Discussion

• @gergelyorosz Gergely Orosz on x

  Token usage is part of perf evaluations at Meta. This is just smart people (Meta only hires smart folks) hitting targets they assume leadership wants them to hit so they get that exceeds expectations (or above) rating + avoid below expectations (Perhaps this is Meta's goal btw)

• @alexolegimas Alex Imas on x

  Focusing on the input and not the output is literally the most Meta thing to do.

• @frank_liquid Frank on x

  Meta is one third of Anthropic revenue? 60T tokens / mo = $900M / mo = $10B ARR for Anthropic 🤯? This is also the largest enterprise contract in history. [image]

• @nickwingfield Nick Wingfield on x

  Orange Theory for nerds? @jyoti_mann1 shows how Meta employees can earn ‘Token Legend’ status and other leaderboard achievements by maxxing out their AI usage. https://www.theinformation.com/ ...

• @melindabchu1 Melinda B. Chu on x

  How is tokenmaxxing a measure of productivity or value?  I can write some bad code which causes an infinite loop and use up millions of tokens.  What is the output of this tokenmaxxing which has resulted in good products or positive outcomes for Meta?  I totally understand R&D in…

• @mitsuhiko Armin Ronacher on x

  I heard some stories about this in the last few weeks and it did not sound like a great work environment. https://x.com/...

• @joshgholder Josh Holder on x

  Every Meta employee I know, to a man, is desperately trying to leave the company

• @qurrie_com @qurrie_com on x

  @mitsuhiko imagine being stupid enough to tie the “performance” of your employees to something that costs you money

• @petergyang Peter Yang on x

  I don't know measuring productivity by token usage sounds almost as dumb as measuring by lines of code written.

• @theinformation @theinformation on x

  Exclusive: Meta employees are competing internally to become “Token Legends,” ranking themselves by how much AI compute they consume. The leaderboard reflects a new status game where token usage is tied to productivity and influence. https://www.theinformation.com/ ...

• @rhorider @rhorider on x

  Many such cases. Companies are measuring productivity to max token burn...even tying pay incentives to it Likely a reason Anthropic has been slamming the brakes on token usage. I imagine there's serious internal panic over models for true compute scale / cost needs rn

• @aakashgupta Aakash Gupta on x

  60 trillion tokens in 30 days. The math on Meta's “Claudeonomics” leaderboard is one of the wildest things I've seen this year. At Anthropic's Sonnet pricing ($3 per million input tokens, $15 output), even a conservative blended rate puts that north of $180 million a month. On

• @shakoistslog Shako on x

  From a game theoretic sense, I wonder if treating this as a KPI, but awarding max value to the 85th percentile would work, and penalizing people below it linearly, and above it non-linearly, would work.

• @maxleiter Max Leiter on x

  Texting my friend at meta [Screenshot of conversation: “True?”  Reply - “Yes, its sad"]

• @kakashiii111 @kakashiii111 on x

  “...competing on an internal leaderboard called “Claudeonomics”...” Apologies, Meta's FCF.

• @theprimeagen @theprimeagen on x

  The only winner here is Anthropic Definitely not the product Definitely not the engineers / staff Definitely not the users

• @thdxr Dax on x

  we do the same thing except we make fun of whoever is #1

• @jyoti_mann1 Jyoti Mann on x

  The highest ranked individual user averaged 281 billion tokens, which could cost millions of dollars, depending on the type of model used. https://www.theinformation.com/ ...

• @jyoti_mann1 Jyoti Mann on x

  Exclusive: Meta employees are “tokenmaxxing” and competing on an internal leaderboard called “Claudeonomics” for status as a token legend. Over a recent 30-day period, total usage on the dashboard topped 60 trillion tokens.

• @theinformation @theinformation on x

  Exclusive: Employees at Meta Platforms who want to show off their AI superuser chops are competing on an internal leaderboard for status as a “Session Immortal”—or, even better, “Token Legend.” Reporting from @jyoti_mann1 👇 https://www.theinformation.com/ ...

• @thdxr Dax on x

  if these are for coding agents it's likely that 99.8% of them are cached input tokens these don't cost 0 but they cost like 8x less and on top of that i'd guess they probably have a 20% discount if not more still crazy numbers though

• @zephyr_z9 @zephyr_z9 on x

  Bruh ur math is completely wrong

• @ericjang11 Eric Jang on x

  from my recent blog post: “Based on my own usage patterns, it's beginning to dawn on me how much inference compute we will need in the coming years. I don't think people have begun to fathom how much we will need.”

• @metacriticcap @metacriticcap on x

  @TheStalwart Very early in its development, the Party would set GDP growth goals for each province in China. As you know, you can do lots of silly things to boost GDP growth and the Party officials certainly did. But the opportunities for China's development were so vast, that si…

• @mikeisaac Rat King on x

  @kevinroose my q is how long does meta tokenmaxxing last before they realize this is a big waste of resources!

• @mikeisaac Rat King on x

  this feels......misguided

• @mikeisaac Rat King on x

  funny — meta product growth director sent out an internal memo today with the title “token usage is NOT impact” after the twitter chatter the past 24 hours after @jyoti_mann1's story on “tokenmaxxing”