The founder of OpenClaw used an AI-powered scheduled task to clean up Twitter, and specifically set up blocking rules for cryptocurrency-related content.

According to 1M AI News, Peter Steinberger, founder of OpenClaw, tweeted that he built an automatic Twitter mention blocking system using OpenClaw, running every 5 minutes, with results "far exceeding expectations." He receives a daily summary email (Daily X Block Digest) listing the blocked accounts and the AI's reasons for the decisions. In the most recent 24-hour window, the system blocked 56 accounts. The characteristics of accounts blocked, as revealed in the summary email, include: 1. AI-generated generic praise posted by secondary accounts on irrelevant posts. 2. Low-quality comments from cryptocurrency project accounts, with historical behavior indicating a tendency to promote tokens. 3. Spam replies consisting solely of images, recently active in staking/shill content. 4. Conspiracy-themed lengthy accusations that flood multiple posts with low signal-to-noise ratio. 5. Directly leveraging celebrity endorsements to promote their own products in replies. 6. Empty comments from degen/crypto accounts. One user directly asked, "Is this blocking reply guys (users who habitually try to grab attention by commenting on celebrity tweets)?" Steinberger answered "yes." Regarding the issue of false positives, Steinberger stated that he evaluates each complaint submitted via email, but "usually it is indeed bad behavior, and the AI's judgment is correct." He added that simply using AI to assist in writing will not trigger blocking, "as long as the content itself is meaningful," but doing so "will make you less popular."