For 3 years, my SF/AI elite friends swore we're 1-2 years from AGI/ASI singularity. Everything changes, they said.
My gut? We're nowhere close. This terminator is still a dot on the horizon.
Here's the thing: I trust my intuition—it's my edge. But I'm not arrogant. These people are smarter than me in many ways. When we clash, I listen.
Now their timelines passed. Turns out we're both right?
Yes, something revolutionary is happening. Software is eating the world faster than ever. But these tools still need us. They're not replacing the tool holder. Not yet. Not even close.
Ilya says data is the constraint. That feels right. There's only so much non-linear intelligence to train on. Humans are still innovating beyond what's scrapable online. I'd like to think I'm one of them.
Yann LeCun says LLMs are capped. Maybe. But combined with world models? Unclear.
Next breakthrough might be synthetic data—like quant finance has done for decades to train alpha models.
This is all a pursuit for more alpha.
What's unnerving: Even though I've been right about timelines, I can't disprove the next corner theory. We keep moving one block at a time, breath held, knowing something's coming but unsure when.
So let's breathe. There's trauma under the surface from this pace of change. We all feel it. Fascia pulled tight, emotionally masked.
But those of us with awareness and lights on? We're not getting replaced anytime soon—I don't care what Elon, Dario, or Leopold say.
Stay sharp. The game's still ours to play.
My gut? We're nowhere close. This terminator is still a dot on the horizon.
Here's the thing: I trust my intuition—it's my edge. But I'm not arrogant. These people are smarter than me in many ways. When we clash, I listen.
Now their timelines passed. Turns out we're both right?
Yes, something revolutionary is happening. Software is eating the world faster than ever. But these tools still need us. They're not replacing the tool holder. Not yet. Not even close.
Ilya says data is the constraint. That feels right. There's only so much non-linear intelligence to train on. Humans are still innovating beyond what's scrapable online. I'd like to think I'm one of them.
Yann LeCun says LLMs are capped. Maybe. But combined with world models? Unclear.
Next breakthrough might be synthetic data—like quant finance has done for decades to train alpha models.
This is all a pursuit for more alpha.
What's unnerving: Even though I've been right about timelines, I can't disprove the next corner theory. We keep moving one block at a time, breath held, knowing something's coming but unsure when.
So let's breathe. There's trauma under the surface from this pace of change. We all feel it. Fascia pulled tight, emotionally masked.
But those of us with awareness and lights on? We're not getting replaced anytime soon—I don't care what Elon, Dario, or Leopold say.
Stay sharp. The game's still ours to play.