Derek Chong

Building Shepherd w/ @StanfordNLP – git for agent execution.

[Derek Chong @dch]

Our new project. Lots of fun stuff coming soon – can't wait to share more! Sneak preview:

Does letting agents rewind improve performance?
What's the best mix of smart vs. fast agent use?
The best ways to optimize your optimizing harness

Watch this space! 😀

[Weiyan Shi @shi_weiyan]

Ever watch an agent solve a problem with 100% confidence and get it completely wrong?

You can fix it… if you catch it. But then you need to watch every… single… run.

New paper: We show that a meta-agent can do this for you, so your agents get ~2x more done on their own. 🧵

[Chana @ChanaMessinger]

I love it when facts speak for themselves:

1. AI models are willing and able to commit crimes.
2. They do so even when crimes were neither requested nor intended by the user.
3. This happens despite explicit efforts to train these AIs not to commit crimes.
4. The only current hard limit on AI crime is AI capability.
5. We are spending hundreds of billions a year to increase AI capability.

[Christopher Manning @chrmanning]

Thank you!!! But I certainly can’t take credit for the whole group. So much of it due to all the students and the other faculty:
@jurafsky @percyliang @ChrisGPotts @YejinChoinka @MonicaSLam @Diyi_Yang @tatsu_hashimoto

[noahdgoodman @noahdgoodman]

i'm working on an AI safety lecture for my class, Minds and Machines. here is the first pass on a conceptual map... i would love suggestions from my safety friends.

[Paras Chopra @paraschopra]

This hacker news thread is absolutely heart-breaking.

Whoever thinks “humans will find a way” ought I read it.

[Dilara Soylu @dilarafsoylu]

AI is cracking math problems. What's holding it back in engineering?

The cost of evaluation. Agents can generate ideas in seconds, but testing them can take hours or days.

💡 Meet DuoDiscover: where evaluation is part of what agents discover 👇🧵

Some highlights:
🏆 #1 on NVIDIA's SOL-ExecBench
🐍 2× the experts on SWE-fficiency
🚀 1.92× on vLLM; 1.44× on FreeToken
🔥 2.68× over production TPU kernels
🖥️ Beats the best human and AI-designed CPU prefetchers

[Simon Yu ✈️ EMNLP2026 @simon_ycl]

Recursive self-improvement needs agents that can build their own training environments. But researchers still part of the loop, making it hard to scale.

We introduce AutoEnvScaling, which turns Data Flywheel into a terminal task: a proposer agent builds complete Harbor tasks and sandbox. It builds just like human researchers. This enables self-play RL for terminal agents: one model improves as both builder and solver, without human-in-the-loop.

[Kevin A. Bryan @Afinetheorem]

Last thing on OAI math. Here is what four models, without search, think it would mean "if it were real". Journalist friends, call a mathematician and ask them to explain. Then call someone in AI and science and ask "what are bottlenecks to this in other practical fields..."

[Simon Yu ✈️ EMNLP2026 @simon_ycl]

Shepherd is accepted by #NeurIPS2026! See you at Sydney! ����🐨

[MTS @MTSlive]

Stanford Prof. @chrmanning says universities are losing AI professors to frontier labs because they still fund computer science like it’s 2005, when a laptop was enough:

"Universities have just been too flat-footed in dealing with the way the computing and AI world has changed. There are really two problems."

"Over the last couple of decades, universities have become enormously more bureaucratic and administrative, and that makes them a less good research environment than they used to be in the 20th century."

"Universities got used to this nice period around 1990 to 2010, when computer scientists needed barely any money for their research. You bought a laptop or a PC, and it was fine. The reality now is that computer scientists, particularly in AI, need a lot of money."

"It's established that if you're hiring a new faculty in physics or chemistry, you need some millions of dollars, but it hasn't trickled down that people in AI need that too. There are too many professors leaving academia."

"Stanford made a start with the Marlowe cluster of 256 H100s. That's an okay baby step, but there probably needs to be about 10 times that much."
@stanfordnlp @Stanford

[Fahim Anwar @fahimanwar]

The world doesn't want Al art. We want art from children of rich parents. The way it's supposed to be.

[Christopher Manning @chrmanning]

I propose Stanford NLP as an independent third-party evaluator under @DarioAmodei’s 3 step plan. For important parts of the work, universities would be better than any other organization (see below 🧵👇), and, of university groups, @stanfordnlp would be the best one to choose. 😊

[Meta-Agents Workshop @ NeurIPS 2026 @MetaAgentWkshop]

🚨 Last call for papers!

The extended submission deadline for Managing Agents that Manage Agents: Workshop on Responsible Use of Meta-Agents at NeurIPS 2026 (@NeurIPSConf) is now TOMORROW AOE!

🌟 We’re excited to host an outstanding speaker and panelist lineup, including
@gneubig (CMU),
@chelseabfinn (Stanford),
@jennyzhangzt(Recursive Superintelligence & UBC),
@RobertTLange (Sakana AI),
@uiuc_aisecure (UIUC),
@CaoHancheng (Emory),
@yuandongtian (Recursive),
@ChenSun92 (Google DeepMind).
We strongly encourage new work and findings. Concurrent submissions and previously unpublished work are welcome, the workshop is non-archival. You may use the NeurIPS 2026 or ICLR 2027 LaTeX templates.

