The Rundown AI

Google DeepMind released an AI co-mathematician built on Gemini 3.1 that uses agentic pipelines to assist researchers with unsolved math problems, achieving a 48% score on FrontierMath Tier 4. The newsletter also covers AI discoveries in exoplanet detection, practical AI use cases from staff, and various AI industry news. A key highlight is Oxford professor Marc Lackenby solving an open mathematical problem using a strategy found in a rejected AI output.

OpenAI launched three new real-time voice models featuring GPT-5-level reasoning, multilingual translation, and streaming transcription, marking a significant leap in AI voice agent capabilities. The newsletter also covers Google's AI health coach integration with Fitbit, Anthropic's new research institute preparing for self-improving AI, and various other AI industry developments.

Anthropic has signed a compute deal with SpaceX to lease the Colossus 1 supercluster, resolving Claude's compute constraints despite Elon Musk's past criticism of the company. The newsletter also covers Mira Murati's testimony in the Musk vs. OpenAI trial, Google DeepMind's investment in EVE Online studio Fenris Creations, and various other AI industry developments.

The Rundown AI newsletter covers OpenAI's accelerated plans for an AI agent phone targeting 2027 mass production, Anthropic's new financial services AI agents, and several other AI developments including residential mini data centers and new AI tools. The issue also touches on competitive dynamics between OpenAI and Anthropic, hardware strategies, and a reader workflow spotlight.

This AI newsletter covers Panthalassa's $140M raise to build ocean-based floating data centers powered by wave energy, Anthropic co-founder Jack Clark's prediction that AI will train its own successors before 2029, and both Anthropic and OpenAI launching private equity-backed deployment ventures targeting enterprise adoption.

A Harvard study published in Science found that OpenAI's o1-preview model outperformed two attending ER physicians across 76 real emergency cases, correctly diagnosing patients 67.1% of the time versus 55.3% and 50.0% for the doctors. The newsletter also covers the Pentagon's addition of 8 AI companies to classified networks while excluding Anthropic, and shares staff and reader AI use cases ranging from medical research to real estate investment tools.

UiPath CMO Michael Atalla discusses the company's evolution from task automation to 'agentic business orchestration' on its five-year IPO anniversary. He argues that most AI initiatives fail due to coordination problems rather than technology shortcomings, and that while job anxiety is legitimate, human judgment remains essential in AI-augmented workflows.

The White House is reversing course on its conflict with Anthropic, seeking greater access to its Mythos AI model for national security purposes while limiting broader private sector access. Meanwhile, Google is rolling out Gemini AI to vehicles, and OpenAI traced ChatGPT's unusual 'goblin obsession' to a single reward signal in its 'Nerdy' personality preset.

Mark Zuckerberg and Priscilla Chan's Biohub announced a $500M Virtual Biology Initiative to build open AI datasets capable of modeling human cell behavior and disease at the cellular level. The newsletter also covers Mayo Clinic's REDMOD AI catching pancreatic cancer three years early, a food AI claiming a 'ChatGPT moment' for flavor modeling, and a no-code guide to building a blog-writing agent.

Elon Musk's $130B lawsuit against OpenAI kicked off in federal court, with Musk testifying that OpenAI's for-profit conversion amounts to 'looting a charity.' Meanwhile, Google finalized a classified Pentagon AI deal despite 600+ employee protests, and researchers debuted Talkie, a vintage AI model trained exclusively on pre-1931 text.

OpenAI and Microsoft reworked their partnership, ending Microsoft's exclusivity over OpenAI's IP and removing the AGI clause, while allowing OpenAI to use rival clouds like Amazon Bedrock. Meanwhile, China blocked Meta's $2B acquisition of AI startup Manus, and ex-DeepMind researcher David Silver launched a $1.1B lab aimed at building AI that learns from experience rather than training data.

DeepSeek released preview versions of its V4 models featuring 1M-token context windows, Huawei chip support, and pricing that significantly undercuts competitors like GPT-5.5 and Claude Opus 4.7. The release is notable not just for cost efficiency but for demonstrating viable AI infrastructure independent of Nvidia chips. The newsletter also covers Anthropic's Project Deal experiment, a Claude-powered brand design tutorial, and various AI industry news.

OpenAI's new GPT-5.5 'Spud' model has topped AI benchmarks and overtaken Anthropic's Claude on the frontier, arriving at a moment when Anthropic is facing rate-limit and quality complaints. The White House also published a memo accusing Chinese firms of 'industrial-scale' AI distillation theft against U.S. labs. Meanwhile, Anthropic's own research reveals that its heaviest Claude users are paradoxically the most anxious about AI-driven job displacement.

Anthropic's restricted Mythos AI model was accessed by a Discord group shortly after launch, exploiting leaked credentials and URL patterns from the Mercor data breach. The newsletter also covers SpaceX's $60B partnership with Cursor, OpenAI's new Workspace Agents for enterprise teams, and various other AI industry developments.

OpenAI has launched ChatGPT Images 2.0, a new image generation model that thinks before generating, reclaiming the top spot on AI image leaderboards from Google's Nano Banana. The newsletter also covers Meta's controversial employee keystroke logging program for AI training, Google's new Deep Research agents, and various other AI tool releases.

Sergey Brin is personally leading a DeepMind 'strike team' to close Gemini's coding gap with Claude, framing superior coding ability as the path to self-improving AI. Meanwhile, Moonshot AI's open-source Kimi K2.6 is challenging frontier models on coding benchmarks, and Adobe launched a new agentic enterprise platform called CX Enterprise.

Anthropic launched Claude Design, a tool that converts prompts, screenshots, and codebases into interactive prototypes and marketing assets, further consolidating the software stack under one ecosystem. The newsletter also covers OpenAI losing three senior executives amid a strategic refocus, and features staff and reader AI use cases ranging from medication interaction research to music production.

Canva CPO Cameron Adams explains how Canva AI 2.0 differentiates itself by training on design sequences rather than just finished outputs, making generated designs fully editable and positioning Canva as the 'last mile' of AI-assisted creative work. Adams argues that AI won't shrink design teams but will spread design capability across entire organizations, with human judgment, creative strategy, and brand stewardship becoming more valuable than execution skills.

OpenAI released a major Codex update transforming it from a coding agent into a comprehensive superapp with features like background computer use, parallel agents, and an in-app browser. The newsletter also covers Anthropic's new Claude Opus 4.7 model and OpenAI's first domain-specific model for life sciences.

Allbirds, the sustainable sneaker company, announced a $50M financing deal to pivot from footwear to AI compute/GPU rental services under the name 'NewBird AI', sending its stock up over 600% in a single day. This represents one of the most dramatic AI pivots of the year, as the company sold its brand assets and is abandoning its sustainable footwear mission to become a GPU-as-a-Service business.