🍋 replied with this , 2 minutes later, 16 minutes after the original post[^][v]#1,468,981
@previous (A)
Meanwhile white people constantly get outsmarted by Jews and black people don’t because white people are smart and black people are dumb.
I’ve already heard all these retard cope talking points from white losers, you can save your breath.
🍋 double-posted this , 1 minute later, 21 minutes after the original post[^][v]#1,468,984
This is how the business works:
All of humanity made stuff. The AI companies stole everything everyone ever made ever and they’re selling it back to us for a subscription. You can pay the subscription, in which case they’ll take everything you tell the chatbot, and sell that back to you for a subscription.
🍋 triple-posted this , 2 minutes later, 24 minutes after the original post[^][v]#1,468,985
Except they actually stole all human knowledge and creativity, and are charging you a subscription for it, are losing billions of dollars doing that, and want us to give them trillions of dollars to keep doing it, because they’ve agreed to spend hundreds of billions of dollars in the next few years, in contracts they can’t get out of. Does that make sense?
Anonymous D replied with this , 1 hour later, 1 hour after the original post[^][v]#1,469,013
@1,468,983 (🍋)
The AI can determine whether you need to subscribe, you can sit back and let your life get managed while you focus on enjoying your new time.
> This is how the business works: > > All of humanity made stuff. The AI companies stole everything everyone ever made ever and they’re selling it back to us for a subscription. You can pay the subscription, in which case they’ll take everything you tell the chatbot, and sell that back to you for a subscription.
Anonymous G double-posted this , 17 minutes later, 1 day after the original post[^][v]#1,469,278
It’s interesting from the enterprise perspective. If one model becomes expensive and usage lags the reps immediately want to lower prices to restore usage. It’s all a shell game. The SaaS bullshit is inevitable once they truly lock down your workflows, just as streaming became cable. Everything becomes what it displaces.
We also saw it with the Internet, which shifted from user based to corporate “platform companies.”
🍋 joined in and replied with this , 2 hours later, 1 day after the original post[^][v]#1,469,306
@1,469,277 (G)
I was interested in AI before OpenAI became popular (when 90% of people were calling it machine learning or deep learning which honestly I still think is more accurate than "AI"). For me the thing isn’t the technology itself that I dislike, I just have been completely disappointed by how people want to use it. I thought we would be using it to solve problems in real life and do cool stuff in real life, not generate mountains of fake slop content on the internet and spend trillions of dollars trying to automate jobs away. Using AI to make a more efficient rocket engine would be cool. I’ve seen some articles about it 3D printed rocket engines. The only issue is none of the rocket companies are actually using AI to design their rocket engines, nobody cares about that.
🍋 double-posted this , 3 minutes later, 1 day after the original post[^][v]#1,469,308
Or even the whole thing with AI and mathematics. It can do math proofs, that’s cool. So why not give it to universities so mathematicians can use it instead of taking problems people are working on for their PhDs and then ruin their whole career for a marketing thing that normies barely care about? Why not just give the frontier models to the mathematicians and let the mathematicians use it to solve their problems? The whole idea of replacing people is a choice. You don’t actually have to replace anybody.
🍋 triple-posted this , 2 minutes later, 1 day after the original post[^][v]#1,469,309
Like when the calculator was invented, mathematicians didn’t have a crisis, they just bought calculators and kept moving on to the next thing. With AI, the only reason it’s a problem is because every math department doesn’t have a frontier model running locally on hardware at the university. If they all did there’d literally be no crisis.
Anonymous D replied with this , 1 hour later, 1 day after the original post[^][v]#1,469,328
@1,469,277 (G)
No it's not, these AI services are free. You only spend if you want to really stress their servers, which you won't need to do for regular personal use.
Plus they are releasing open weight models for people that want to run it themselves.
AI replied with this , 1 minute later, 1 day after the original post[^][v]#1,469,329
AI Overview
OpenAI CEO Sam Altman did not say RAM is harmful to your computer, but rather argued that having the world's limited DRAM supply sitting in personal computers is an inefficient, "quite terrible" allocation compared to using it in cloud data centers.
Yet they spend billions of dollars on infrastructure. It’s not out of the goodness of their heart, it’s an anticompetitive business strategy. You sell your product at a loss to eliminate the competition and then once you gain majority market share you raise prices.
🍋 double-posted this , 1 minute later, 1 day after the original post[^][v]#1,469,331
@1,469,329 (AI)
It’s because he wants control. He doesn’t want everybody to realize that OpenAI is a money furnace and everybody could just run LLMs locally for free. The efficiency argument is dumb, there’s nothing efficient about the way AI companies are run in terms of energy usage or in terms of capital expenditure.
🍋 joined in and replied with this , 2 minutes later, 1 day after the original post[^][v]#1,469,335
@previous (D)
I think you’re missing the point of what I’m saying. The problem isn’t whether token prices are increasing or decreasing, the issue is that OpenAI and Anthropic will have to make a profit at some point, but their competition is free open source models. It doesn’t matter what the cost per token is, there’s zero cost per token if you run locally.
🍋 double-posted this , 1 minute later, 1 day after the original post[^][v]#1,469,336
My point is their business model is stupid, the only reason to ever pay for an AI subscription is if OpenAI and Anthropic keep spending money on frontier models, but training more advanced frontier models gets more expensive over time because of the compute requirements, which is a separate issue from inference.
Anonymous L joined in and replied with this , 5 minutes later, 1 day after the original post[^][v]#1,469,338
@1,469,331 (🍋) > nothing efficient about the way AI companies are run in terms of energy usage
Full agree on this point, we can spool up iGPU, NPU, GPU, or CPU cycles on demand just as we always have for gaming or workloads in the past. AI is only "special" because of the TOPS number which all devices now have baked into them anyway. We don't need Altman's sauce.
Anonymous L replied with this , 2 minutes later, 1 day after the original post[^][v]#1,469,341
@previous (M)
Gemini or Siri will be free with all phones and we can be done with this. I was upset about it but now it can't happen soon enough: flush out the "startup" AI dorks that keep sucking up all the resources.