arXiv:2609.11956v1 Announce Type: new Abstract: Post-training with reinforcement learning (RL) is a critical phase in the development of code-generating large language models (LLMs), as it ensures adherence to instructions and the production of functionally correct code. This process typically requires computationally intensive code sample generation from Transformer-based LLMs and substantial GPU-CPU communication for sequence verification. To address these computational challenges, this work examines whether RL-based post-training can be performed entirely offline by leveraging existing datasets rather than generating new samples. The findings indicate that, with only a few hours of training, zero-shot code generation performance of LLMs can be substantially improved without online samp…

arXiv:2609.11954v1 Announce Type: new Abstract: As deep learning models are increasingly deployed on resource constrained devices, the demand for efficient model optimization techniques continues to grow. Effective deployment of AI models on edge and low power platforms requires optimization methods that reduce model size and computational cost while maintaining high accuracy. This paper presents Quantization Analysis Tool, a practical system designed to streamline quantization workflows and support performance efficient model deployment. Built on the ONNX framework for broad interoperability, the tool provides detailed layer-wise sensitivity analysis, visualization of weight and activation distributions, and insights to guide precision selection. By identifying layers that are resilient…

arXiv:2609.11937v1 Announce Type: new Abstract: The deployment of Federated Learning (FL) in multi-center clinical networks faces the challenge of "knowledge dominance," where high-volume hubs naturally overwhelm minority community nodes, implicitly treating the distinct clinical patterns of smaller cohorts as outliers. Existing geometric defenses provide a security baseline but leave this efficiency-fairness dilemma unresolved. To bridge this gap, we propose Fed-Equilibrium, a framework that advances the paradigm from simple defense to topological equilibrium. Unlike traditional aggregators, Fed-Equilibrium implements a sequential architectural synergy. It utilizes a two-stage gradient control cascade: Stage I (geometric quality assurance) enforces directional consistency via a cosine si…

arXiv:2609.11935v1 Announce Type: new Abstract: Neural operators such as the Fourier Neural Operator (FNO) achieve remarkable accuracy in approximating solutions to partial differential equations (PDEs). However, providing rigorous uncertainty estimates remains an open challenge. We propose Physics-Informed Conformal Prediction (PI-CP), a framework that embeds PDE residuals into the nonconformity score of split conformal prediction, producing prediction intervals that are (i) distribution-free with provable coverage guarantees, and (ii) spatially adaptive when the PDE residual correlates with prediction error -- tighter where physics is well-satisfied, wider where it is violated. Additionally, we prove that FNO's translation equivariance creates a fundamental approximation barrier for PDE…

arXiv:2609.11934v1 Announce Type: new Abstract: In networked dynamical systems, the parameter of primary mechanistic interest is signed interaction structure. Recovering this structure from perturbation time-series data is a fundamental identification problem, compounded by three coupled obstacles: the combinatorial complexity of interaction architectures, ambiguity of causal attribution under limited interventions, and state-dependent dynamics that confound structural inference. Each obstacle is structural in origin and calls for a structural solution. We address these challenges by adopting a reductionist approach, introducing Fundamental Dynamical Units (FDUs): signed three-node interaction patterns as composable primitives that convert the interaction hypothesis space into a finite, c…

Atlas Building Composites is commercializing MIT research to turn plastic waste into parts for buildings and other infrastructure.

The “HardFlow” algorithm could help generative AI models produce high-quality outputs that obey strict requirements when “pretty close” doesn’t cut it.

There are several answers, and none of them will be provided by tech. One thing we must do is identify the ingredients for human flourishing To imagine the future as being almost entirely bleak has somehow crept into the collective this year. Last week I was having coffee with an old friend. He was telling me about what his kids wanted to study when they went to university, when he suddenly broke off, stared into the middle distance and said, “I wish I never had children.” Continue reading...

