You must log in or # to comment.
Nuh uh, it targets babies and refugee tents like no one’s business! And that’s priceless.
Don’t forget girls’ schools.
Pfff, they don’t need to learn, do they?
Well no fucking shit. That the whole point, right?
Cool, now you need us back to fix the shit you broke / permanently disabled.
Reasonably sized LLMs solved the NLP problem, like YOLO + SSD solved permanently the object detection problem, why we didn’t celebrate that and stopped there or switched to a new endeavor like JEPA architecture? why they had to fucking push LLM more and over invest? my stupid ass could see this a thousand kilometer away, why they couldn’t foresee this?
Yes. That’s what I’ve been saying for a while. Without VC money to subsidize it, AI is fucking expensive. Literally the only reason anyone is paying for it is because it’s so cheap it’s barely worth it. Once the price goes up 10x, nobody’s gonna be interested.
Price per intelligence benchmark has been going down though but openai killed 4o mini which was the absolute cheapest smartish model.
The writing is not on the wall just yet, but with some players dropping out and some enshittification I could see models becoming very expensive.
But the benchmarks are semi useless. Performance at new problems in your precise use case matter and that’s only something one can find out you using it. This can differ wildly from what benchmarks suggest. AI can be an ability enhancer when for example someone who has no idea hos to code suddenly manages to create crappy code but that manages to do a job where for example success can be easily verified. But if the result should ve good or even reliable … well…
I mean eventually open weight models will be good enough to do the few things AI is actually good at, and then there’s another reason to not pay astronomical prices.
What’s your use case? To me, lazy api doc search, and grammar/flow check
I don’t use AI for anything.
Oh wait. I have started using it for something. I use it to look over contracts as a first pass and ask it to point out any red flags. I still read the contracts, but it can help find obvious unacceptable terms.
I switched las month from GPT-OSS-120B to Gemma4 31B. For some simple scripts I found the latter considerably mor efficient, less verbose and with better results. At the same time the sycophancy is much worse.
No way I’ll use cloud based models and feed the data base. But also local models clearly like to burn energy.
without?
VC - venture capital.
‘rich people with more money than sense who will give you their money if you can generate for them a stupendous return on investment by massively raising prices 3-5 years down the line’
That may be true, but a human will also do a better job.
Ehhhh, the majority of humans are pretty awful at their jobs. AI is worse than any competent human who gives a shit, but in modern office culture that’s more rare than it ought to be.
I really want to downvote you, but yes, if you pay people peanuts and treat them like shit, they don’t give you their best.
AI doesn’t have that problem so it’s great for the kind of soulless corpo who doesn’t value quality.
That does not have to be true. A human can also do a very bad job if they want.
On these wages?
They’ve invested so much money and infrastructure into AI that it’s impossible not to push it, even if it is more expensive with worse results. They’re too deep in the rabbit hole now and they know it. Everyone was betting on the next step of AGI and it turns out the step is a mountain that they may never scale. And of course, when the bubble pops, the executives will be fine and we’ll be left to foot the bill.
There will be no AGI, just automated surveillance.
deleted by creator
But I think there’s an important psychological dimension to this: bosses are especially easy to trick with AI when they’re being asked to believe that they can use AI to fire workers who are in a position to tell them to fuck off.
… What a relief it would be to fire everyone who is professionally required to tell you to fuck off when you want them to do stupid and/or dangerous things …
This also explains why media bosses are so anxious to fire screenwriters and actors and replace them with AI. … The difference is that the writers will call you a clueless fucking suit and demand that you go back to your spreadsheets and stop bothering them while they’re trying to make a movie, whereas the chatbot will cheerfully shit out a (terrible) script to spec. The fact that the script will suck is less important than the fact that swapping writers for LLMs will let studio bosses escape ego-shattering conflicts with empowered workers who actually know how to do things.
It also explains why bosses are so anxious to replace programmers with chatbots. … Tech companies had business-wide engineering meetings where techies were allowed to tell their bosses that they thought their technical and business strategies were stupid.
I’ve never seen gas prices so high and with sticking power:
$6.60 is the highest I’m seeing in the Seattle area for regular. Diesel and premium are basically $7.00 and beyond. The orange turd has basically fucked up the thing that keeps people buying stuff. So anyway, I wonder how AI is going to extract more from us for the rich to get richer.
That’s okay. Big cities have great public transport, right?
Right?
“AI - pay more for more incompetence”
shocked_pikachu.jpg
Tell me, can your AI employees and the AI employees of other companies buy your products?
Is it just me or does reality start to leak through occasionally now even in mainstream financial media?
