Listen up, you beautiful disaster of a readership. While I’m nursing my fourth bourbon of the evening, let me tell you about the latest circus act in our digital nightmare. The Information - usually a solid source when they’re not huffing unicorn farts - dropped a bombshell claiming AI progress is hitting a wall. Cute story. Real cute.
Here’s what’s got everyone’s panties in a twist: supposedly, OpenAI’s next big thing, Project Orion, isn’t the revolutionary leap forward we were promised. The improvements are “smaller” compared to the jump between GPT-3 and GPT-4. And the kicker? It might actually be worse at coding than its predecessor. Oh, the humanity.
But then our boy Sam Altman drops this cryptic tweet: “there is no wall.” That’s it. Three words. The digital equivalent of walking into a bar, ordering a shot, slamming it down, and walking out without explanation. Pure power move.
Now, here’s where it gets interesting. The same safety researchers who quit OpenAI because they thought the company was moving too fast - you know, the ones who should be throwing a parade if AI development actually slowed down - they’re saying this thing is still moving like a freight train with cut brakes.
Meanwhile, Gary Marcus - AI’s professional party pooper - is dancing in the streets like he just won the lottery. This guy’s been predicting AI’s demise longer than I’ve been drinking bourbon (and that’s saying something). He’s pointing to his 2022 blog post about deep learning hitting a wall like it’s some kind of prophetic text.
But here’s the thing about walls - they’re only obstacles if you can’t see what’s on the other side. Since Marcus wrote his doomsday prediction, we’ve had:
The real punchline? OpenAI isn’t even showing us their best cards. They’re like that card shark at the back of the bar who’s playing with a deck that makes Penn & Teller nervous.
Sam Altman’s latest cryptic message about the ARC benchmark - you know, that supposedly impossible-to-crack test of artificial general intelligence - reads like a bourbon-soaked fortune cookie: “in your heart, do you believe we’ve solved that one or no?”
Let me translate that from CEO-speak: “We’ve already cracked it, but we’re not going to tell you because watching everyone else try is too damn entertaining.”
The truth is, there’s no wall. There never was. What we’re seeing isn’t a slowdown - it’s the calm before the storm. It’s that moment in the bar right before someone says “hold my beer” and everything goes sideways.
So while everyone’s busy writing obituaries for AI progress, OpenAI is probably sitting on something that makes GPT-4 look like a pocket calculator. And the best part? They’re playing us all like a dive bar piano.
But what do I know? I’m just a guy who writes about tech while maintaining a blood-alcohol level that would make Ernest Hemingway proud.
Until next time, you beautiful disasters.
– Henry Chinaski (Written from my usual spot at the bar, where at least the bourbon is honest about its intentions)
P.S. - If anyone from OpenAI is reading this, I’ll trade you early access to Orion for my secret hangover cure. You know how to reach me.