You'll still get a few years before the software becomes remotely disabled, though. This story about Office 2019 losing functionality follows Office 2019 losing support in 2023. If that's the rate things go, then maybe Office 2021 will lose functionality either 2 years from now (7 years after release) or 3 years from now (3 years after losing support).
And, as I understand it, Anthropic hasn't committed as much spending to building out new data centers, and has setup their operations to be GPU agnostic, so they can keep flexibility between NVIDIA GPUs, Google TPUs, and Amazon Trainium, and play the data center pricing game. Anthropic is better positioned to survive an AI winter (and I believe it's coming soon).
The economics of it don't add up and the growth rate of the curve of improvement over time has already significativelly fallen which looking at the historical curves for other technologies is a very strong indication that it's approaching the limits of how far it will go even though it's nowhere close to the hype.
Yeah, I'm convinced that they've maintained the illusion of continued exponential improvement from 2024-2026 by sneaking in exponential increase in resources (hardware complexity, power consumption), to prop things up past what should have been a plateau.
AI has an interesting economic trait in that it's very, very expensive to deploy, and made very fast progress from 2022 to 2024. That caused investors with money to believe that:
Pushing the frontier was going to cost a lot of money. More than any other purported revolutionary tech.
Extrapolation of past improvement meant that whoever was on the cutting edge may end up with a product with a huge paying market.
So whoever wins this race would be rich, and the investment would have been worth it for them.
But since 2024, we've seen that the cutting edge got even more expensive much faster than expected, and much of the improvements in performance now come from inference rather than training, which represents a high ongoing cost.
Now, if we extrapolate from that trend line, we'll see that the market will be much smaller for AI services at the cost it takes to provide that service, and the question then becomes whether the industry can make its operations cheaper, fast enough to profitably provide a service people will pay for.
I have my doubts they'll succeed, and we might just be looking at the industry like supersonic flight: conceptually interesting, technically feasible, but just a commercial dead end because it's too expensive.
ActivityPub is the fediverse protocol, lemmy and piefed are software implementations of that protocol.
In a similar way, email is a protocol, and Gmail and Exchange/Outlook are software implementations of that protocol. You can use Gmail to send email to Outlook users. Different people can administer their own Exchange servers on their own domains.
And some features of the software work best with other people who use the same version of the same software, although most things kinda work between different software. Like how calendar invites sometimes act weird between users of different software, but for the most part the core functions work OK.
So lemmy and piefed (and mbin and sublinks and some others) are different software trying to speak to other fediverse services through the ActivityPub protocol. It mostly works, but some of the details don't work exactly the same between each type of software.
One real concern I have is that there are now automated tools that can read a patch, and the maintainer's release notes with a description of a security vulnerability fixed by that patch, and then create a working exploit of the pre-patch vulnerability.
In that particular moment, you know that a vulnerability exists and that it was serious enough to be described in release notes, and you can compare two code versions, one that is secure and one that is not. From there, any AI coding agent is working towards something that definitely exists, with a bunch of description of what it might be.
So that means that the window between when a patch is released and when users actually apply that patch is going to be more important than ever. Downstream maintainers will be under a lot of time pressure to implement changes from upstream, because every new security patch will create a race to create 1-day exploits for everyone using that software.
The ones that power spacecraft generate less than 5000W of heat at max power (while producing 300W of usable electricity).
In order to power a single server rack of 72 Blackwell GPUs, which takes about 130,000 watts, you'd need about 430 of those RTGs, and need to manage cooling requirements of 430 times as much (plus however much additional power will be required by the cooling system itself, too).
Companies are building entire workflows around AI, but they are building them under the assumption that they won't ever be charged per token.
Or worse, where the AI models underpinning a workflow breaks or degrades in some way to reduce token usage and then starts behaving in unexpected ways, in a process/workflow that assumes a particular type of behavior.
Counterpoint: sometimes the best still shot requires a particular moment captured with a particular, consciously arranged setup.
This interview of a veteran NBA photographer breaks it down of how he only has a single shot per shot because of how he necessarily relies on strobes set up to not distract the players or interfere with the broadcast. As a result, he scouts/studies each player and team so that he knows when the right moment is to actually capture the shot, because he can't exactly ask players to do it again.
If you read interviews of Pulitzer photography winners, they'll often say a lot of the same things: being prepared and being lucky and having that convergence of having incredibly high skill/expertise/understanding of the setting, while being able to capture in every opportunity presented.
