Issue #162: OpenAI’s Code Red and the Race to Beat Gemini 3

PLUS: Claude’s 50% engineer boost, Google’s no-code agent builder, Runway 4.5’s video leap and Thrive’s OpenAI partnership

Nathan Lands
December 05, 2025

Good morning, welcome to this week’s Lore Brief, your 3-minute brief of the most important moves in AI and tech.

This issue is brought to you by Factory, the fastest way to ship software with autonomous engineering agents.

OpenAI declares internal “Code Red” over Gemini 3 → Sam Altman paused non-core projects to rush “Garlic”, a reasoning-heavy model that already beats Gemini 3 and Opus 4.5 internally. Launch expected early 2026. Read more here →
Anthropic internal study: Claude boosts engineers 50% → 59% daily usage, 2-3× output, enables full-stack work, but cuts peer collaboration and risks skill atrophy. Read more here →
Google Workspace Studio launches no-code agent builder → Drag-and-drop Gemini agents for Gmail, Calendar, Drive, and 50+ third-party apps like Salesforce. Shareable like Docs. Read more here →

— (@)

— (@)

— (@)

DeepSeek Math V2 claims open-source IMO gold → The 685B MoE model nails theorem-proving with self-verifying loops, scoring gold on IMO 2025. Weights are free on Hugging Face for math agents and proofs. Read more here →
Mistral unveils Mistral 3 family → 675B MoE Large 3 matches closed frontier models on reasoning and multilingual tasks; smaller Ministral 3/8/14B run on phones. All open source and on Hugging Face. Read more here →
Anthropic preps for 2026 IPO → Targets $300B+ valuation after latest $13B round backed by Google and Amazon. One of the biggest AI listings ever coming. Read more here →
Kling O1 launches first unified multimodal video model → Single engine takes text, images, or video as input for generation and editing with strong character/scene consistency. Read more here →
Anthropic acquires Bun as Claude Code hits $1B ARR → Fast JS runtime stays open-source; deal boosts coding agents. Read more here →

See you next week!