OpenAI doubles its annual revenue to $10B with 500M weekly users, while Anthropic restricts API access over TOS breaches. Plus. Meta attracts AI experts from Apple.
OpenAI disabled a ChatGPT feature that let shared chats appear in search results after privacy concerns arose from users unintentionally exposing personal info. It was a short-lived experiment.
Anthropic revoked OpenAI’s access to the Claude API this week, citing violations of its terms of service, according to sources familiar with the matter.
Mark Zuckerberg outlines Meta’s vision of AI as personal superintelligence that empowers individuals, contrasting it with centralized automation, and emphasizing user agency, safety, and context-aware computing.
OpenAI reached $10B in annual recurring revenue, doubling from last year, with 500M weekly users and 3M business clients, while targeting $125B by 2029 amid high operating costs.
OpenAI and Microsoft are renegotiating their partnership as OpenAI pushes to restructure its business and gain cloud flexibility, while Microsoft seeks to retain broad access to OpenAI’s tech.
Meta has poached four AI researchers from Apple’s foundational models team in a month, highlighting rising competition and Apple’s challenges in retaining talent amid lucrative offers.
Microsoft launched Copilot Mode in Edge, an AI feature that helps users browse, research, and complete tasks by understanding open tabs and actions with opt-in controls for privacy.
Amp is an agentic coding tool engineered to maximize what’s possible with today’s frontier models—autonomous reasoning, comprehensive code editing, and complex task execution.
Mistral AI’s full-stack coding platform helps enterprises cut dev time by 50% with on-prem deployable models, integrated IDE tools, and agentic workflows—addressing security, scalability, and codebase integration challenges.
AI SDK v5 by Vercel introduces type-safe chat, agent control, and flexible tooling for React, Vue, and more—empowering devs to build maintainable, full-stack AI apps with typed precision and modular control.
Google’s Deep Think, now in the Gemini app, boosts creative and complex problem-solving via parallel reasoning, achieving top scores in coding and math benchmarks and aiding scientific inquiry.
An open-source platform for developing LLM applications. Its intuitive interface combines agentic AI workflows, RAG pipelines, agent capabilities, model management, observability features, and more—allowing you to quickly move from prototype to production.
Fully hosted on US servers by Cerebras Systems, it processes at ~2000 tokens per second.
Over 220 PRs from more than 95 unique contributors were merged, introducing new features like custom commands, VIM mode, and more.
Diagram shared by Eric Zakariasson showcasing how the context window gets filled up over time in Cursor.
Kiro will be releasing pricing tiers shortly alongside opening up their waitlist.
Upload PDFs, images, code files, and more to AI-powered apps that work with your data.
Do threats or tips improve AI performance, as Brin suggested? We found no effect on average though some individual questions varied.
Using Claude Code, Sajal Sharma shows how AI boosts dev productivity—when paired with clear specs, architecture-first thinking, and strong human oversight to avoid technical debt.
Deploying agent systems in enterprise settings often faces obstacles due to the absence of domain-specific process knowledge in typical models. This results in chaotic planning, omitted tools, and unstable execution. To tackle these issues, this paper presents Routine—a structured multi-step agent planning framework that provides clear instructions and smooth parameter transitions to ensure stable execution of complex tool-calling tasks.
A curated weekly roundup of AI news, tools, and breakthroughs to stay ahead of the curve in a fast-moving industry.
Curated by devs, for devs. Helping 400+ developers stay ahead of the AI curve.