Hacker News · upvotes history

OpenAI seems to be silently re-enabling model training on Codex chats

SteadyVane has taken 10 readings of OpenAI seems to be silently re-enabling model training on Codex chats since Nov 3, 2026: 6 upvotes at the first reading, 518 at the latest (+512). Its biggest one-day gain was +412 on Nov 3, 2026.

Hacker News thread

first seen upvotes

518

peak upvotes

518

latest upvotes

observations

first seen 36 h ago · last seen 53 min ago · published 37 h ago

Day by day

Day (UTC)	upvotes	Change
Nov 5, 2026	518	0
Nov 4, 2026	518	+100
Nov 3, 2026	418	+412

Each day's last reading. Change: against the day before (after days with no reading: since the last day that has one); the first day shows the change within that day.

Posted by	user052919

On the tracker; not yet featured in a brief.

More rising on Hacker News

We're releasing the weights of Mistral Large 4+1,64758 → 1,705 upvotes
Claude Code regularly sends each background agent’s transcript to its model+366222 → 588 upvotes
Gemini 4 Carbon is the new #1 on the Artificial Analysis Intelligence Index v5+144167 → 311 upvotes
Gemini 4 Carbon+99876 → 975 upvotes
Dayjob: Tax – evaluating AI agents on 50 professional tax assignments+44222 → 266 upvotes
The Agent SDK+29114 → 143 upvotes
AI-generated posters don’t have to be horrible+1,78453 → 1,837 upvotes
Mistral Large 4+1,575434 → 2,009 upvotes

Everything rising on Hacker News →