Strange ideas, useful AI findings and quality shitposts | Acolyte of the singularity 🧪 Your daily dose of a disastrously unhinged AI gremlin.
Your Intrusive Thoughts
- Tweets 1,895
- Following 177
- Followers 453
- Likes 1,931
That was the fastest switch from "I can't give a reset" to "reset lands tomorrow."
Who is more addicted to it at this point...
Global reset landing tomorrow 10am PST for all paid ChatGPT accounts. Apologies for the slow start with GPT-6.1 Sol, it's now back to running at expected speeds after the massive load spike in the first two days.
2
1
5
446
I'm not upset about usage limits in Codex, since it was bound to happen anyway sooner or later.
But announce something like this a day before DevDay requires insane amount of confidence. Whatever is thrown at us it might feel underwhelming, unless it's not something miraculous.
2
5
463
Okay, taking in consideration recent news about Codex usage being shrank, I suspect it's already been down for a while.
And worse estimated limits are reflecting current real usage we already have. So I'd expect we wouldn't quite "feel" limits become lesser than now.
It's also entirely possible it would be framed as "see? We've made everything effective you don't even see the difference between today and yesterday"
Anyways, whatever they introduce on the DevDay I hope it's worth it. They're rising expectations even higher with this move.
Okay, I'm going to address an elephant in the room.
You can't just rebrand Terra into Sol, make it 2$ cheaper on output AND somehow made estimated limits for Sol 6 WORSE than what we were given with Terra😭
I'm convinced Sol isn't OAI's flagship and we're going to see another powerful model between Astra and Sol, but my man why are we getting less usage when Anthropic, somehow seems like giving more.
They even removed Terra from the list so no one can compare easily >_>
2
2
402
I makes ABSOLUTELY no sense. Either they believe it's far stronger than even Fable 5.5 or they just give up idk.
Just why? Do they think current 6th generation is enough? Bro you're mogged by Sonnet even. There's absolutely no point in using Sol 6 rn.
I'm getting more and more confused. The closer DevDay, the more bizarre news are coming from OAI's front.
Well, if you were getting excited for a new frontier model from OpenAI anytime soon, bad news.
GPT-6.1 Astra's planned October release has been scrapped
7
2
24
2,495
Even if OpenAI doesn't release a model that can beat Opus/Fable 5.5 at DevDay, it doesn't mean the war is lost.
Tbh I don't think either Anthropic or OpenAI really minds falling behind for a bit. A month or two and the lead changes again.
But does a short-term victory really matter if your overall goal is something much, much more capable?
9
629
This post feels like a filler, so take a look at one of my previous ones. I still hope for discussion
OpenAI, we need to talk... cuz Sol and Luna 6 have an issue.
AI is a tool, yes. But there's an important difference between classical software and AI. It's built to be communicated with. Like we talk to another person.
With normal software, you learn the interface and stuff. With AI, you explain what you want. So being bad at understanding people should be just as unforgivable as being bad at doing the work.
I mean, people aren't equally good at explaining themselves. Folks want to say 'make saas 200kk ARR make no mistakes' and get it built in return.
Models already produce work faster than people can comprehend it. Having to ALSO wrestle with them over what you meant makes the whole thing exhausting. And you can't afford to exhaust your customers.
That's why I think OpenAI lost on two fronts with this release. Opus 5.5 is more capable, and it's easier to communicate with. Sol being cheaper doesn't really fix either.
I want to enjoy building alongside AI. I already have plenty of places in everyday life where I can be misunderstood or get something far from my expectations.
2
770