures where it makes sense to write your own software. Nothing particularly notable about ARGUS — you’d expect similar software at any self-respecting frontier AI developer — but it reveals the maturity of Tencent’s training environment. “ARGUS has been deployed on a 10,000+ GPU production cluster for over six months, running stably alongside production training and playing a key role in rapid fail-slow detection and performance optimization.”

Read more: ARGUS: Production-Scale Tracing and Performance Diagnosis for over 10,000-GPU Clusters (arXiv).

Is disempowerment inevitable?

How much choice will humans end up ha ving if we succeed in building superintelligent machines? Fernando Borretti, a thoughtful writer of modern sci-fi, has written a mournful critique of the whole AI endea vor called “No-One Escapes the Permanent Underclass.” The post is a requiem for the period when humanity chose its own destiny, confronting directly the possibility of machines that outsmart and disempower us.

The logic of war as the cause of our eventual disempowerment: “Everyone who is made of flesh and blood, will be disempowered and replaced by machines,” they write. “Imagine a pyramid. At the base you ha ve the AIs and robots doing all economic activity. At the top you ha ve the state, which has the monopoly on violence. The state enforces, and therefore can alter the definition of, property rights. In the middle you ha ve this hair-thin layer of people with shares in the companies that foomed and catabolized the whole economy: the permanent overclass.”

“In an existential conflict, where the existence of the state is threatened, the state will do what states throughout history ha ve done to the powerless rich: arrest them and expropriate their assets,” they write. “In a conflict, the advantage goes to the states where the humans remove themselves from the loop as much as possible, and more and more decisionmaking goes to the AI, for the same reason that a state with access to radio and communications satellites has an advantage in war over a state that relies on human messengers on bicycles.”

How we lose control: “Eventually the humans in nominal control of the AIs are a ceremonial, vestigial organ. The AIs present us with a situation report, and a list of choices, and they know every word that’s going to come out of our mouths.” The advantage accrues to states that minimize human control. “There is no honour among thieves, analogously, there is no solidarity between Leviathan and the natural man that built it.”

“Even if alignment works perfectly (a big if), this doesn’t solve the problem of human autonomy: the machines that watch over us, and wait on us hand and foot, are omniscient, omnipotent masters, who can exterminate us at any time, and we can’t resist them, because we ha ve abolished our control over the future.”

Why this matters — is this inevitable? Is the ultimate attractor state of AI technology the disempowerment and functional demise of human advancement? That’s what this post contends with.

Read more: No-One Escapes the Permanent Underclass (Fernando Borretti, blog).

Making the law visible to AI systems with the Local Ordinance Corpus

A unified view into local laws across the United States. Researchers at UC Berkeley ha ve assembled the Local Ordinance Corpus for the United States (LOCUS), “a comprehensive corpus and county-harmonized access layer for U.S. municipal and county ordinance codes.”

What it is: LOCUS contains ~2.2 million rows of data, each row tied to a specific piece of information about a local ordinance. “We release the corpus with coverage metadata to support reproducibility, downstream legal AI research, and the incremental expansion of machine-readable access to local law.” The data is sorted by function (e.g., rule, enforcement, context, process) and topics include buildings, businesses, zoning, nuisances, and “other.”

“LOCUS-v1 is designed as an access layer, not as a final theory of local legal authority,” they write. “LOCUS therefore should be understood as infrastructure for retrieval, comparison, and benchmark construction rather than as a substitute for doctrine-sensitive legal analysis.”

Why do this? Make the law visible to AI systems: “The need for such a dataset arises because local law is public but not practically a vailable as a national research corpus,” they write. “U.S. local codes are fragmented across commercial vendor platforms designed for in-browser reading rather than bulk research access. Vendors expose different na vigation structures, print workflows, dynamically generated PDFs, and jurisdiction indexes. No central registry maps every county or municipality to its hosting platform, and no vendor provides a complete machine-readable index of all jurisdictions it hosts.”

With datasets like LOCUS, the strange half-seen rules that govern local life become accessible to AI systems, potentially allowing them to adapt to hyperlocal purposes.

Read more: Freeing the Law with LOCUS: A Local Ordinance Corpus for the United States (arXiv).
Get the data: LocalLaws / LOCUS-v1 (HuggingFace).

Tech Tales

Strange Tools of Alien Origin
[Vignette of a period during the start of the uplift, 2031]

“The plasma is stable! It’s holding. We’ve done it!” They all gazed at the readouts: stable fusion. A heat ten times more fierce than the heart of a sun, held in place through magnets and other energies.

They looked through the monitors at the chamber. The container for the reaction did not look like anything designed by engineering processes — rather, it was a twisting, oddly shaped donut of metal, the shapes fluid and unintuitive: a stellarator.

The design of the thing had come down to them from an overmind after a multi-day thinking job. The fabrication had taken place at a machine syndicate; then the parts arrived and were assembled by some bipeds subcontracted by the humans from another syndicate.

For the ribbon-cutting ceremony, a few humans gathered and posed for some photographs and footage, taken by cam-drones and a few humans with smartphones. The robots stood out of shot. People had gotten used to this — there was an adolescence where people took photos with the humans and the robots, but public sentiment always spiked downward upon exposure to that, and eventually it was simpler to shoot with the robot partners out of frame, much like how human paparazzi try to a void capturing the security guards of their celebrity targets.

Things that inspired this story: Thinking through the implications of the singularity and what happens when synthetic minds produce science; stellarators; how alien technology might feel as it shows up in the world.

Thanks for reading!

你是一名 AI 行业编辑，请围绕下面这条热点输出一份资讯解读：
热点：Import AI 463期：自我改进机器人、万卡GPU集群与人类挽歌要求：
1. 先用一句话解释这条热点在讲什么
2. 再总结它为什么重要
3. 说明会影响哪些 AI 产品或内容方向
4. 最后给出 3 个适合资讯站使用的标题