How a CDN ended up serving trillion-parameter models: Cloudflare's pivot story
thursdai_pod · x · 2026-11-15
ThursdAI interviews Cloudflare's Michelle Chen on how a CDN ended up serving trillion-parameter models:
- Early GPU scarcity: "people were flying around the world with GPUs in their suitcases"
- Cloudflare bet models would get smaller; they got bigger, forcing a pivot into inference serving
- The company now trains its own models
The same episode also covers Fable 5.2, Anthropic IPO chatter, and Gemini 4 rumors.
Related event: Cloudflare recounts journey from CDN to training trillion-parameter models(2 posts)→
More from Companies & People
- Microsoft France CTO: AI literacy means managing agents, not just writing prompts — JFPuget · 2026-11-15
- California professor teaches 30 classes across 9 colleges for $800K/year using pre-recorded lectures — alvelda · 2026-11-15
- Nvidia bio author Stephen Witt recalls Jensen Huang berating him in interview and erupting at dissent — oran_ge · 2026-11-15
- Ilya's SSI: $32B valuation, zero models or papers after 2.5 years of silence — haider1 · 2026-11-15
- Anthropic CEO Dario Amodei says he hopes IPO will 'underpromise and overdeliver' — KateClarkTweets · 2026-11-15
- Vals AI hackathon kicks off: 3 tracks, $60K+ prizes, live leaderboards — RishiBommasani · 2026-11-15