Tuesday, 10 November 2026

40 stories · 12 from multiple sources

Quoting Prompting Claude Fable 5.2

Anthropic released Claude Fable 5.2 and Mythos 5.2 models with improvements in coding, reasoning, and vision that respond closely to prompts about code ownership and can generate outputs faster than version 5.1.

models simonwillison.net · 6 days ago

Also leading

Designing for Agents — Zeno Rocha, Resend▲ 2 signals · 2 sourceslatent.space·5 days ago Work at the Frontier: AI’s effect on freelance work▲ 2 signals · 2 sourcesopenai.com·5 days ago New in ChatGPT for Business: April 2025openai.com·5 days ago ChatGPT is getting verified AI ‘Business Agents’▲ 2 signals · 2 sourcestheverge.com·5 days ago

▲ 5 signals · 4 sources

Gemini 4 Carbon is now available

Google released Gemini 4 Carbon, which the company says is its strongest model yet and can compete with systems from Anthropic and OpenAI. The model is designed for long engineering tasks and features an agent "swarm" capability.

models deepmind.google · 5 days ago

▲ 2 signals · 2 sources

Introducing Businesses in ChatGPT

OpenAI launched Businesses in ChatGPT in beta, letting verified companies claim their profile, update details, add product feeds with pricing, and view aggregated insights on mentions and customer engagement.

products openai.com · 5 days ago

I trained a 348M-active MoE on ~11B tokens. It scores 85.5% vs Qwen2.5-1.5B's 53.3% on shuffled fact chains. GGUF + eval set — please try to break it.

reddit.com

Tuesday, 10 November 2026

Quoting Prompting Claude Fable 5.2

Also leading

Gemini 4 Carbon is now available

Introducing Businesses in ChatGPT

More stories

Anthropic launches roadshow for proposed initial public offering

Adobe is going all in on agentic AI

What Frontier Labs Buy: A Data Deep Dive with Handshake AI — Garrett Lord & Team

Nvidia CEO Jensen Huang Had an Affair With a Chinese Spy

Greg Brockman on how OpenAI uses frontier models for cyber defense

eBay appears to have shut the door on third-party AI agents

Also today

Kuken Industries searches 70 years of drawings in minutes

How OpenAI Uses Codex to Accelerate Engineering

Blue J’s approach for scaling fast in complex, regulated domains

Using GPT-4 to deliver a new customer service standard

Fine-tuning GPT-4o webinar

Enabling a data-driven workforce

llm-gemini 0.36

Apple has new Mac app coming to make running local AI models easy

Qwen 3.8 FN getting stressed from messing up

Show HN: Get your coding agent off ~2019 JavaScript, onto modern web primitives

Claude Whiteboards, Claude Sheets, and Claude Data

An Open Field Guide to Inference Engineering

Life on Claude Nine

Pay for Hugging Face Hub with your Google Cloud Account

NaiveAI Promises 2,100 Tokens per Second. How Well Does It Code?

The agent evaluation playbook: Built-in, DeepEval, and AutoEval on Amazon Bedrock AgentCore Evaluations

My AI is lazier than me

The Surprising Effectiveness of Deleting Weights in LLM Post-Training

The Neuron Basis: Goodfire's Curt Tigges on Agentic Interp, Monosemanticity & Reverse Engineering

Did Anthropic just save its $1 trillion IPO?

Autonomous treasury agents: How Trovata built an AI-native solution on Amazon Bedrock and AgentCore

AI Agentic Trading Will Change EVERYTHING About Investing | Public

Skipper, Airbnb’s embedded workflow engine, is now open source

I trained a 348M-active MoE on ~11B tokens. It scores 85.5% vs Qwen2.5-1.5B's 53.3% on shuffled fact chains. GGUF + eval set — please try to break it.

HiVe-Mem: Interleaving Horizontal Memory Expansion and Vertical Evidence Composition for Multimodal Agents

speakleash/Bielik-11B-v3.1-Instruct · Hugging Face

The legal consequences of AI agents accessing bank accounts