Why Meta Could Actually Keep Your Agent Data Out of Training — and Frontier Labs Can't
ivan_bezdomny · x · 2026-09-16
ivanbezdomny argues Meta genuinely could exclude agentic user data from training: it has abundant other data and ran end-to-end encrypted 1:1 messaging, so it knows how to honor such commitments. Labs competing purely on best models with no separate product, by contrast, will by default leak user IP into training — that's their business. Google, like search, has always treated your clicks as training data. Counterpoint: everyone now assumes Claude/ChatGPT sessions leak into training anyway.
Related event: Industry Assumes Chat Logs End Up in AI Training Data(2 posts)→
More from Models
- Users accuse OpenAI of silently degrading models for "suspicious" accounts — LiquidVolatility · 2026-09-16
- Rumored mid-tier model Terra looks dead: work splits between Sol/Astra and cheap Luna — kevinkern · 2026-09-16
- IFM's 7B K2-Horizon nearly matches 27B models, tops GPT-5 on BrowseComp, fully open-sourced — karminski3 · 2026-09-16
- Salesforce launches Koa, its first CRM reasoning model, built on NVIDIA Nemotron 3 Super — NVIDIA Blog · 2026-09-16
- Stealth Startup Unveils Non-Chat Model Claiming 100x Speed and Cost Edge Over LLMs — hardimanjames · 2026-09-16
- Analyst argues Meta could actually exclude user data from training, unlike OpenAI's lawyerly wording — ivan_bezdomny · 2026-09-16