I keep hearing these wildly divergent takes on AI that have no relationship to observable reality.
-
I keep hearing these wildly divergent takes on AI that have no relationship to observable reality.
On one side, you’ve got people insisting AI will replace all human creativity. On the other, you’ve got people claiming AI is useless and barely qualifies as a calculator. Both camps are out to lunch.
AI still needs a human prompt, which is—brace yourself—a form of human creativity. And as a gamer, I’ve had AI opponents in FPS games since the 90s. This isn’t new magic.
So what’s going on? Nothing mysterious. It’s the Gartner hype cycle doing exactly what it’s done since 1995. AI is following the script with uncanny precision.
Right now we’re sliding from the Peak of Inflated Expectations into the Trough of Disillusion. You can see the split in real time.
The laggards still insist AI is the new Industrial Revolution. The culture at large is already exhausted and disillusioned. And the vanguard—the few paying attention—are starting to understand what AI might actually be good for.
In short: AI is never replacing all knowledge work. Not because it’s incapable, but because it’s unaccountable. As IBM put it back in 1979, “A computer can never be held accountable therefore a computer must never make a management decision.” If a human needs to be accountable for an outcome, the human is staying in the loop.
But AI can absolutely do other work. DLSS lets AI generate frames between frames to boost FPS. AI can auto-transcribe audio. It can translate languages in real time. These are all useful and practical tasks.
But who’s accountable for the output? Not AI. It can’t be.
And that’s the industry’s real problem. Everyone is building output frameworks. Almost nobody is building accountability frameworks. Without those, AI is—putting it extremely mildly—guaranteed to disappoint.
We’re watching that disappointment play out right now. It’s not a failure of the tech. It’s just the hype cycle doing what the hype cycle does. Personal computers went through it. Video games went through it. The Internet went through it.
AI is no different.
-
I keep hearing these wildly divergent takes on AI that have no relationship to observable reality.
On one side, you’ve got people insisting AI will replace all human creativity. On the other, you’ve got people claiming AI is useless and barely qualifies as a calculator. Both camps are out to lunch.
AI still needs a human prompt, which is—brace yourself—a form of human creativity. And as a gamer, I’ve had AI opponents in FPS games since the 90s. This isn’t new magic.
So what’s going on? Nothing mysterious. It’s the Gartner hype cycle doing exactly what it’s done since 1995. AI is following the script with uncanny precision.
Right now we’re sliding from the Peak of Inflated Expectations into the Trough of Disillusion. You can see the split in real time.
The laggards still insist AI is the new Industrial Revolution. The culture at large is already exhausted and disillusioned. And the vanguard—the few paying attention—are starting to understand what AI might actually be good for.
In short: AI is never replacing all knowledge work. Not because it’s incapable, but because it’s unaccountable. As IBM put it back in 1979, “A computer can never be held accountable therefore a computer must never make a management decision.” If a human needs to be accountable for an outcome, the human is staying in the loop.
But AI can absolutely do other work. DLSS lets AI generate frames between frames to boost FPS. AI can auto-transcribe audio. It can translate languages in real time. These are all useful and practical tasks.
But who’s accountable for the output? Not AI. It can’t be.
And that’s the industry’s real problem. Everyone is building output frameworks. Almost nobody is building accountability frameworks. Without those, AI is—putting it extremely mildly—guaranteed to disappoint.
We’re watching that disappointment play out right now. It’s not a failure of the tech. It’s just the hype cycle doing what the hype cycle does. Personal computers went through it. Video games went through it. The Internet went through it.
AI is no different.
@atomicpoet one of the problems in the general discussion is the term "AI" in itself.
It throws a lot of different tech in the same bucket. How does one discuss based on such an inaccurate term?
In this bucket is proven tech that is being used for ages (bots, DLSS,classification / pattern recognition) and new tech like LLMs.
And LLMs only get that much attention and money because they mimic human speech so well that many people confuse this with real understanding and thinking.
-
@atomicpoet one of the problems in the general discussion is the term "AI" in itself.
It throws a lot of different tech in the same bucket. How does one discuss based on such an inaccurate term?
In this bucket is proven tech that is being used for ages (bots, DLSS,classification / pattern recognition) and new tech like LLMs.
And LLMs only get that much attention and money because they mimic human speech so well that many people confuse this with real understanding and thinking.
Whatisgoingon So true!