You should capture a lot of photos and examine them to understand how to make them better, and increase your skill level and understand your subject so that you can still optimize for the very best shot possible.
Original reporting by Bloomberg is here or here for an archive.is version.
Sounds like the talks are stalled about revenue guarantees from the Kenyan government, if demand for the data center's capacity never shows up. The power infrastructure isn't really an issue yet, since the first phase is going to be 100MW and there are plans to build geothermal plants sufficient to cover several multiples of the country's current power usage, including whatever demand comes from this data center.
Yeah if I were starting now I'd be looking at jellyfin. But I paid the lifetime plex pass, and inertia/laziness what it is, so I haven't found a reason to actually switch yet.
Transformers are like blockchain: an interesting use of mathematical principles to solve certain problems in a novel way, where the hype around that core attracts charlatans and scammers and combinations of the two traits who claim that it will go on to solve totally different problems in such a way as to revolutionize the world we live in.
NFTs were the end of that line for blockchain where the machine started to eat itself. I can see a future, stable use of blockchain in some limited contexts, but cryptobros have always overstated the contexts in which that particular type of digital ledger can be more useful than other types of digital ledgers.
We'll see where the end of the road is for transformers, and what's left at the end. I believe that computer inference will always be useful in some contexts, and that the advances in huge models with absurdly large numbers of parameters have unlocked some previously impractical tasks, but I could also see that settling into a general background existence as just another technological tool for doing things in a world that still looks pretty similar to the world today.
Hey now, that infrastructure is good for, like, 3 years, so it's really like spending $145 billion to save $81 billion. And that doesn't even start to get into how much it costs to operate that infrastructure.
Seems like performance per unit cost necessarily has to spread between performance per dollar that the actual computer hardware costs and performance per watt-hour of ongoing energy consumption. Obviously each generation of chips shows exponential growth in performance, but how much is that advancement offset by increased power consumption and increased price of the chip itself?
That's what I mean. I'm interested in seeing actual numbers, like seeing how costs differ in specific model/hardware combinations, with certain assumptions on the price of electricity and maybe interest rates or amortization schedules.
He flagged the issue with flat rate subscriptions not making any sense for the underlying token pricing and usage by users, and predicted that a lot of the AI startups that act as some kind of subscription middleman would feel the squeeze and eventually impose rate limits/quotas, degrade the quality of their offerings (i.e., push users towards cheaper models), or fail. I think that's a pretty good summary of what has been happening at the user/pricing level with Perplexity, Lovable, and Cursor. Microsoft's Copilot plans are also seeing a lot of changes to pricing and rate limits, as well as model choice, in ways that user complaints have gotten louder in the past month or two.
He was a skeptic on Stargate right out of the gate, and I think that external visibility into how that loose collection of projects under that banner has been going over the past year shows that something inside is fundamentally wrong. That isn't necessarily an indictment of the broader AI ecosystem as a whole, but Zitron's most pointed financial criticism has been directed at OpenAI and Oracle, and the costs of data center construction. Those criticisms have looked especially prescient this calendar year (and generally fits into my preconceived notions that building physical stuff is slow and expensive and that we Americans aren't very good at keeping megaprojects on schedule and under budget).
I'm a money guy. I don't have any special expertise in industry trends and how money will be spent in the future on industries where I'm not an insider (i.e., AI), but I find Zitron's accounting of how money is being spent in the present to largely seem accurate. So that's why I'm in this thread asking people about how they see the present and the future of spending/pricing/volume, to see if those projections of revenue needed are actually feasible.
Not everyone needs a Lamborghini or Concorde to get where they are going.
I agree with that. Still, Lamborghinis are still being built, operated, and maintained, while Concordes are not.
I'm wondering whether the future of AI looks like the last 50 years of aviation, where there aren't that many generational advances because the cost of developing new stuff becomes prohibitively expensive, but where the commoditization of what has already been invented makes it so that the experience for the average person really isn't that different between 2026 and 1976, where the sweet spot for cost effectiveness isn't at the bleeding edge at all.
And for my own curiosity on this line of thinking, I wanted to know whether the day-to-day cost of running these models is going down, and in which contexts.
You'll still get a few years before the software becomes remotely disabled, though. This story about Office 2019 losing functionality follows Office 2019 losing support in 2023. If that's the rate things go, then maybe Office 2021 will lose functionality either 2 years from now (7 years after release) or 3 years from now (3 years after losing support).