Y'all never heard of and it shows.

They'd have been in a much better place if they'd called *this* GPT-4.5.

Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.

It was only this year that I started believing Gary Marcus, but I'm glad I jumped on the train before the other trains hit walls.

My GPT-5 has the same "personality" that my GPT-4o did. I don't understand why it's different for anyone else. Just tell it to act like it did. It remembers.

Ha ha, losers! I still have access to 4o! How d'you like that nostalgia?. Who needs GPT-5 on mobile? Not me, apparently!.

If they'd called *this* GPT-4.5, I think nearly everyone would be raving about it.

6 seconds. ? For that?. Yeah, uh oh.

Uh oh.

Plus user: . Still not seeing the new agentic capabilities in ChatGPT. Thought that was releasing Monday. ?.

If GPT-5 isn't an amazing leap forward, I think I'll fully side with the people who say we're basically reaching the limit of LLMs in general. (Nothing has wow'd me since 4o's image generation, and that's more of a fun side feature.).

xAI said just: "We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts.". Why would they release the "MechaHitler" Grok on the world without doing some in-house testing? How could this have been a surprise?.

The transformation to Musk's ideology may have already happened. Before, I don't believe Grok would have taken a strong position here. Grok has also been thinking quite a long time about the second question.

Gemini Pro (*Pro* mind you) still gets easy sports questions wrong. For instance, I asked when Man City's upcoming game is, and it insists it's on Sunday; it's Monday. It also said their actual opponents didn't qualify. That's. that's bad, right? It's Google.

All the major LLMs telling me obviously false things about the US bombing Iran, making up a variety of false sources. And I'm talking about the better, more advanced reasoning models. Very disappointed. One said President Biden condemned it from the White House. FFS.

The number of people who initially embraced Grok - then were embarrassed by its truthful responses and call it "liberal" now - and now hate it is hilarious.

Friendly-ass reminder to treat every single new image you see of a protest or a warzone as AI generated until you can confirm otherwise. I love AI in general, but this is certainly a huge downside. People will waste time at best, and buy complete bullshit at worst.

An unfortunate thumbnail crop, to be sure.

The Gentle Singularity We'll see. As for AGI (let alone ASI), I'll believe it when I see it. Whatever "it" is, exactly. In any case, I enjoy a hopeful vision of humanity's near future. I only worry about the disruption in the meantime.

You guys should probably hold off on commenting on the day's moves until after the closing bell.

Highly recommend trying this prompt out. I've done this several times in different chats, and the style is never radically different, and a horned animal has showed up more than once.