Hacker News Reader

Comment by dev1ycan | Hacker News Reader

I don't understand the idiocracy we live in, it is beyond obvious not just that the stock market is a bubble but ESPECIALLY the AI related stocks are a massive bubble, when it pops, and it will, it is going to be very very ugly, yet people keep pouring in, as Sabine said it, it's starting to look like particle physics where they keep asking for bigger colliders, just because you have a bigger collider, if your methodology is flawed you aren't gonna get any more significant returns.

Eventually they will run out of exponential cash to pour in, and investors will start asking questions, stocks are already valued at 60x+ their earnings, whenever it pops you don't want to be the one who bought the top.

Guess it's still gonna take a while more for the layman to realize the issues with LLMs, but it'll happen.

View on HN

>if your methodology is flawed you aren't gonna get any more significant returns.

The problem with this statement is that predictions made about scaling 5 years ago have held true[1]. We keep adding parameters, adding compute, and the models keep getting more capable.

The flaws of LLM's from 2024 are not what is relevant. Just like the flaws of LLMs from 2021 were not relevant. What is relevant is the rate of change, and the lack of evidence that things won't continue on this steep incline. Especially if you consider that GPT4 was sort of a preview model that motivated big money to make ungodly investments to see how far we can push this. Those models will start to show up over the next 2 years.

If they break the trend and the scaling flops, then I think a lot of air is gonna blow out of the bubble.

[1]https://arxiv.org/pdf/2001.08361

we added a lot of parameters.

We added a LOT of data.

The resulting models have become only slightly better. And they still have all of their old problems.

I think this is proof that scaling doesn't work. It's not like we just doubled the sizes, they increased by a lot, but improvements are less and less each time. And they've already run out of useful data.

They are very literally asking for trillions and even nuclear powered data centers, pretty sure we've gotten to the point where it's not sustainable.

Those are roadmap items being asked for, but the next gen models are already in training. If they keep moving along the same trend line, like all the previous models have, then they probably will be able to find the investors for the next next gen. Even if it's a few trillion dollars and a few nuclear power plants.

This doesn't even factor in the tech inertia. We could stop making new models today, and it would probably be 4-5 years before integration slowed down. Google still hasn't even put Gemini in their home speakers.

Exactly. I was assuming that the by now the default answer to "LLMs sort-of do this, but not very well" should be "OK, wait a few months".

Computers have been able to do mathematical calculation and logical deduction cheaply and perfectly for decades, and it's not really required for generative AIs to be able to do it for them to be useful. It's good enough if they can write and execute some python code to do it, and generally they are fairly capable of that.

The question of whether they can do it is interesting in an academic sense, but has nothing to do if they're useful or not. They also don't need to be true AGI to be useful.

Reply on news.ycombinator.com