Live data from Hacker News

After the AI boom: what might we be left with?

blog.robbowley.net

411–420 of 489 posts

Re: After the AI boom: what might we be left with?

#411

Earlier quoted context omitted.

I feel like on the one hand you're exactly right that this is kind of like a cold war level investment... and it's not necessarily being pursued through a business return lens; something deeper and more visceral is driving it. However your conclusions are what throw me off.. you kind of have this Doom and Gloom mindset which may be fair but I don't really think it's related to this particular bubble.. in other words…

That would be my take under different circumstances as well, but there are two key differences in this situation: 1. The debt bomb. Not dealing with this could cause a great depression by itself. Having it go off at the same time as we're underwater on bad capex with a misaligned economy could produce something the likes of which we've never seen. 2. We have an authoritarian president that has effectively captured th…

Are you sure the debt bomb is a bomb? Private debt can be a major concern, but government debt may not work the way you think it does. In fact, the great depression was preceded (and possibly in some ways triggered) by a significant paying down of public debt, counterbalanced by an increase in private debt that kept the money supply growing through the roaring 20s.

Re: After the AI boom: what might we be left with?

#412

Earlier quoted context omitted.

Inference is not that expensive. I'd argue that most models are already useful enough that people will pay to run them.

At $20/month for Claude, I'm satisfied. I'll keep paying that for what I get from it, even if it never improves again.

Of course, but my point is that I don't think it's economically sustainable. If innovation/funding in AI stalls, those $20 will likely skyrocket fast.

Re: After the AI boom: what might we be left with?

#413

Earlier quoted context omitted.

I'm actually from a science, government and defense background, I've never worked in the valley. It might be possible that it can't succeed, but we don't know that and the evidence in terms of what happens if it fails is pretty compelling, so I think morality dictates we have to at least try, at least until we have more information or another path emerges.

Morality would dictate that we stop spending enough money to end hunger in not just our country but the entire planet on this stupid venture. We are gutting our future entirely on our own and hoping that AI will save us. That’s moral suicide, not an imperative.

I didn't argue against that at any point, but you have to understand the political realities we're operating within.

Re: After the AI boom: what might we be left with?

#414

Earlier quoted context omitted.

I'm actually from a science, government and defense background, I've never worked in the valley. It might be possible that it can't succeed, but we don't know that and the evidence in terms of what happens if it fails is pretty compelling, so I think morality dictates we have to at least try, at least until we have more information or another path emerges.

You're making the (imo incorrect) assumption that a revolutionary moment isn't the exact thing the US needs right now. Morality dictates we have to try and force one.

We need a change, but I don't think we have the unity and will to force a revolution right now. I think our best chance is a massive win in the 26 elections, but barring that I think we're struck bailing out the oligarchs for now, and striking back with collaborative economic warfare.

Re: After the AI boom: what might we be left with?

#415

Earlier quoted context omitted.

It's a US/China centered article, because that's the game, Europe is not a meaningful player and everyone else is going to get sucked into orbit of one of the superpowers. If you read the article carefully, I work hard to keep my priors and the priors of the people in question separate, as their actions may be rational under their priors, but irrational under other priors, and I feel it's worth understanding that nua…

I read your article. My cynical take is that this is the US committing economic suicide, based on a misguided belief in something that'll never happen. The new superpowers will be the EU, which was smart enough not to make the same gamble, and China, which will structurally survive it. I also disagree with your conclusion of a moral imperative to make sure that AI succeeds. I believe it's the opposite. AI failing wou…

I don't see odds on a good outcome from a revolution. Keep in mind which faction in the united states is generally militant. The best possible scenario there is broad civilian unrest that the administration tries to forcefully quell, triggering a military coup, but it's unlikely that the coup would be unified, and right wing militias and hardcore trump supports would go down fighting.

We need political Aikido to hold this country together.

Re: After the AI boom: what might we be left with?

#416
post #217
post #96

Earlier quoted context omitted.

> Speech to text and vice versa exists for over a decade. Indeed. I was using speech to text three decades ago. Dragon Naturally Speaking was released in the 90s.

Then you hopefully remember how “natural” it actually was.

The old voice bank for the redhead with drills had more soul.

Re: After the AI boom: what might we be left with?

#417

Earlier quoted context omitted.

Honestly I've never wanted to be more wrong in my life. There were so many moments when I was doing research and analysis for this article that I said to myself: "We're so fucked"

And you didn't even touch on the "if anyone builds it, everyone dies" aspect of super intelligence.

Honestly, I think more likely scenario is that the AI that is built to enslave us is accidentally over-aligned and sets up a digital coup.

Re: After the AI boom: what might we be left with?

#418

Earlier quoted context omitted.

This is written by somebody who’s never been to Louisiana and Mississippi. The saying in Russian that they never had it any good in the first place comes to mind. Where’s thriftwy when you need them?

The south is pretty modern all things considered. You wanna see real pain? Go ask Chinese people who had to sacrifice these past few decades to make China the powerhouse it is today. Its the reason why they can tolerate anything the US throws at them. Their pain threshold is much higher than whatever the west has. These southerners could have lost their culture, the local industry has long since left, and fentanyl is…

I'm a southerner, and I road trip a lot. Driving through parts of the southeast you will pass through a sea of dilapidated trailers with yards full of junk, run down chain restaurants and industrial lots. It's not India or the Philippines, but it's still pretty third world in a lot of ways.

Re: After the AI boom: what might we be left with?

#419

Earlier quoted context omitted.

That would be my take under different circumstances as well, but there are two key differences in this situation: 1. The debt bomb. Not dealing with this could cause a great depression by itself. Having it go off at the same time as we're underwater on bad capex with a misaligned economy could produce something the likes of which we've never seen. 2. We have an authoritarian president that has effectively captured th…

Are you sure the debt bomb is a bomb? Private debt can be a major concern, but government debt may not work the way you think it does. In fact, the great depression was preceded (and possibly in some ways triggered) by a significant paying down of public debt, counterbalanced by an increase in private debt that kept the money supply growing through the roaring 20s.

I understand the Keynesian dynamics of maintaining money velocity through government spending, but it's a balancing act, and I'm pretty sure we're way off balance now.

Re: After the AI boom: what might we be left with?

#420
post #85

> GPUs that have a 1-3 year lifespan In 10 years GPUs will have a lifespan for 5-7 years. The rate of improvement on this front has been slowing down faster then CPU.

There was a video on YT (gamers nexus I think) and an excel sheet comparing jumps in performance gains between each new nvidia generation. They are becoming smaller and smaller, probably now driven by AI boom where most of the silicon is used for data centers. Regardless of that, I have a feeling we are approaching the ceiling of chip performance. Just comparing PS3 and PS4 and then PS4 with PS5, performance jump is…

Most of the performance gains also came from smaller datatypes such as fp4. Going from bfloat16/fp16 to fp8 is easy. fp4 is challenging but possible. fp2 can't exist independently, because it is just the sign bit plus a single bit.

The next frontier would be training directly with block floating point, where you have a shared exponent plus the two remaining bits. It's getting tight.

Maybe it is possible to have mini LoRA blocks where an n times n block is approximated by the outer product of two n sized vectors. For n = 4 the savings would be 50% less FLOPs and for n=8 the savings would be 75% less FLOPs.

Post reply on HN