Anthropic expands Project Glasswing to 150 new partners across 15 countries, granting access to its Mythos AI model, which has already uncovered over 10,000 critical security flaws in critical sectors including power, water, and healthcare.
A new attention variant called Wall Attention launches with per-channel, per-timestep multiplicative decay for independent forgetting rates, backed by optimized Triton kernels enabling efficient autoregressive decoding, full GQA support, attention sinks, sliding windows, and sequence packing with verified numerical accuracy.
OpenAI expands Codex with six new plug-ins targeting non-technical users in sales, analytics, and finance, as non-developers now make up 20% of Codex users and are growing three times faster than developers, signaling a major push to compete with Anthropic for a broader audience.
Microsoft unveils a massive AI overhaul at its Build conference, introducing personal agents, a 35-billion-parameter reasoning model, and open-source security tools as CEO Satya Nadella positions the tech giant as a dominant force in the rapidly evolving enterprise AI market.
Snowflake unveils a sweeping AI overhaul at its annual Summit, launching CoWork, a personal AI agent integrating Gmail and Slack, alongside new security features, an autonomous coding assistant, and a deepened partnership with Anthropic aimed at accelerating trusted enterprise AI adoption.
Microsoft unveils Project Solara at Build, an Android-based AI-agent platform powering next-gen enterprise devices, with major retailers like Best Buy, CVS Health, and Target already lining up to pilot the technology alongside chip partners Qualcomm and MediaTek.
OpenAI denies making any political donations or operating a super PAC, clarifying that co-founder Greg Brockman and his wife's personal ties to the Leading the Future PAC are entirely private, while the company reaffirms its stance against political astroturfing and calls for transparent, responsible AI advocacy.
Meta's Chief Information Security Officer Guy Rosen is departing after more than a decade with the company, with Meta actively searching for his replacement as Rosen stays on until a successor is found.
Trump signs an executive order requiring AI companies to voluntarily submit powerful models for government review 30 days before release, establishing an NSA-overseen benchmarking process amid growing concerns over advanced AI systems already uncovering critical cybersecurity vulnerabilities.
Microsoft launches 'Autopilot' AI agents, starting with Microsoft Scout, which autonomously manages scheduling, flags risks, and learns user priorities across Microsoft 365 apps like Teams, Outlook, and SharePoint, now rolling out in private preview.