[{"data":1,"prerenderedAt":30},["ShallowReactive",2],{"nr-en-meta-ingenieure-ki-nutzung-bewertung":3},{"slug":4,"title":5,"dek":6,"date":7,"time":8,"publishedAt":9,"updated":10,"updatedAt":10,"dateFmt":11,"updatedFmt":10,"kind":12,"tier":13,"author":14,"authorName":15,"topics":16,"tracker":22,"trackerLabel":23,"headlineStat":24,"image":25,"ogImage":26,"imageAlt":5,"csv":10,"minutes":27,"words":28,"html":29},"meta-ingenieure-ki-nutzung-bewertung","Meta Stops Rating Engineers by AI Tool Usage – End of Tokenmaxxing","The tech giant will no longer measure engineers' performance by token consumption, but by work output. A course correction that shows how problematic gamifying AI tools can become.","2026-09-08","11:57","2026-09-08T11:57:00+02:00","","September 8, 2026","news","standard","ideal-syka","Ideal Syka",[17,18,19,20,21],"AI Integration","Corporate Culture","Performance Measurement","AI Costs","Meta","\u002Fki-preis","AI Pricing & Cost Management","Meta discontinues token-based performance ratings","\u002Fnewsroom\u002Fimg\u002Fmeta-ingenieure-ki-nutzung-bewertung.webp","\u002Fog-nr\u002Fmeta-ingenieure-ki-nutzung-bewertung.en.png",3,508,"\u003Cp>Meta is reversing course: The company will no longer measure its engineers' performance based on AI tool usage. This was announced by Meta executives \u003Cstrong>Maher Saba and Santosh Janardhan\u003C\u002Fstrong> in an internal memo reviewed by The Information. \u003Cstrong>AI dashboards and token counters\u003C\u002Fstrong> will no longer factor into performance reviews – instead, \u003Cstrong>quality, speed, and complexity\u003C\u002Fstrong> of work will take center stage.\u003C\u002Fp>\n\u003Ch2>Quick Facts\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>Meta had introduced AI usage as a performance metric, leading to the phenomenon of \u003Cstrong>"tokenmaxxing"\u003C\u002Fstrong>: employees burned through massive amounts of AI tokens just to rank higher on internal leaderboards\u003C\u002Fli>\n\u003Cli>Internal AI usage is on track to cost Meta \u003Cstrong>billions in 2026\u003C\u002Fstrong> – a key driver of the policy shift\u003C\u002Fli>\n\u003Cli>Starting in \u003Cstrong>2027\u003C\u002Fstrong>, Meta plans to introduce budgets and a centralized AI dashboard to control costs\u003C\u002Fli>\n\u003Cli>In parallel, Meta is testing its new AI agent tool \u003Cstrong>Hatch\u003C\u002Fstrong>, which is designed to autonomously complete computer-based tasks\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>How Tokenmaxxing Happened\u003C\u002Fh2>\n\u003Cp>The principle seemed sound at first: if you want to measure how much your engineers use AI tools, count the tokens consumed. But as with many metrics, it led to unintended consequences. Employees optimized their work not for better results, but to consume as many tokens as possible – a classic case of misaligned incentives. The result: \u003Cstrong>excessive consumption\u003C\u002Fstrong> without corresponding productivity gains.\u003C\u002Fp>\n\u003Cp>Meta apparently caught the problem early. Financial pressure helped: when a company of Meta's scale sees internal AI usage costs heading toward the billions, it's time to take a hard look at what those expenses actually deliver.\u003C\u002Fp>\n\u003Ch2>Hatch Tests Face Privacy Pushback\u003C\u002Fh2>\n\u003Cp>Parallel to the shift in evaluation criteria, Meta is testing its new AI agent tool \u003Cstrong>Hatch\u003C\u002Fstrong>, according to WIRED. The system is designed to autonomously handle computer-based tasks – essentially an autonomous assistant for repetitive or complex workflows. But internal testing reveals friction points: \u003Cstrong>some employees hesitate to connect Hatch to personal accounts due to privacy concerns\u003C\u002Fstrong>. An understandable reservation when dealing with agents that require access to sensitive data and systems.\u003C\u002Fp>\n\u003Cdiv class=\"tbl-scroll\">\u003Ctable>\n\u003Cthead>\n\u003Ctr>\n\u003Cth>Aspect\u003C\u002Fth>\n\u003Cth>Old Model\u003C\u002Fth>\n\u003Cth>New Direction\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\u003Ctr>\n\u003Ctd>\u003Cstrong>Performance Metric\u003C\u002Fstrong>\u003C\u002Ftd>\n\u003Ctd>Token consumption\u003C\u002Ftd>\n\u003Ctd>Quality, speed, complexity\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Cstrong>Cost Control\u003C\u002Fstrong>\u003C\u002Ftd>\n\u003Ctd>None\u003C\u002Ftd>\n\u003Ctd>From 2027: budgets & centralized dashboard\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Cstrong>Employee Incentive\u003C\u002Fstrong>\u003C\u002Ftd>\n\u003Ctd>More AI usage\u003C\u002Ftd>\n\u003Ctd>Better work outcomes\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\u003C\u002Ftable>\u003C\u002Fdiv>\n\u003Ch2>What This Means for German Companies\u003C\u002Fh2>\n\u003Cp>Meta's reversal is a cautionary tale: companies integrating AI tools into their culture shouldn't simply measure usage metrics and treat them as success indicators. That leads to gaming the metric rather than achieving real productivity gains. German mid-market firms launching AI pilots should set clear goals from day one – not "more AI usage," but "better results, faster, with less effort." At the same time, Meta's approach with budgets from 2027 onward shows that even tech giants have realized: AI costs need governance. For German companies, this is a signal not to wait until invoices explode before acting.\u003C\u002Fp>\n\u003Ch2>Sources\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Fthe-decoder.de\u002Fschluss-mit-tokenmaxxing-meta-bewertet-ingenieure-nicht-mehr-nach-ki-nutzung\u002F\">The Decoder (DE)\u003C\u002Fa>\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cem>Editorially owned by \u003Ca href=\"\u002Fen\u002Fautor\u002Fideal-syka\">Ideal Syka\u003C\u002Fa>. Sources and method: \u003Ca href=\"\u002Fen\u002Fredaktion\">Newsroom & method\u003C\u002Fa>. Tips and corrections: \u003Ca href=\"mailto:ai@i6eal.de\">ai@i6eal.de\u003C\u002Fa>.\u003C\u002Fem>\u003C\u002Fp>\n",1788874732420]