Meta Adjusts AI Performance Review Criteria, Drops Token Usage Metric
Summary
According to reports, Meta told engineers that performance reviews will no longer include AI token usage, shifting focus to work quality and actual impact. Meta stated that 93% of internal code changes are now assisted by AI agents, and the goal is not AI adoption rate itself but higher-quality delivery and faster iteration. Additionally, Meta noted that internal AI usage is growing rapidly, and the AI tools used by employees this year could cost tens of billions of dollars.
Bubble analysis
This news is only tangentially related to the AI bubble question, as it mainly concerns Meta's internal performance management adjustments rather than investment or capital expenditure. Although it mentions AI tool costs could reach tens of billions of dollars, it lacks specific figures and investment cycles, so it cannot directly serve as bubble evidence.
Related signals
linked by shared tags
- Tech Giants Rapidly Update AI Models, Highlighting 'Model Fatigue'#meta
- 'Model fatigue' sets in as AI labs race to roll out new versions at frenetic pace#meta
- IFA 2026 On-site: Skip Embodied AI and AI, Europe Doesn't Believe in 'Black Tech'#ai-adoption
- Meta Offers Data-Sharing Discounts for Muse Spark AI Models, Priced at 1.3-8% of Original#meta
- Meta says Muse Spark 1.3 has frontier performance — but its best results come from a model developers can’t broadly use yet#meta