Anthropic's Claude Fable 5.2 doesn't generate images, it paints them

Anthropic's Claude Fable 5.2 doesn't generate images the way other AI models do. Instead of computing pixels, the language model writes Python code that places hundreds of thousand

Medium Just Built a Place to Think Before You Publish. That Might Be the Future of Blogging.

Its new TK app combines private writing, AI assistance, and small communities. But its biggest idea may be that the internet needs fewer… Continue reading on Generative AI »

Why Your AI Transformation Will Fail

A handful of CEOs have stopped trying to add AI to their companies and started rebuilding the companies around it.

Firm linked to Jensen Huang's ex-lover reportedly brokered nearly $12 billion in Nvidia chip orders for China

Megaspeed, the firm owned by ousted Nvidia CEO Jensen Huang's former lover Alice Huang, arranged nearly $12 billion in Nvidia chip orders for Chinese companies between 2024 and 202

AI Proved Signal’s Code Secure. The 26,000 Lines Humans Trust Matter More

Google DeepMind’s agents wrote about 6.5 million lines of Lean proof for Signal’s core library, checked against the compiled binary. The… Continue reading on Generative AI »

With Gemini 3, Google Begins Rebuilding How Work Gets Done

Today, Google released Gemini 3.0. It’s easy to file that under “yet another model release”: higher scores, nicer demos, slightly smarter chatbots. But that framing misses what act

Australia's central bank finds AI hitting young workers first, especially in software and customer service

Australia's central bank sees early signs that AI is hitting young entry-level workers in highly exposed jobs like software development and customer service. Since ChatGPT launched

AI Character Expression Sheets: One Sheet or One Expression at a Time?

OpenAI Releases Open-Weight Safety Models That Rewrite Policy Rules on the Fly

OpenAI released gpt-oss-safeguard, open-weight reasoning models that let platforms define and update safety policies at inference time instead of retraining classifiers—part of a b

Meituan releases a 1.6-trillion-parameter agent model that never touched an Nvidia GPU

Meituan has released the weights for LongCat-2.5, a 1.6-trillion-parameter model built for AI agents and trained entirely on Chinese hardware. On Terminal Bench, it scores 88.4 poi

I Built a RAG System With No AI. It Broke in the Same Places Yours Does.

Most RAG bugs happen before the language model ever sees the question. A 60-line retriever in plain Python shows exactly where, and what… Continue reading on Generative AI »

OpenAI Completes Restructuring: Nonprofit "Control" with Unlimited Investor Returns

OpenAI completed a long-running recapitalization that puts a newly named OpenAI Foundation in control of a public-benefit for-profit arm—while handing the nonprofit equity worth ab

Anthropic CEO Amodei admits "there's a huge uncertainty in what's going to happen next year" days before a record IPO

Days before what could be the largest IPO in history, Anthropic CEO Dario Amodei called buying compute 'the deepest business dilemma' his company faces. Anthropic's commitments now

Can Decision Models Speed Up AgenticRAG? An End-to-End Evaluation

This article covers the implementation of Jev-based search over a Corpus2Skill index. Continue reading on Generative AI »

Thinking Machines Claims 30x Cost Cut for Training AI Models

Thinking Machines published a detailed recipe for “on-policy distillation,” showing how per-token grading from a teacher model can push small students to strong math and assistant

Gemini 4 Carbon takes the MathArena crown but costs 15 times more than GPT-6.1 Sol

For the first time in a while, no OpenAI model leads the MathArena AI math benchmark. Google's Gemini 4 Carbon and Anthropic's Fable 5.2 have knocked GPT-6.1 Sol from the top, thou

Plus AI vs Gamma AI: Which Tool Is Better for AI Presentations?

Qualcomm Announces AI Inference Chips to Challenge NVIDIA

Qualcomm has announced the AI200 and AI250 accelerator solutions for data-centers, aiming to disrupt Nvidia’s dominance in AI inference infrastructure and diversify beyond mobile c

Anthropic says outdated prompting habits are slowing down Claude Opus 5.5

Anthropic has made Claude Opus 5.5 the default model on all paid plans and published a guide that does away with old prompting habits. The main takeaway is to drop "think hard" and

Harness Engineering for Swift: Making AI Coding Agents Follow Your Engineering Standards

OpenAI Launches ‘Company Knowledge’ for ChatGPT

OpenAI introduced company knowledge for ChatGPT Business and Enterprise, connecting workplace apps like Slack and Google Drive using a specialized version of GPT-5, but the feature

OpenAI reassigns robotics researchers to other teams as general-purpose models might make specialized robotics AI unnecessary

OpenAI is dissolving its standalone robotics team for the second time, but not because it lacks data. General-purpose models like GPT-6 Astra can apparently control robots directly

Beyond Prompting: Harness Engineering for AI-Generated Maestro Tests

Anthropic Secures 1M Google TPUs While Keeping Amazon as Primary Training Partner

Anthropic announced today it will deploy up to one million TPU chips via Google Cloud—worth “tens of billions of dollars”—to deliver over a gigawatt of compute capacity in 2026, in

Anthropic can already retrain Mythos with "a fraction of the compute originally needed," says former researcher Coxon

Former OpenAI and Anthropic researcher Jacob Coxon points to a slide from Anthropic's IPO roadshow as a reason he left, saying it shows the company's algorithms are improving not j

I Gave an AI Bot My WooCommerce Store. Here’s What It Got Right and Wrong

The bot is the easy part. The store data is where it goes wrong. Continue reading on Generative AI »

Microsoft Launches Near-Identical Browser Days After OpenAI's Atlas

Microsoft expanded Copilot Mode in Edge with new AI features including Actions and Journeys, days after OpenAI launched ChatGPT Atlas browser.

Runway's open-source Praxis-1 turns video AI into a robot controller

Runway releases its robotics model Praxis-1 with open weights and source code under the Apache 2.0 license. The model uses knowledge from video training for robot control and leads

Does AI Make You Dumber? What the Studies Actually Measured

OpenAI Acquires Sky, to Bring a Mac-Native AI Interface Into ChatGPT

OpenAI acquired Software Applications Incorporated, makers of Sky, a natural language interface for Mac that can understand screens and control apps directly.

Anthropic Brings Project-Based Memory to Paid Claude Users

Claude Pro and Max users can now enable memory that remembers conversations across sessions, with Anthropic's twist: separate memory spaces for each Project to prevent context blee

Wonder Studios Raises $12M to Build AI Production Company

London-based Wonder Studios raised $12 million to scale its AI-native production company model, combining commercial work, IP partnerships, and original content amid a crowded fiel

OpenAI Launches ChatGPT Atlas Browser With Built-In Agent Mode

OpenAI released ChatGPT Atlas, an AI-powered browser that integrates ChatGPT throughout the browsing experience and includes an agent mode that can control websites on behalf of us

Adobe's AI Foundry Lets You Train Custom Models on Corporate IP and Brand Guidelines

Adobe launched AI Foundry, a new enterprise offering that builds custom generative AI models trained on companies’ proprietary IP, marking a major expansion of its Firefly platform

Anthropic Launches Claude for Life Sciences with Benchling, PubMed Integration

Anthropic introduced Claude for Life Sciences, adding scientific tool connectors and specialized features for researchers from discovery through commercialization.