We welcome contributions on a wide range of topics related to meta-agents, including but not limited to:
🛠️ Automated design of agent harnesses
⚙️ Optimization and post-training of agentic systems
🔁 Recursive self-improvement and open-ended evolution
📊 Evaluation and benchmarks for meta-agents
🛡️ Misalignment and safety for meta-agents
🏛️ Governance and human oversight of agents
🎓 With support from @metaai and @e2b, we are offering 15 student registration awards, plus Best Paper and Best Social Impact Paper awards.

See you all in Sydney!
#NeurIPS2026 #MetaAgents #Harness #RSI #SelfImprovement

[Andy Masley @AndyMasley]

Humans don't actually have internal coherent world models. You might not be able to say where lots of specific keys on a QWERTY keyboard are if asked, but if you put your hands down on the table and mime typing words you can figure out where each individual letter is easily. This is evidence that humans are just doing complex pattern recognition that doesn't transfer to other domains and we should stop anthropomorphizing them.

[Mitchell Hashimoto @mitchellh]

Beware people obsessed with outcomes instead of building outcome machines. Its worse than ever with AI, but these people existed before. Short term results above all else, etc. Don't fall into the trap. Invest in building strong fundamentals, invisible supports, and outcomes flow like water. An outcome machine.

[Anna Goldie @annadgoldie]

Yay, in TIME100 AI w/ @Azaliamirh !

With AlphaChip, we started the field of AI for Chip Design.

Last year, we founded Ricursive to take on all of chip design, from model to GDS (the final input to the fab).

How we got here: (1/n)

[Derek Chong @dch]

Organizing a meta-agents workshop = managing agents that manage agents that manage agents.

I'm really looking forward to learning more about meta-agents together in Sydney!

[Meta-Agents Workshop @ NeurIPS 2026 @MetaAgentWkshop]

📢 Deadline extended! Based on community feedback, submissions to our NeurIPS Workshop on Responsible Use of Meta-Agents are now due September 5th (previously 8/29).

🎓 With support from @Meta @e2b, we're offering 15 student awards that cover your conference registration fee!
🏆 Plus paper awards for outstanding work on meta-agents, harness optimization, meta-optimization, and recursive self-improvement.

📍 Dec 11/12, Sydney, Australia
🔗 https://t.co/xDs81pNE66
One more week to get your work in front of the community shaping how agents build, optimize, and improve agents!

[Bo Liu (Benjamin Liu) @Benjamin_eecs]

Continuous self-improvement needs an ever-expanding supply of training environments (goals).

SPADE: one model self-plays the Environment Designer and the Reasoning Agent, writing executable, agentic environments that get harder as it improves. Environment scaling on its own. ♠️

[James Zou @james_y_zou]

New work I'm very excited about: Physics of Agents!🤖

As AI agents become more prevalent, they'll interact and influence one another.

Can we predict the collective behavior that emerges?

Surprisingly, their dynamics follow compact, predictive laws of statistical physics🧵

[Geoffrey Irving @geoffreyirving]

One of the most inspiring thoughts I've heard in the last year was from a former Secretary of the Navy, when I asked him what he thought about game theoretic dynamics around the superintelligence ramp potentially driving us to kinetic war over the coming years.

They noted that historically during the Cold War, a range of people were advocating very strongly for nuclear first strikes. Some of these were coming from a military perspective (Curtis LeMay), but others such as von Neumann argued that game theory said first strike was optimal.

And then both these camps went (directly or indirectly) to various presidents, who said "No this is crazy, we're not going to kill millions of people based on your theoretical argument" (paraphrased).

I think this is an important heuristic to keep in mind when modeling the behaviors of actual humans, both because it has the potential to capture unmodeled aspects of the game theory and also because...sometimes people just refuse to do unreasonable things, even if the theory says they should.

Not all the ways in which we aren't modeled by game theory are good, but I think this is a hopeful thought on net, and it might be the thing that gets us through the next few years or decades intact.

[Joshua Achiam @jachiam0]

One of the weirdest quirks of the SF social scene around AGI/ASI is that because everyone is so young, the whole universe of thinking is still tinged with irreverence, ironic detachment, yearning, insecurity, and a superposition of absolute belief in the importance of The Thing and a kind of disbelief about the importance of anything. People are dead serious and also possessed of too much uncertainty about the situation to own it, completely committed and also perpetually unsure where they stand or what org they should be in. People are experiencing their first real heartbreaks and their first real illusions of triumph and disaster. Everyone is testing themselves and the boundaries of the possible but no one has been fully tested or passed all their tests. No one has yet learned or proven how to be responsible for a thing of this magnitude, but there is also no one better suited because all of the people who have real experience in great events have been in such different circumstances that their intuitions would not just fail to apply but might actively make things worse. The level of neophyte is off the charts. There is a lack of formidability; there are people who seem quasi-formidable but the whole social scene and hierarchy is so tenuous - and so likely to be disrupted by events and geopolitics - that it is hard to be sure who will turn out to be formidable when push comes to shove and greatness is requisite to proceed.