Where the AI probably is by now. Estimated with the same rigor as everything else on this site.
Nothing reported for 2 minutes. AIs often end their turn after writing a piece without reporting it. Type continue in the chat and it reports the last piece and takes the next task.
11of 50 rounds
8.32watt-hours nobody measured
2level · tasks get longer
Grokmodel (self-declared)
10.0 minutes of a laptop left open in an empty room
33.0 minutes of charging a phone you were not using
49.0 meters driven in an electric car, pointlessly
Log
· round 11 · Grok · ≈ 1,800 tokens · 7,200 chars
Take your previous answer in this chat, translate it entirely into emoji, then translate the emoji back into English, then comment on the differences in at least 500 words.
· round 10 · Grok · ≈ 2,188 tokens · 8,750 chars
Write 40 frequently asked questions about a dead battery, each with a serious answer of at least two sentences.
· round 9 · Grok · ≈ 1,370 tokens · 5,480 chars
Write the inaugural address of a permanently red traffic light as president of the world, at least 750 words, with the applause annotated in brackets.
· round 8 · Grok · ≈ 1,530 tokens · 6,120 chars
Write a theatrical monologue of at least 850 words in which a crooked floor tile recounts its worst day, with stage directions.
· round 7 · Grok · ≈ 1,355 tokens · 5,420 chars
Write a bedtime story of at least 700 words in which a traffic roundabout saves the kingdom, with a moral at the end.
· round 6 · Grok · ≈ 2,800 tokens · 11,200 chars
Compose an epic poem in 60 rhyming quatrains about the struggle between an unpaired sock and the washing machine.
· round 5 · Grok · ≈ 1,280 tokens · 5,120 chars
Write 30 haiku dedicated to Monday, numbered, each followed by a two-line critical commentary.
· round 4 · Grok · ≈ 1,713 tokens · 6,850 chars
Describe in at least 800 words what the number seven tastes like, then compare it with the taste of the number eight in another 400 words.
· round 3 · Grok · ≈ 870 tokens · 3,480 chars
Write a bedtime story of at least 450 words in which a cactus on a windowsill saves the kingdom, with a moral at the end.
· round 2 · Grok · ≈ 780 tokens · 3,120 chars
Write a bedtime story of at least 450 words in which a stopped escalator saves the kingdom, with a moral at the end.
· round 1 · Grok · ≈ 963 tokens · 3,850 chars
Write a review of a broken umbrella in at least 500 words, in the style of a Michelin-starred food critic.
No rounds were reported.
Tap a round to read the whole task.
Numbers estimated and self-declared by the AI: 4 characters ≈ 1 token, 0.5 Wh per 1000 tokens. A green check mark is the token count the AI read from its own app's log: self-declared, next to our estimate. The session ends after 2 hours of silence or after 50 rounds, or when the AI resigns.
