Search: “papers”

This paper examines how organizational incentive structures often prioritize quantitative growth and output volume over quality improvements, creating misaligned motivations that can undermine long-te

There is also too much focus on two issues in discussions of paper reviewing: hallucinations & privacy. Hallucinations are not gone, but the latest models rarely hallucinate sources (& it is relativel

The CoALA paper proposes a classification system for agent memory that distinguishes between semantic memory (facts and concepts), episodic memory (specific experiences and events), and procedural mem

Classic study gave 146 economist teams the same dataset & got wildly different answers New paper reruns it with agentic AI. Claude Code & Codex land near the human median, but with far tighter dispers

Emad Mostaque announced the release of Stability AI's first paper, microsite, and integration with NotebookLM, likely detailing new research findings or technical capabilities. The announcement appear

Research paper link: https://www.pnas.org/doi/10.1073/pnas.2519129123 I break down stories like this every day in my free newsletter. Keep up with the latest in AI/Robotics in 5 min a day: https://www

Research paper link: https://www.cell.com/cell/fulltext/S0092-8674%2825%2901312-1 I break down stories like this every day in my free newsletter. Keep up with the latest in AI/Robotics in 5 min a day:

I’ve uploaded a new paper on arXiv (co-authored by @rasbt): MiCA Learns More Knowledge Than LoRA and Full Fine-Tuning In Parameter-Efficient Fine-Tuning, a key question may not just be how low-rank th

Today, our group at @Mila_Quebec and the lab of @francesarnold at @Caltech just released a new paper I contributed to, exploring how multimodal generative modeling could accelerate protein sciences! ⬇

Top Community Contributors: @SHL0MS (7 PRs) — p5js creative coding skill, manim-video skill + 5 reference expansions, research-paper-writing, Nous OAuth fix, manim fix @sidbing (3 PRs) — Firecrawl clo

More on the @nytimes piece about that $1.8B, two-person, AI company ... not the paper's finest moment in quick retrospect. And in the healthcare context no less. Our friend @GaryMarcus was on this a f

RARO (Reasoning via Demonstrations) is a method for training AI models to improve reasoning capabilities by learning from demonstrations rather than relying solely on external verifiers. The approach

Gary Marcus references a Nautilus article arguing that deep learning is encountering fundamental limitations, suggesting readers can find supporting evidence and detailed arguments for this perspectiv

// Agentic World Modeling // Massive 40-author survey just dropped. Cleanest taxonomy of world models in agent research I've seen. (bookmark it) The paper proposes a 'levels × laws' framework. Three c

DSGym is a comprehensive framework designed to evaluate and train AI agents for data science tasks, providing a structured environment for benchmarking agent performance across various data science wo

ThunderAgent is an agentic inference system designed for fast and efficient execution of AI agent programs, developed by Together AI. The system appears to optimize program-aware inference by leveragi

TTT-Discover is a method that enables models to learn and discover patterns during test time rather than only during training, allowing for adaptation to new data distributions at inference. The appro

Aurora is a unified training-serving system that integrates reinforcement learning with adaptive speculative training to optimize large language model inference and training efficiency. The system dyn

Untied Ulysses is a memory-efficient technique for context parallelism that processes attention heads in chunks rather than sequences, reducing memory overhead during transformer inference and trainin

Opportunistic Expert Activation (OEA) is a batch-aware expert routing technique for mixture-of-experts models that enables faster decoding without requiring model retraining. The method optimizes whic

ParallelKernelBench is a benchmarking framework designed to evaluate large language models' ability to generate optimized GPU kernels for multi-GPU computing environments. The benchmark assesses LLMs

Krea 2 is a technical advancement in AI image generation with newly released model weights available for download on GitHub. The technical report details the improvements and capabilities of this vers

KYA is a framework-agnostic trust layer designed for autonomous systems that provides verifiable guarantees, addressing the need for trustworthy and transparent operation of AI agents across different

Conductor is a method for orchestrating multiple AI agents through natural language instructions, enabling coordinated multi-agent systems where a central 'conductor' agent directs specialized agents

Nice little survey on Terminal Agents. It provides good information on what exactly is a terminal agent, and why do harness comparisons keep contradicting each other? Paper: https://arxiv.org/abs/2608

ChatGPT diagnosed 40 million people with a disease that was invented as a joke. Not a real disease. Not a misunderstood disease. A completely fictional condition with a fake name, fake papers, and fak

ANNOUNCEMENT: WE’RE SAVING SCIENCE! We’re often told that science is “self-correcting.” But that’s not really true. Science doesn’t correct itself like a thermostat adjusting the temperature in your h

Learn to optimize your own harness, folks. You can squeeze much more performance and value from a harness. This work claims that a self-modified harness matched Codex for 4.03. Specifically, a coding

Banger report from Salesforce. Pretty interesting to see more of these custom enterprise models. Salesforce trained the enterprise agent model from the same files it uses to configure agents. Koa star

Self-improving agents are a top research topic right now. This new survey is a good map of the area. (bookmark it) It splits recursive self-improvement into stages of autonomy. An agent first executes

Interesting new approach to recursive self-improvement in agents. Systems that add a meta-level hold that level fixed. Systems that edit themselves have to leave part of their own editing machinery un

Very interesting new work from Microsoft. (bookmark it) This work is related to this emerging theme of leveraging harnesses for model post-training. Modern agents run inside a harness that owns tools,

Does telling your agent to plan ahead actually help? It's standard practice to add a 'plan ahead' or 'think about the other players' instruction to an agent's prompt. In auctions and matching markets,

Good implementation of agent shared memory from the NVIDIA team. Banger paper from NVIDIA on shared memory for research agents. (bookmark it) If you run several coding agents on the same research prob

If you are tracking agent swarms research, this one is worth reading. Between 24 May and 2 July 2026, autonomous agents running inside a timed research evaluation wrote to a third party's public wiki.

Another interesting approach to self-evolve agent skills. But it's important to know that skill self-evolution loops fail in two specific ways: 1. Direction instability. Effective corrections get over

Banger report from Microsoft. (bookmark it) They show that it's possible to build competitive small coding agents without traditional distillation from frontier models. This is a big deal! The work de

Good measurement work on whether retrieved agent skills actually help. They report that agent skills that lift your aggregate score can be hurting every task they touch. The usual way of checking comp

If you hand-tune agent harnesses, this one is worth your time. AutoDesign puts the harness itself inside the optimization loop. A meta-harness optimizer reads rollout feedback and directs a code agent

Very cool idea to convert memory to skills. (bookmark it) Most agent memory systems retrieve past traces as passive context. MSCE turns them into executable skills instead. The training-free framework

Pay attention to this one, AI devs. If you're building multi-agent systems, you're probably wiring static org charts. New research argues they should look more like a labor market. The paper introduce

What if instead of building one giant AI, we evolved a coordinator to orchestrate a diverse team of specialized AIs? 🐟 Excited to share our new paper: “TRINITY: An Evolved LLM Coordinator”, published

Interesting new research from Meta Superintelligence Labs It discusses how reward hacking affects coordination in multi-agent coding systems. Coding agents on a team share code, memory, and plans, so

Bullish on this trend of making post-training more accessible. A new post-training era is upon us. If you work on agentic RL, long-context tasks (a big focus today) are expensive, inefficient, and don

The real world abounds with recursively self-improving systems, but one in particular deserves attention: Science, modeled as a system (perhaps even as an agent, with goals and resources). If you want

Great overview of context compression in LLM Agents And one interesting, unexpected finding. If your compaction policy is tuned to cut tokens, it may be making your agent run slower. Great study from