91–100 of 218 posts
This article is pretty confusing, doesn't really have a thesis, just listing some stats. Maybe that's the intent.
Certainly it suggests that “this time is different” without saying it in a quotable fashion.
The metrics it provides seem useful. What are the metrics it is missing?
I worked at a mom and pop ISP in the 90s. Lucent did seem at the forefront of internet equipment at the time. We used Portmaster 3s to handle dial up connections. We also looked into very early wireless technology from Lucent. Something I wanted to mention, only somewhat tanget. The Telecommunications Act of 1996 forced telecommunication companies to lease out their infrastructure. It massively reduced the prices an…
You're implying only 4 years of regulation was enough to shift the balance of power between telecoms and smaller ISPs." If it's true that this regulation was what helped jumpstart the internet it's an interesting counterpoint to the apocalyptic predictions of people when these regulations are undone. (net neutrality comes to mind as well) I've never heard anyone claim before that just having these laws on the books f…
Why would it be enough? This legislation prevents monopolies from abusing position, therefore we will repeal it the moment it turns out to be useful?
Yeah, it takes time to consolidate power again, that does not mean the legislation is not good.
Fiber networks were using less than 0.002% of available capacity, with potential for 60,000x speed increases. It was just too early. I doubt we will see unused GPU capacity. As soon as we can prompt "Think about the codebase over night. Try different ways to refactor it. Tomorrow, show me your best solution." we will want as much GPU time at the current rate as possible. If a minute of GPU usage is currently $0.10, a…
Earlier quoted context omitted.
> I doubt we will see unused GPU capacity I'd argue we very certainly will. Companies are gobbling up GPUs like there's no tomorrow, assuming demand will remain stable and continue growing indefinitely. Meanwhile LLM fatigue has started to set in, models are getting smaller and smaller and consumer hardware is getting better and better. There's no way this won't end up with a lot of idle GPUs.
>Meanwhile LLM fatigue has started to set in Has it? I think there is this compulsion to think that LLMs are made for senior devs, and if devs are getting wary of LLMs, the experiment is over. I'm not a programmer, my day job isn't tech, and the only people I know who express discontent with LLMs are a few of programmer friends I have. Which I get, but everyone else is using them gleefully for all manner of stuff. An…
I think it's important to remember that a good bunch of this is going to be people using it as an artificial friend, which is not really productive. Really that's destructive, because in that time you could be creating a relationship with an actual person instead of a context soon to be deleted.
But on the other hand, some people are using it as an artificial smart friend, asking it questions that they would be embarrassed to ask to other people, and learning. That could be a very good thing, but it's only as good as the people who own and tune the LLMs are. Sadly, they seem to be a bunch of oligarchs who are half sociopaths and half holy warriors.
As for compute, people using it as an artificial friend are either going to have a low price ceiling, or in an even worse case scenario they are not and it's going to be like gambling addiction.
Earlier quoted context omitted.
Not sure why you're getting downvoted. If you speak with AI researchers, they all seem reasonable in their expectations. ... but I work with non-technical business people across industries and their expectations are NOT reasonable. They expect ChatGPT to do their entire job for $20/month and hire, plan, budget accordingly. 12 months later, when things don't work out, their response to AI goes to the other end of the…
> their expectations are NOT reasonable. They expect ChatGPT to do their entire job for $20/month and hire, plan, budget accordingly. This is entirely on the AI companies and their boosters. Sam Altman literally says gpt 5 is "like having a team of PhD-level experts in your pocket." All the commercials sell this fantasy.
I think we are at the PS3/Xbox 360 phase of AI. By that I mean, those were the last consoles where performance improvements delivered truly new experiences, where the hardware mattered. Today, any game you make for a modern system is a game you could have made for the PS3/Xbox 360 or perhaps something slightly more powerful. Certainly there have been experiences that use new capabilities that you can’t literally put…
But the way how it stayed niche shows how it's not just about new gameplay experiences.
Compare with the success of the Wii Sports and Wii Fit, which I would guess managed it better, though through a different kind of hardware that you are thinking about ?
And I kind of expect the next Nintendo console to have a popular AR glasses option, which also would only have been made possible thanks to improving hardware (of both kinds).
I get the demand for new applications, which require inference, but nowadays with so many good (if not close to SOTA) models available for free and the ability to run them on consumer hardware (apple M4 or AMD Max APUs), is there any demand for applications that justify a crazy amount of investment in GPUs?
Knowing history of past bubbles is only mildly informative. The dotcom bubble was different than the railroads bubble etc. The only thing to keep in mind is that all of this is about business and ROI. Given the colossal investments, even if the companies finances are healthy and not fraudulent, the economic returns have to be unprecedented or there will be a crash. They are all chasing a golden goose.
This time they have fiat money and government on their side. so that is also different.
Knowing history of past bubbles is only mildly informative. The dotcom bubble was different than the railroads bubble etc. The only thing to keep in mind is that all of this is about business and ROI. Given the colossal investments, even if the companies finances are healthy and not fraudulent, the economic returns have to be unprecedented or there will be a crash. They are all chasing a golden goose.
In reference 14 we read > However, what’s become clear is that OpenAI plans to pay for Nvidia’s graphics processing units (GPUs) through lease arrangements, rather than upfront purchases I wish someone here could explain it to a dummy like me. Nvidia tells OpenAI: heres some GPUs, can you pay for them over 5 years. How is this an "investment" by Nvidia? That reference keeps calling this an investment, but what they d…
That Nvidia has to front the costs of the product at the beginning and arguably the risk that the allocation of assets end up not being paid off (bankruptcy, etc.) By carrying those costs early and the associated risk, Nvidia expects a return on that. If the risk is realized they'll lose but otherwise they'll gain. That has all the hallmarks of an investment.
NVidia could protect itself against OpenAI bankrupcy by adding a clause to the lease saying that if OpenAI goes bankrupt, Nvidia gets its GPUs back. So the risk would only be that the lease would be aborted sooner than expected.