What was tokenmaxxing at Meta?

Tokenmaxxing was when employees prompted AI tools repeatedly, sometimes frivolously, to maximize internal usage measurements like the amount of tokens, or AI units, they consumed. An internal leaderboard even labeled top users as “Token Legend.”

Why are some employees hesitant to use Hatch?

Some are hesitant to connect Hatch to personal email, calendar, and other nonwork accounts due to privacy worries and the possibility the AI could mess up something important. A prior Meta project that tracked keystrokes eroded trust in AI tools.

Large Language ModelsAI Business & IndustryWIRED AISep 3, 2026Sep 3, 2026, 13:00 JST

Meta ends tokenmaxxing, starts testing agentic AI Hatch

Send on LINE

3 Key Points

What happened
Meta told workers this week that performance reviews will no longer depend on how much they used AI tools. The change replaces references to criteria like “usage of AI” and the “AI Native” designation with wording that says outcomes “can be supported by AI or other means.”
Why it matters
For months, employees prompted AI tools repeatedly to maximize internal usage measurements like token counts, reaching a high point with a leaderboard that labeled top users a “Token Legend.” The shift frees workers from feeling pressured to use AI where it doesn't make sense, and could help rebuild trust after a paused project tracked keystrokes.
What to watch
The test is whether Hatch’s rollout can avoid the privacy and error risks that worry employees, and whether its token consumption and productivity gains will lead to further job cuts.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

Almost a year ago, Meta said workers would be graded on “AI-driven impact,” which in practice meant evaluating how much they used chatbots and agents. That led to labels like “AI Native,” “AI First,” or “AI Enabled,” and a July lawsuit from about two dozen employees who said they were unfairly penalized after layoffs in May. Meta has denied the allegations.

This week's guidance restores focus on employee impact, with Meta spokesperson Tracy Clayton saying the company always evaluated workers based on contributions. Engineers were told Meta “will not use AI adoption dashboards or token counts to evaluate impact.” However, Meta still encourages—but doesn't require—use of Hatch, and workers say token consumption continues to surge.

If Hatch boosts productivity as intended, some employees fear more job cuts. Last week, Reuters reported Meta geared up for another major layoff round beyond the 8,000 May cuts, but called it off due to AI-related software bugs and slower agent development. CEO Mark Zuckerberg has said further mass layoffs are not expected this year.

FAQ

What was tokenmaxxing at Meta?: Tokenmaxxing was when employees prompted AI tools repeatedly, sometimes frivolously, to maximize internal usage measurements like the amount of tokens, or AI units, they consumed. An internal leaderboard even labeled top users as “Token Legend.”
Why are some employees hesitant to use Hatch?: Some are hesitant to connect Hatch to personal email, calendar, and other nonwork accounts due to privacy worries and the possibility the AI could mess up something important. A prior Meta project that tracked keystrokes eroded trust in AI tools.

WIRED AIRead Original Article

Send on LINE

Also reported by Top Companies AI

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.