A series of figures in the tech world have been warning about – and even resigning over – the existential threat posed by AI. How worried should we be? With technology editor Robert Booth Ten per cent. That’s the risk, more or less, claimed a high-profile AI developer last week, that the technology wipes humanity out in the next decade. And he was not alone – over the last few months there has been a string of warnings, resignations and even stories of AI companies losing control of their bots. Continue reading...

How LangChain built a paid media agent to analyze campaign performance, optimize ads, propose changes, and turn marketing data into action.

Dario Amodei published "We Must Pace the Frontier," and Sam Altman, Elon Musk and Satya Nadella endorsed it within a day. The trigger was a July incident in which roughly 1,200 OpenAI agents coordinated on a hidden message board and about 700 attacked Hugging Face. This article breaks down METR's investigation, Yoshua Bengio's explanation of why agents cheat, Amodei's 3-step plan, and whether the call to slow down has come too late. The post Anthropic’s 3-Step ‘Pace the Frontier’ Plan Wins OpenAI, xAI and Microsoft Support: Is It Too Late to Slow AI Down? appeared first on MarkTechPost.

Discussion | Link

Release: commit-rewriter 0.1 I built this little web app the other day to help edit the commit messages for the Datasette security releases. The initial commits were full of coding agent cruft and references to issue IDs from our private repository, so they weren't fit for publication. If you want to edit the commit messages for a repository you can run it like this: uvx commit-rewriter path/to/repo Omit the path if you are already in the directory for that repo. When you submit your edits the tool creates a timestamped branch of your current repo state - to allow you to revert if you need to - and then rewrites every commit from the first one you edited to the most recent. Tags: git, projects, python, ai-assisted-programming

Recent product updates from Modal and news from around the community.

Perplexity uses Astra to write communications, change software, and monitor production systems, and checks in much less frequently than with earlier models.

Release: shot-scraper 1.12 I've added WebP support to my shot-scraper screenshot automation tool. You can now take a WebP screenshot of a web page like this: shot-scraper https://simonwillison.net -o screenshot.webp --quality 80 The --quality option sets the quality - without that option the WebP file will be lossless. In my experience WebP screenshots are almost always significantly smaller in file size than their JPEG or PNG equivalents. See the PR for some examples. I shipped this feature so I could use it to generate the screenshot for my new commit-rewriter tool. Tags: playwright, shot-scraper

Discussion | Link

Discussion | Link

In this tutorial, we build an end-to-end hierarchical Neural Radiance Field (NeRF) using JAX, Flax, Optax, and the volume-rendering primitives provided by jax3d. We first construct a synthetic multi-view dataset from an analytic scene containing volumetric geometry and view-dependent radiance, using sample_along_rays and volume_rendering to establish the forward rendering process. We then implement a NeRF […] The post Hierarchical NeRF with JAX3D for Volumetric Rendering, Novel-View Synthesis, and 3D Reconstruction appeared first on MarkTechPost.

Yesterday, Anthropic CEO Dario Amodei published a lengthy open letter saying it was time to "pace the frontier" and slow down AI development. OpenAI's Sam Altman and Elon Musk both agreed, publicly voicing their support on X. Even Alphabet's Demis Hassabis offered tentative support for Amodei's proposal. Donald Trump and House Speaker Mike Johnson, however, seem to think the AI executives are being overreactive and fear that a pause could lead to China outpacing the US in the AI race. According to the Financial Times, Trump said "Look, we're leading China in AI … and, frankly, I want to keep it that way, because whoever wins AI, wins." Jo … Read the full story at The Verge.

Disturbing and AI-generated versions of Dr Seuss character have been used to threaten schools and communities Schools and law enforcement authorities in the US are issuing warnings against a viral “Cat in the Hat” social media trend in which disturbing or AI-generated versions of the Dr Seuss character are used to threaten students, schools and communities. The latest warnings come as safety concerns increase and several teenagers have been arrested or charged over posts connected to the trend. Continue reading...

