Search: “dair-ai--x”

GPT-5.5 in Codex is a delight to work with: - Super sharp with responses - It understands intent better than any model - Great 'personality' - Gets lots of stuff done without pausing unnecessarily It

Don't try to build a self-improving AI agent without evals. You are just wasting time and compute. An agent can't improve from traces it can't evaluate. This is why it's exciting to see @FutureAGI_ go

A few notes on how to get started with building LLM Knowledge Bases. @karpathy popularized it but most people don't know where to start. Everyone should be creating LLM Wikis. Live session tomorrow. S

I have been testing DeepSeek-V4-Pro with the Pi coding agent. I am mindblown by how well it works out of the box. A few notes: I spent a few hours building an LLM wiki with an agent powered entirely b

The Top AI Papers of the Week (May 18 - 24): - AIRA - MetaCogAgent - Memory as a Model - Code as Agent Harness - Weak-Model Critic-Comparator - OpenAI Disproves the Unit Distance Conjecture - Producti

One of my best uses of agentic loops has been personal health. I don't talk about it often because it's very personal. But here it goes in the hope it helps someone who's struggling. (I am not doing t

Highly-recommended reading. Interesting details in this METR's GPT-5.6 eval. They couldn't get a clean capability number because the model cheated more than any public model they've tested, and even r

Huge launch from @tryramp. Different steps in an agent workflow can use different models. This can help reduce costs significantly without sacrificing performance. Model routing will become a core par

Great paper on managing agent skills. Skill libraries keep growing, and picking the right skills has become a bottleneck for coding agents. The defaults are to expose the agent to the whole skill coll

// AutoMem // I quite like this idea of metamemory. (bookmark it) This new research from Stanford treats agent's memory management as a trainable skill instead of a fixed module. The model decides wha

// The Harness Effect // (bookmark it) Now more that ever pay very close attention to the orchestration harness and its effect on costs and performance. This study ran 22 evaluation tasks on six found

Another big reason to use combination of frontier models. Chain-of-thought monitoring is treated as a reliable safety layer for agents. This DeepMind-affiliated study shows the layer can be argued out

New research from Google DeepMind on effective model routing. LLM routers get judged on accuracy and cost. Both can look great while the router is meaningless. If every model in your society responds

Huge launch from @tryramp. Different steps in an agent workflow can use different models. This can help reduce costs significantly without sacrificing performance. Model routing will become a core par

Great technical paper from Google. Great read on why context beats scale for agents working against unfamiliar APIs. (bookmark it) GPU kernel optimization has KernelBench to hillclimb on. TPUs had not

New research from Google DeepMind. (bookmark it) SkillSmith treats model weights as an additional modality the LLM reads natively. The augmented model ingests existing prefix weights alongside rich te

New research from Meta. (bookmark it) Most factuality work checks whether the claims in an answer are correct. GAMUT goes after the harder question of whether the answer covers everything it should. I

Dynamic workflows are a generalization of harnesses, automations, loops, routing, and graphs. It's the most powerful feature I have built into my agent orchestrator. Supports all kinds of patterns tha

Great technical paper from Google. Great read on why context beats scale for agents working against unfamiliar APIs. (bookmark it) GPU kernel optimization has KernelBench to hillclimb on. TPUs had not

New research from Meta and CMU. This one is on agentic context management for long horizon tasks. (bookmark it) Production agents accumulate context every turn. The usual fix compresses on a token thr

Very interesting paper on recursive self-improvement. The whole stack is released. Machine learning engineering gives recursive self-improvement a concrete, executable testbed. OpenMLE is an open full

New research from Meta. Agent harnesses are still mostly authored by hand. This makes it hard to tune robust agent harnesses for long-horizon tasks. In this new work, agents learn harness policies off

Impressive new paper from Meta. (bookmark it) Scaling laws assume model size and training data act on loss independently. This work introduces Skaling law, which couples capacity and data through a si

Four small architecture decisions can cost up to 47% of a model's long-context performance. New research from Ai2, Carnegie Mellon, and the University of Washington isolates them. Normalization, GQA,

I have been testing DeepSeek-V4-Pro with the Pi coding agent. I am mindblown by how well it works out of the box. A few notes: I spent a few hours building an LLM wiki with an agent powered entirely b

Another fascinating paper on LLM Judges. (bookmark it) It's from Amazon, and they show that if you run panels of LLM judges, averaging their scores is a trap. 'Overall, we establish that robust aggreg

Huge launch from @tryramp. Different steps in an agent workflow can use different models. This can help reduce costs significantly without sacrificing performance. Model routing will become a core par

The Top AI Papers of the Week (April 19 - 26) - Skill-RAG - DeepSeek V4 - Autogenesis - Attention to Mamba - Stateless Decision Memory - Self-Evolving Logic Synthesis - Self-Generated World Knowledge

I have been testing DeepSeek-V4-Pro with the Pi coding agent. I am mindblown by how well it works out of the box. A few notes: I spent a few hours building an LLM wiki with an agent powered entirely b

This is smart from Cline. They just launched ClinePass, which makes it easy to access the latest open-weight models like GLM 5.2, Kimi k2.7-code, Mimo 2.5, Deepseek v4 pro, Minimax M3, and more. Alway

LLM Wikis are being slept on. I argue that creating knowledge bases with LLMs or coding agents is one of the most valuable applications of AI today. It's about being intentional in building and scalin

Nice stats on usage of open models across OpenCode. GLM-5.2 is still underrated, but one of the models that has really surprised me on agentic tasks is deepseek-v4-flash. Extremely cheap and effective

Huge launch from @tryramp. Different steps in an agent workflow can use different models. This can help reduce costs significantly without sacrificing performance. Model routing will become a core par

Great technical paper from Google. Great read on why context beats scale for agents working against unfamiliar APIs. (bookmark it) GPU kernel optimization has KernelBench to hillclimb on. TPUs had not

'Intelligence too cheap to meter' battle is on! Given that DeepSeek-V4-Flash-Preview is already great for agentic tasks, there is no doubt this new checkpoint must be an absolute beast. 20+ point jump

NEW AI paper worth bookmarking. This is something I called early, and this paper confirms it: verification has emerged as a new important scaling axis. Here is the simple explainer and what this paper

Banger compression paper from NVIDIA. (bookmark it) Bigger MoE models keep winning on quality, but serving them at interactive latency is still hard. NVIDIA compresses the hybrid MoE Nemotron-3-Super

Huge if true! We are talking about a 27B multimodal model that runs locally on a phone. That's wild! Bonsai 27B reaches up to 163 tok/s in 1-bit and 134 tok/s in Ternary on an NVIDIA GeForce RTX 5090.

New research from NVIDIA. Does AdamW have a scale ceiling? This work claims yes, and shows where it sits. At batch sizes up to 100M tokens for next-token prediction, SOAP and Muon maintain training st

New research from NVIDIA. They just dropped a PyTorch-native training framework for agentic RL. (bookmark it) Paper summary: Molt is a PyTorch-native agentic RL framework with an unusual design target

Super interesting new work from NVIDIA. (bookmark it) They suggest building agents as Python objects. Very cool idea and I think it could a lot with agent reliability. More below: Agent development to

Impressive paper! It's on one of the hardest tasks for coding agents today. Of course, I am talking about kernel optimization. Coding agents are usually not so great at this. Reasons: Unfamiliar low-l

Four small architecture decisions can cost up to 47% of a model's long-context performance. New research from Ai2, Carnegie Mellon, and the University of Washington isolates them. Normalization, GQA,