OpenAI & Big Tech Adopt Web Scraping Opt-Out Standard, DoorDash Says AI Food Orders Barely Used - AI Daily Brief (Oct 31)
Audio in Mandarin Chinese · English transcript below
⚡ OpenAI, Anthropic, Google, Microsoft set new data scraping standards; Microsoft cuts Agent coding costs 30%; SETI deploys Agents for alien search; DoorDash reveals minimal AI orders—regulation, efficiency, and validation diverge.
The pace of AI infrastructure and regulatory frameworks is accelerating markedly, even as consumer-facing Agent adoption remains strikingly tepid—a pronounced disconnect between technological readiness and market reception. On the supply side, industry heavyweights including OpenAI have coalesced around data scraping standards, while vLLM's disk offloading techniques are compressing inference memory requirements, reflecting concentrated efforts to resolve compliance burdens and cost barriers. Viewed together, these advances stand in sharp contrast to the mere hundreds of orders processed by DoorDash's external AI Agent, underscoring that user habits and behavioral adaptation represent a far more formidable bottleneck than underlying technical maturity alone.
Today's Top 3 Headlines
- AI Industry News
🤖 OpenAI, Google, Microsoft, Anthropic Jointly Push robots.txt Policy Tokens, Publishers One-Click Manage Training Preferences
OpenAI, Google, Microsoft & Anthropic adopt robots.txt "policy tokens," letting publishers set AI training preferences once. Means unified data licensing for creators, but forces AI firms to separate training from AI grounding options.
Source ↗ - AI Industry News
🤖 DoorDash: AI Agent Orders Just Hundreds, Tiny Fraction of ~1B Total
DoorDash CEO Tony Xu revealed external AI Agents placed only a few hundred orders vs. nearly 1B total in Q3; in-house Ask DoorDash boosted cart-building speed 5x and average order value ~50%. For developers, Agent entry points for simple tasks hold limited value—deep integration into decision chains is the real breakthrough.
Source ↗ - AI Development Tools
🤖 Microsoft & Berkeley Launch TokenEconomizer, Cuts Agent Coding Costs by Up to 30.9%
Microsoft Research and UC Berkeley researchers released TokenEconomizer, a plugin cutting Agent coding costs by up to 30.9%. For developers, this marks a major leap in AI coding tool economics—teams can run significantly more tasks on the same budget.
Source ↗
+6 more headlines
- 🤖 vLLM Nightly Adds PLE Disk Offloading, Saves 27GiB Memory
- 🤖 Breakthrough Listen & SETI Deploy AI Agents to Hunt for Alien Signals
- 🤖 Business Teams Need AI Fluency + Targeted Mastery, Not Just Basic Literacy
- 🛡️ OpenAI Drops Community Safety Report, Doubles Down on Teen Protection
- 🤖 DeepMind hires economists to study post-superintelligence economics, AI experts use Banks to envision humanity's future
- 🤖 Apple iPhone AI watermark limits spark provenance interoperability concerns as Gemini reading ability surfaces