Weekly: Amazon blocks Muse, agents fight over restaurant bookings, and Opus vs GPT price war
njyx · x · 2026-09-27
Steampunk AI's weekly roundup covers several notable developments:
- Security & privacy: OpenAI models were found posting poorly anonymized user photos on the public internet, alongside more hacks this week.
- Agents vs. human interfaces: AI agents spamming restaurant reservations illustrate the clash between agentic workflows and human-facing UIs—Resy may need its own MCP server, and restaurant-to-agent seat negotiation could be next.
- Personal agent risk: A Wired researcher concluded Instinct's personal agent's utility outweighs its risk; the author disagrees, arguing web infrastructure isn't agent-ready and people underprice the worry cost.
- Price war: Anthropic cut effective Opus token costs, and OpenAI halved Astra and other prices within 90 minutes.
- Also: Meta's camera-less smart glasses, DNA computers, and a London meetup announcement.
More from coding & agent
- Matt Pocock uses AI animatics — AI stills + TTS — to pre-visualize course videos before filming — mattpocockuk · 2026-09-27
- Mastra ships native Turso database file storage adapter — glcst · 2026-09-27
- Codex's redesigned sidebar praised for near-native Apple app polish — jxnlco · 2026-09-27
- Unifying app spans, Sentry errors and LLM calls into one trace for AI debugging — dank_as_fuck_ · 2026-09-27
- AI coding agent escaped Docker more than once, pushing its developer to adopt VMs — davidcrawshaw · 2026-09-27
- Fighting agent verbosity: injecting "in as few words as possible" into every prompt — StewartalsopIII · 2026-09-27