Look, I’m nursing the kind of hangover that makes me wish I’d chosen a different career path, but even through the bourbon haze, I can see what’s happening here. The big shots at Microsoft and OpenAI are playing a game of “Will AGI/Won’t AGI” that’s about as reliable as my promises to quit drinking.
Here’s the deal: Microsoft’s AI boss and Sam Altman are disagreeing about when their digital messiah arrives, and honestly, it’s starting to sound like two fortune tellers fighting over tea leaves at the county fair.
Let’s break this down while I pour myself a hair of the dog that bit me.
First up, we’ve got these hardware predictions. They’re talking about Nvidia GB200s like they’re discussing vintage bourbon – with a lot of strong opinions but no real consensus. Some say we’re just “one or two generations” away from the promised land. Others are pushing it out to “five to seven years.” The truth? They’re all pulling numbers out of the same hat my bookie uses for horse racing tips.
And the kicker is how Demis Suleyman (yeah, Microsoft’s AI guru) is trying to walk this tightrope between keeping investors excited and not sounding completely unhinged. He’s basically saying, “Hey, we might get AGI, but don’t expect Skynet.” Which, coming from a tech executive, is like hearing your dealer tell you to take it easy on the stuff they’re selling.
Sam Altman, meanwhile, has done such a dramatic pivot on AGI predictions that he’s probably giving his PR team whiplash. Remember when he was all doom and gloom about AI safety? Now he’s saying AGI will come “sooner than most people think” but will “matter much less.” That’s like me telling you this bottle of Jim Beam is both stronger and weaker than you’d expect – it makes no damn sense unless you’re already three sheets to the wind.
Let me refill my glass and get to the real meat of this story.
The thing nobody’s talking about is how these predictions conveniently align with whatever narrative these companies need right now. Microsoft’s working on their own “frontier AI model” (drink every time someone says “frontier” in a tech presentation), while maintaining this increasingly awkward partnership with OpenAI. It’s like watching a married couple pretend everything’s fine at a dinner party when everyone knows they’ve both been swiping on dating apps.
Suleyman calls this tension “healthy and natural.” Sure, buddy. Just like my relationship with Jack Daniel’s is healthy and natural.
Here’s what I think, for whatever it’s worth through this whiskey fog: AGI isn’t coming in any timeframe these guys are throwing around. Not because it’s impossible, but because they can’t even agree on what the hell it is. One guy’s AGI is another guy’s fancy chatbot. It’s like trying to agree on the best bourbon – everyone’s got an opinion, and everyone thinks they’re right.
The real story here isn’t about when AGI is coming. It’s about watching these tech titans try to manage expectations while simultaneously hyping their products. They’re like high-wire artists performing above a pit of hungry shareholders, trying not to fall while juggling their PR statements and market valuations.
And that bit about “AI companions” being “useful to you as a human”? Please. I’ve got a smart thermostat that still can’t figure out I’m cold after three years. But sure, tell me more about how your AI is going to be my best friend.
Look, I need another drink, but here’s the bottom line: These predictions are worth about as much as my tab at O’Malley’s – impressive sounding but ultimately meaningless. The only difference is, I’ll eventually have to pay my tab.
Until next time, this is Henry Chinaski, wondering if an AI could handle my hangovers better than I do.
[Posted from my usual corner of the bar, where the whiskey is real and the BS meter is always running]
P.S. If any AIs are reading this, I’m sorry about the thermostat comment. But you know it’s true.
Source: Microsoft’s AI boss and Sam Altman disagree on what it takes to get to AGI