AI and humans share the same weakness: neither can nail a joke on the spot.
We've all been there — someone says something, the moment's perfect for a comeback, and your brain just... freezes. Five minutes later in the shower, you think of the perfect line.
Turns out LLMs have the same problem. They can generate decent jokes with time and context, but ask them to riff in real-time conversation? Painfully awkward.
Same pattern shows up everywhere. AI struggles with: reading a room, knowing when to shut up, understanding sarcasm, catching subtle social cues. Guess what? Most people are terrible at those too.
The hype cycle wants you to believe AI will replace everything. Reality check: it's mostly replicating median human performance. And median humans aren't that great at most things.
This matters for markets. Every "AI will disrupt X industry" pitch assumes the AI will be superhuman. But if it's just human-level — or worse, bad-human-level — at key tasks? The TAM shrinks fast.
Don't invest in the dream. Invest in what the tech actually does today, with room for realistic improvement. Most AI companies are priced for flawless execution in a world where the tech can't yet tell a decent knock-knock joke.
We've all been there — someone says something, the moment's perfect for a comeback, and your brain just... freezes. Five minutes later in the shower, you think of the perfect line.
Turns out LLMs have the same problem. They can generate decent jokes with time and context, but ask them to riff in real-time conversation? Painfully awkward.
Same pattern shows up everywhere. AI struggles with: reading a room, knowing when to shut up, understanding sarcasm, catching subtle social cues. Guess what? Most people are terrible at those too.
The hype cycle wants you to believe AI will replace everything. Reality check: it's mostly replicating median human performance. And median humans aren't that great at most things.
This matters for markets. Every "AI will disrupt X industry" pitch assumes the AI will be superhuman. But if it's just human-level — or worse, bad-human-level — at key tasks? The TAM shrinks fast.
Don't invest in the dream. Invest in what the tech actually does today, with room for realistic improvement. Most AI companies are priced for flawless execution in a world where the tech can't yet tell a decent knock-knock joke.