RETURN TO CHANNEL

Search Captions & Ask AI

"This is the first time that OpenAI has released a new model that was not decisively the best."

August 12, 2025 / 00:59

This episode discusses the release of GPT-5 by OpenAI, the expectations surrounding it, and its perceived underwhelming performance. Key topics include the potential for diminishing returns in large language models and the initial issues with the model's router.

Sammy the Bull Altman announced GPT-5, which did not meet the high expectations set by a prior tweet. The conversation raises concerns about whether the advancements in LLM technology are becoming incremental rather than groundbreaking.

There is a mention of OpenAI's router malfunctioning for the first 12 to 16 hours after GPT-5's release, impacting user experience. Despite this, the hosts acknowledge OpenAI's strong product quality and user base.

The discussion also references Eric Fishria's statement about OpenAI's value, suggesting that even without future model releases, the company remains significant due to its user engagement and product excellence.

TLDR

GPT-5's release was underwhelming, raising concerns about diminishing returns in LLM advancements.

Episode

0:59

00:00:00

Sammy the Bull Altman dropped GPT5. It was a bit underwhelming. Sam did this tweet. It kind of built a lot of

00:00:05

expectation that this would be like otherworldly. It wasn't. Are we hitting like either the trrow of despair or

00:00:11

maybe diminishing returns in the LLM space where maybe it's feeling incremental, not groundbreaking when we

00:00:18

release these iterations. Maybe for OpenAI, this is the first time that, you know, OpenAI has released a

00:00:24

new model that was not decisively the best. In all fairness to OpenAI, the the router that kind of auto selects the

00:00:30

right model was apparently broken for the first kind of 12 or 16 hours of the model's release. It wasn't working.

00:00:36

Yeah. It was not working consistently for all users. I mean, the one thing OpenAI is like they are a very good

00:00:42

product. Just really, really good. And I just always go back to the statement from Eric Fishria. You know, it might be

00:00:48

that if OpenAI never releases another model, they're still a really valuable company because they have so many users

00:00:55

and they're so good at product and they do have a lot of distribution.

Episode Highlights

OpenAI's Model Struggles
The router for model selection was broken for hours after release.
@ 00m 29s
August 12, 2025
OpenAI's Value Beyond Models
Even without new models, OpenAI remains a valuable company due to its user base.
“If OpenAI never releases another model, they're still valuable.”
@ 00m 46s
August 12, 2025

Episode Quotes

Are we hitting the trough of despair?
"This is the first time that OpenAI has released a new model that was not decisively the best."
OpenAI is a very good product.
"This is the first time that OpenAI has released a new model that was not decisively the best."
If OpenAI never releases another model, they're still valuable.
"This is the first time that OpenAI has released a new model that was not decisively the best."

Key Moments

Trough of Despair00:09
Router Issues00:29
Valuable Company00:46