Why is this needed? Why not just have agents that scrape sites and interact with them as a human would? I feel like these standards will come with gotchas and limitations that favor a few big companies.