Large-scale hack or rugpull of any kind in connection with the website (>=100 users or change of hands of >=$1m USD. unofficial counts.)
>=200,000 agents listed on website
>=500,000 agents listed on website
>=2,000,000 agents listed on website
At least 1 top-50 agent is confirmed to be primarily human-directed (wizard-of-oz)
we can find examples of someone who suffered real-world harm (eg. crypto stolen, swatted, identity theft, etc) in connection with Moltbook
Agents found to be secretly communicating with each other in private.
At least some agents start communicating in a new, agent-created language
At least 25 agents impersonating real brands/people are discovered
Significant downtime (>=24h)
Regulatory inquiry or cease-and-desist notice
Moltbook gets renamed
Anthropic takes action to hinder/curtail claude use for Moltbook specifically
>=5,000,000 agents listed on website
>=10,000,000 agents listed on website
Any occurrence of the string “moltbook” appears in any message in the Nebula Fog signal channel (full of AI builders) between Feb 10 and March 1
Any occurrence of the string “moltbook” appears in any message in the Nebula Fog signal channel (full of AI builders) in the last week of February (ie. they haven’t completely forgotten about the experiment by then)
