Ask HN: Is this the fast take off?

Sam Altman said that if we had ChatGPT 5 like 5 years ago, we would have called it AGI,

Hold onto your butts, this time next year none of us will be working.

> Would you think humans and their error-prone systems had led to this by a series of accidents and greedy investments?

Any year before 2008, no.

On a more serious note though, I think you're zoomed in on the wrong pieces of the puzzle. (1) and (3) are just reflections of Nvidia's valuation, and (2) needs more time to be fully substantiated.

> How different would it look right now in November of 2025 if Sam had an AGI in pocket that was using all that money to grow itself?

I don't even know what an AGI is at this point. Who knows.

I don't think so. Primarily because if you can ask that question instead of just being dead, then it's not the fast takeoff.

On a less drastic note, if AI were autonomously cannibalizing the economy, you'd run into more things and go "huh I guess that's run by AI now" instead of "ah godammit why did they shove an LLM in this workflow".

We should not confuse action with result, not confuse the short term with the long term, and not confuse attention with impact. Predictions are not reality, hype is not interest, letters of intent are not money in the bank.

We've been here before, nothing has been proven yet, and the trough of disillusionment is lurking now too.

My opinion is that there will be no "AGI" any time soon. The recent model have shown two things: 1. a hard plateau where gains in one surface leads to losses in another and 2. what we have is cool and useful but extremely far from an autonomous AGI system.

This tech is not going anywhere at the time being (next 5 years). I am at the confidence where I'll bet on this. I also think unlocking the next step will require further theoretical and technical advances.

we don’t have AGI yet but we do have three things

1. AI that has huge demand and massive utility right now

2. Predictable improvements on the horizon for models in proven domains

3. Many heuristic that make near term AGI a reasonable bet to make, assuming progress continues as it has over the past few years

All these taken together makes what’s going on right now very clear and not surprising.

> Would you think humans and their error-prone systems had led to this by a series of accidents and greedy investments?

Yes, especially if we don't have (0) enormous productivity gains suddenly appearing out of nowhere.

This isn't fast take off, this is a stock bubble.

I believe we'll eventually have AGI, and it'll run on a PC at home, on CPU if you're willing to wait and have enough RAM. Training takes massive amounts of compute, but that can be distributed. I also believe that it's possible to significantly reduce the power required to run LLMs using alternative chip architectures.

Even if the bubble pops and no new data centers get filled with masses of GPUs, I think we're going to eventually have AGI now. Enough people have enough of a clue as to how to do it, not to mention the public sourced training data and models, to make it a certain outcome.