Meta's Project OT: Planned 60% Team Cuts via AI, Called Off at Last Minute
Reuters reports Meta's Project OT (Organization Transformation) — a plan to cut team sizes by up to 60% through AI adoption. Called off at the last minute, the plan reveals how far Meta was willing to go and the fallout engineers already faced.
Key Details from the Plan
Formed in January at Zuckerberg's Hawaii retreat, Project OT envisioned an "AI native" Meta where smaller, "talent-dense" teams oversee AI workers. Internal documents showed many teams could shrink by 60% via layoffs and reassignments. HR projected total layoffs exceeding the 2022-2023 round (25% of staff). The plan called for two waves: May and November.
What Actually Happened
Hours before the May 19 cuts, Zuckerberg "blinked" — canceling November planning. The May cuts only reached 10% of staff, but some teams still faced 30-40% cuts. Engineers with critical domain knowledge were reassigned to AI data labeling, causing knowledge loss and outages like the Instagram "zero auth password reset" bug where any account could be hijacked via AI bot.
Why This Matters for Engineers
Meta's move reflects a broader trend: companies betting AI can replace senior roles. But the struggles of remaining teams highlight the risk — AI can't yet replace deep domain expertise required for infrastructure and product work.
For engineers, this signals a push toward "AI-native" workflows; understanding what AI can and can't do in your domain is crucial for job security.
📖 Read the full source: HN AI Agents
👀 See Also

The AI Bubble Is Popping; We Just Don't Know It Yet
Big tech's Q2 earnings reveal wild stock swings, shrinking free cash flow, and massive AI capex. The Register's podcast digs into what it means for IT teams.

Microsoft Ends Revenue Sharing with OpenAI, Impact on AI Agents Unclear
Microsoft will stop sharing revenue with OpenAI, its main AI partner, according to a Bloomberg report. The move could affect how developers integrate AI agents via Azure OpenAI services.

IDP Leaderboard benchmark shows Claude Sonnet 4.6 matches Opus 4.6 for document AI tasks
The IDP Leaderboard tested 16 AI models on 9,000+ documents across OCR, table extraction, key extraction, visual QA, handwriting, and long documents. Claude Sonnet 4.6 scored 80.8 overall, essentially matching Opus 4.6 at 80.3, while Haiku 4.5 scored 69.6.

Claude Code v2.1.216: Sandbox Filesystem Toggle, Quadratic Slowdown Fix, and 30+ Bug Fixes
Claude Code v2.1.216 adds sandbox.filesystem.disabled to skip filesystem isolation while keeping network egress control, fixes a quadratic message normalization slowdown in long sessions, and resolves 30+ bugs including OAuth token expiry, worktree isolation, and background agent persistence.