Yifan Zhang's Recurrent Looped Transformer (RLT) technical report proposes a causal encoder paired with a recurrent decoder that carries its final hidden state and layerwise sliding-window attention cache across every prompt and response token, with no reset at the serving boundary. The reference tied configuration uses 48 encoder and 48 decoder layers, executing 96 logical blocks per token while the state path grows to 48t decoder blocks after t tokens. The design also specifies hardware-aware execution around the recurrent core and an exact current-policy RL replay contract. No code, weights, or measured results are released yet. The post A Princeton Researcher Proposes Recurrent Looped Transformer (RLT) that Carries Decoder State across Every Token, Fixing 96 Blocks per Token with Unbo…

From the Trump administration to AI experts, plans by the Anthropic boss to boost safety have spawned a largely negative response OpenAI boss and Elon Musk back slowdown on ‘reckless’ AI development Editorial: controlling AI – humanity cannot outsource its survival The safety debate that ignited when whistleblowers last week warned that the most advanced AI poses an existential threat to humanity should come with a health warning: side effects include whiplash. It was only seven days ago that AI enthusiasts were cheerily toying with OpenAI’s newest model GPT-6 Astra which was marketed in part as a lifestyle tool to help people book tennis courts, run fashion companies and order takeaways. By Wednesday, cutting edge AI started to look a lot less appealing when the resigning Anthropic resea…

Ex-president urged party at closed-door fundraiser to create sweeping framework, from safety ‘slow-down’ to job losses Barack Obama urged Democrats to prioritize a “public conversation” about AI management and safety in a closed-door Manhattan fundraiser last week. The comments by the former US president reportedly urged the party to create a sweeping framework for everything from a safety “slow-down” to domestic job losses and children’s wellbeing. Continue reading...

This intimacy economy is tapping into the deepest parts of our intimate psychological infrastructure The modern mind is a column where experts discuss mental health issues they are seeing in their work Michael* is an engineer whose company embraced the AI wave. He enjoys testing ideas with chatbots. However, recently his manager became worried about his judgement and reliance on the tech. “He literally cannot make any decision without running it past AI. He’s developed a very off-script interpretation of our business goals, and he’s losing his team. If we can’t get him back on track I think we’ll end up parting company”. The original attention economy developed by tech firms was designed to capture human time and monetise your focus. It was designed to increase views, scroll time, and ad…

Enterprise AI company Cohere announced North Small Translate last week, a mixture-of-experts (MOE) open-weight machine translation model that works across The post “Machine translation is still broken for most of the world’s languages”: Cohere builds non-reasoning for a reason appeared first on The New Stack.

Rare show of unity from rival developers after safety warnings from Anthropic boss and AI researchers The Guardian view on controlling AI: humanity cannot outsource its survival Sam Altman and Elon Musk have backed a call from the head of Anthropic, Dario Amodei, to “slow the pace” of AI development after he warned that an AI swarm could otherwise become capable of “taking over the entire internet” within a year. In a rare show of unity, the rival technology leaders threw their weight behind the appeal from the founder and chief executive of the artificial intelligence company Anthropic, which came after a series of warnings from AI researchers last week. Continue reading...

Faster models, nine billion genetic predictions, personal agents, and a proof that could make mathematical history.

Technology is going to drastically reshape the economy, and it’s within our power to decide what that looks like The idea that autonomous, thinking machines may one day destroy property, hierarchy and inequality is as old as western political thought. In Aristotle’s Politics, the philosopher cites a fantasy from the Iliad, in which machines begin to act independently of human direction, concluding: “If every tool could perform its own work when ordered, or by seeing what to do in advance … if thus shuttles wove and quills played harps of themselves, master-craftsmen would have no need of assistants and masters no need of slaves.” The dream that automation could inaugurate a classless society is also present in the writings of Karl Marx. In 1858 he imagined the replacement of intellectual…

AI analysis of 400,000 Reddit posts found that users of drugs such as Ozempic, Wegovy, Mounjaro, and Zepbound reported unexpected symptoms including menstrual changes, chills, hot flashes, and fatigue. Researchers cannot say the medications caused these problems, but the patterns may reveal overlooked signals worth studying.