AI
Hugging Face TITO Warns Agentic RL Teams About Token Drift
Hugging Face explains how retokenizing tool-using agent rollouts can break gradients, and proposes TITO as a safer training-loop rule.
AI
IBM Agentic CLEAR tracks agent failures across three levels
IBM Research released the Agentic CLEAR paper and open source tool for analyzing agent traces at the system, trace, and node levels.
AI
CoreWeave turns agent training and inference into one loop
CoreWeave’s new W&B-integrated agentic AI platform ties Serverless RL, inference, Weave observability, Skills, and MCP into one operations loop.
AI
Anthropic’s $65B Series H Turns Claude Into a Compute Race
Anthropic’s $65B Series H puts Claude demand, a $96.5B valuation, $47B revenue run rate, and AWS, Google, and SpaceX compute into one story.
AI
OpenAI launches Rosalind Biodefense as a trusted-access test for GPT-Rosalind
OpenAI has launched Rosalind Biodefense, pairing GPT-Rosalind, Codex life-science tooling, and trusted access for public-health defense work.
AI
LFM2.5 8B Brings Tool Calling to Local Agents
Liquid AI released LFM2.5-8B-A1B, a 1.5B active MoE model with 128K context for local tool calling, structured outputs, and agent workflows.
AI
Kog claims 3,000 tokens/s, and coding agents hit a latency wall
Kog KIE tech preview claims 3,000 tokens/s on 8x MI300X. The useful question is what 2B-model, batch-1 latency means for coding agents.
AI
Five LLMs split on 67% of fact-checks, and AI search absorbs the cost
Lenz Research tested 1,000 real fact-check claims across five frontier LLMs and found that 67% did not receive the same verdict.
AI
OpenAI sets o3 and GPT-4.5 ChatGPT sunset dates
OpenAI updated GPT-5.5 Instant, removed Canvas from the GPT-5.5 ChatGPT path, and set ChatGPT-only sunset dates for o3 and GPT-4.5.
AI
OpenJarvis 1.0 brings personal AI onto an Ollama PC
Ollama now supports OpenJarvis v1.0. The release shows how local personal AI changes cost, latency, and data boundaries.
AI
Copilot adds Opus 4.8 with a 15x request multiplier
GitHub Copilot made Claude Opus 4.8 generally available with a 15x premium request multiplier before AI Credits arrive on June 1. Teams should review budgets and model policy before defaulting to it.
AI
Mythos reached ACE on 21 of 41 V8 bugs, Anthropic exploit benchmarks warn defenders
Anthropic published Claude Mythos Preview exploit evaluations and a CVD dashboard. V8 21/41 ACE and 1,596 disclosed flaws reset security triage expectations.