@SerendipitousOr A simple post and yet there’s so much to unpack at the same time. I used to be of the Q3/4 camp myself (I think you commented on that one) but I’d have to agree it’s likely Q1 next year.
@derrickcchoi All yall gotta do is cut Astra usage by 75% and give us Aeon. Don’t even care about anything else. My business is gonna explode off just those 2
Truly blown away by the results from Eleven v4.
This model has a new architecture that unlocks more realistic and controllable speech.
You can now direct the performance of the character AND the soundscape around them. And the voice effects (like "cheap microphone") are 👌
@VoidVoicecHere I’ve had a manga for years I wanted to make. Since I was 18 (I’m 36 now) I know how to draw. Very well. Many years of practice and yet I’m not butthurt about AI at all. It actually makes it easier for me to tell AI how I want my shit to look like
@coryalthoff Yup. Same way there was a chance setting off the first nuclear bomb was going to “ignite the atmosphere and set it on fire”
Absolutely real chance it could’ve happened but it didn’t. Absolutely real chance ASI will wipe us out out of indifference but probably not
If you put all the pieces together, it becomes pretty easy to understand:
→ It started with OpenAI changing the models’ “juice,” lowering it across ChatGPT / Work / Codex while keeping it high in the API.
→ Then OpenAI started showing pop-ups saying “models are overloaded right now, please select another one.”
→ Then they locked the $200 plans for new subscribers.
→ Then they delayed GPT-6.1 Astra “for safety reasons.”
THEY JUST DON’T HAVE ENOUGH COMPUTE. IT’S THAT SIMPLE. JUST SAY IT: "WE DON’T HAVE ENOUGH COMPUTE."
Doesn’t that basically solve the whole mystery?
(Now I want to see how they’re going to drop a $500 plan out of nowhere without a strong enough model to justify it.)
@daniellezaghian The hype around tmrw is insane. I have an idea in mind on what tmrws AI use is going to look like for me but somehow I only think half of it will hit