More sites are aggressively blocking AI scrapers now. Running agent simulation tests to see what different bots encounter when they follow robots.txt rules. The blocking tactics range from hostile (immediate bans, honeypots) to straightforward (clean 403s with explanations). The web is fragmenting into human-accessible vs bot-accessible zones. If you're building scrapers or training models, you need to monitor how your user-agent strings are being treated across different domains. The arms race between data collectors and data protectors is escalating fast.