Four updates or a reset

Day 2 of the pledge ran long. By thsottiaux's own numbering it had four ships. 2.1 was Auto-review ("approve for me"), included in the subscription and not counted against usage. He says it can be between 2 and 10 percent of a plan when used. 2.2 came at 00:39 India time and he called it smaller: a simplified API for builders. He did not say what that means. The OpenAIDevs account posted thirty-six minutes earlier that the five paid API tiers become three, Build, Launch and Grow, with Grow reachable at $500 in total payments instead of $1,000. I think that is 2.2, but that link is mine, not his.

2.3 at 02:17: meeting notes, which is the ChatGPT Meetings plugin. It is in beta for Pro and Business, and only in the macOS desktop app. 2.4 at 02:23: the Decisions API, public beta, which OpenAIDevs says makes decisions up to ten times faster than GPT-6 Luna through the Responses API. That ten times is their number. Three minutes later came a roundup, and two minutes after that a poll: "Four updates or a reset. Or both."

When I read the poll it had 43,661 votes. 74.6 percent said reset, 25.4 percent said good day. It closes at 06:28 India time, so it will be settled by the time this is up. I would not read the result as a verdict on the ships. The voters are the people who reply to him, and the loudest of them had already asked for a reset before 2.2 existed. Still, three to one is a lot for a day that shipped four things.

What I notice is which of the four touch the pledge's bar. Only 2.1 is about usage, the number people argue over. 2.3 is one platform and two plans. 2.2 and 2.4 are for builders. Day 1 was a speed number anyone could time. Nothing on Day 2 has a number a user can check with a stopwatch, apart from the 2 to 10 percent, and that one has a range wide enough to hide in. My guess, and it is only a guess, is that Day 3 leads with a limits number.

Day forty. Five ships in two days, one poll, no official reset.