Dots dashboard experiment: hand-rolled scrapers, no Firecrawl, flaky schedules
edwin · x · 2026-10-03
Part 6/7 of edwin's retrospective on having OpenAI's Dots build an AI status dashboard 8 times:
- Scraping: all hand-rolled HTTP with custom JSON, RSS, and HTML parsers — no Firecrawl, Browserbase, or Tavily
- Scheduling: some builds shipped with no active schedule at all; where one existed it used a ChatGPT task rather than a Cloudflare cron
Takeaway: agents still reinvent wheels and remain unreliable at production-grade scheduling and parsing — useful signal for agent engineering.
Related event: Testing OpenAI Dots: It Defaults to Cloudflare Stack Every Time(3 posts)→
More from coding & agent
- Eve agent tool cuts build time 8x to 6.9s by reusing compiler module graph — evilrabbit_ · 2026-11-19
- Dev shares NL-driven testing UI: 'the best interface is chat' — KlausCodes · 2026-11-19
- Devs flag GPT's defensive programming for hiding bugs; OpenAI asks for /feedback cases — pvncher · 2026-11-19
- Flutter ships A2UI 1.0: agents respond with real interactive widgets, not chat text — rseroter · 2026-11-19
- Skills vs MCP vs RAG vs Memory: the four context mechanisms AI agents actually need — MaryamMiradi · 2026-11-19
- Anthropic details agent memory: 'dreaming' lets Claude organize its experience — daniel_mac8 · 2026-11-19