Sigh... Yeah... This country is fucked.
Still gonna vote, though.
We need to start with voting in such overwhelming numbers this November that Democrats get the House and Senate - ideally with a huge proportion of actual progressive leftists. It needs to be a large enough majority to completely stop any further actions on Trump's part.
Next, there needs to be congressional hearings that expose everything - fucking everything - the orange sack of shit has done to corrupt the government. Make it stay in the news 24 hours a day. Put federal officials under constant congressional subpoena. Make it impossible for the remaining Republicans in the House and Senate to avoid hearing about it.
Finally, when those Republicans have finally had enough, move to impeach and remove both Trump and Vance. Get both those motherfuckers out of there. Let the new Speaker of the House become acting president, and start cleaning up the mess.
At that point, start packing the court. Boost the number of justices to 15. Implement term limits for SCOTUS as well, and make them retroactive. Alito, Thomas, and Roberts: Gone.
If this happens, maybe in 2028 we can start really repairing the damage.
Guys, guys... I'm starting to think maybe the current SCOTUS doesn't have America's best interests at heart.
nvidia, microsoft, google, oracle, etc are all profitable
inference is profitable, training is not
When I don't include the costs of doing business, my business is profitable! That's silly. Inference might be very slightly in the green now when viewed by itself (although that's deeply questionable; no actual GAAP accounting has shown it to be so). But since training is an ongoing expense that frontier model providers have to constantly engage in, their companies are - and will remain - very deeply in the red.
And without seeing GAAP accounting showing where all the money goes in support of inference, I am highly doubtful that it's profitable.
training is what the majority of data centre spend is going towards
if they want to be profitable pull back the training but right now they are competing for market share
They can't. Ever. Pulling back on training means allowing model drift. You need to understand that models are obsolete the moment they're released. Their training data is set in stone. New version of Typescript ships? Some celebrity dies? Big election happens? The model not only doesn't know about any of it, it can't be updated. The best you can manage is throwing MCP and RAG at it in the hopes that the model will pay attention to it, but the point of diminishing returns on that arrives almost instantaneously. You have to train. Constantly.
feel free to look back at all the times lemmy predicted the end of spotify because it wasn’t profitable, now they turn around and cry it’s making money
Bad comparison. Spotify has already been a profitable, publicly-traded company for years.
And - this part's important - I'm not Lemmy. The platform we're having this conversation on has nothing to do with whether or not the AI model providers are profitable.
At work nobody is talking like this, everyone is talking about claude and it makes sense, it’s the best thing since vscode
Anecdotes aren't data. But as long as we're swapping anecdotes, here's mine: I work with actual machine-learning engineers. They're the ones who bag on Anthropic and OpenAI the most. And they use Qwen, Gemma, and a few other small, open-source, open-weight models. Have you looked at Hugging Face? Its community is huge, and growing daily. No one wants to be locked in to Claude Code or any other proprietary development tool when the service has been unstable and the pricing has becoming ridiculous in their desperate attempts to become profitable.
The cost for using Qwen tokens is $0, no matter how many tokens they use.
You say no one talks like this... Are you sure you're listening?
Anthropic is already profitable if you take out the enormous spend they have on training, which if the bubble bursts would leave them as the number 1 ai provider, it’s also insanely in demand and has trouble keeping up with its current product, they also have several products mythos etc lined up
First off... Why in the world would you take out the enormous spend they have on training? Training is an ongoing expense, not a startup expense. If your expenses exceed your income, then you're not making a profit.
Secondly, they had one quarter in which they reported (using non-GAAP accounting) a very slight amount of profit. That same quarter, SpaceX gave them a massive - and temporary - discount on rented compute.
We don't have any reason to think they're actually profitable.
I doubt it, I think it’d be closer to liberation day tariff’s or the oil crisis, it’ll go down for a bit, many articles will be written about how this is the worst thing ever then 6 months later it’ll be back up again
You're way more optimistic than I am. If OpenAI and Anthropic crash, there are a huge number of businesses that have built themselves around their products, and those will crash, too. And I think you're downplaying the damage the tariffs have already done.
As said all the major players in this game are super profitable major companies, that won’t change
Again, not true. OpenAI is not profitable. Anthropic is almost certainly not profitable. Grok from SpaceX is not profitable. Google is profitable, but not from Gemini. Microsoft is profitable, but not from Copilot.
No business that is built entirely on AI is profitable. Not one.
Look... No one's arguing that the coding tools built around AI are entirely useless. They're not (although their capabilities are way, waaaaay over-hyped). The problem is an economic one: Serving up AI models cannot be profitable. There's just no way, especially now that we have small AI models that can be run on local workstations, and offer similar performance to the frontier models.
Qwen, running in a well-designed harness such as OpenCode, with a carefully written AGENTS.md file, is of comparable performance to at least Claude Sonnet, and possibly Claude Opus. All without the massive, ludicrous infrastructure requirements.
How is Anthropic supposed to compete with that? Sure, you can probably get something useful out of Opus faster, but at the cost of thousands of dollars. Using Qwen and similar local models is free.
Which version? There are quite a variety, and it seems like they all have stellar intro songs.
No one thinks companies like Google or Microsoft are going away.
What's going to happen is that OpenAI and Anthropic will ultimately fold because they can't be profitable, Google and Microsoft will scale back their supremely unprofitable LLM operations, Micron and NVIDIA will plunge in value because all of a sudden their bloated prices aren't being paid by anyone, Oracle will suffer because OpenAI can't pay its enormous bills with them, massive data center projects will end before completion, and a whole lot of smaller businesses that embraced all this madness will collapse.
That will have devastating ripple effects throughout the economy. It's going to be a lot larger than the dotcom bubble bursting.
Would have been? EVs are exactly that. They're great. Ones with range can easily replace cars with internal combustion engines for most use-cases. Usually costs me about $5 a month to keep mine charged.
Fully agree on LLMs being expensive messes that aren't very useful, though.
Clever! And when sanity starts reasserting itself at your company, you can claim you're focusing on saving the company money by reducing your token usage and just... make fewer animal slop websites.
Software engineer, here. Yep, the burnout is real. I consider myself fortunate, however; with the skyrocketing cost of AI, my employer has been urging us to do as much as possible by hand lately to cut back on token usage.
I think that's pretty much where the entire industry will go soon.
It's related to the AI bubble. The AI companies are trying to make it as difficult as possible to get a good PC, because they know they're cooked if the general public has access to systems that can run AI models locally, so they're buying everything up as fast as they can in the name of data centers that will never be built.
As soon as the first one fails, it's all over. Prices will tumble and memory makers will come crawling back to Valve (and other hardware makers) begging them to buy.
Yeah, that's the one that stands out the most to me. I had no idea Origa had died. :'(