this post was submitted on 08 Sep 2026
1256 points (99.4% liked)
Fuck AI
8195 readers
1740 users here now
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I don't make a lot of predictions about artificial intelligence, but I will go out on a limb on this one and make one: this will never happen.
The human brain runs on 20 watts of power. Right now I have an AI running on an Intel NUC (Qwen 2.5 9B) that can do most home automation tasks I run through it (setting timers, setting reminders, making to-do lists, turning things off and on), as well as answer general questions ("who were the members of The Beatles?") and be (mostly) correct. It uses about 65 watts of power to do so.
It's nowhere near as smart as I am and it uses 3x the electricity I do. But I couldn't do this even last year. Efficiency gains to this technology have a ceiling so high that I don't think most people believe it because they see all these data centers being built and people like Sam Altman talking dumb shit like this. The future of AI will be local. It will run on your phone and be smarter than you, and in fairly short order.
We are already able to run models locally (Qwen 3.8 27B) that blow last-gen frontier models away (and give current frontier models a run for their money), but you need a high-end setup like an RTX 5090 24G and 64G system RAM.
LLMs and the surrounding code that drives them ("harnesses") are so far from AGI that it will be decades before we see your prediction come true. We need not just one, but many more leaps forward in AI progress to achieve a level of AI that is capable of performing human-level reasoning and planning that will be "smarter than you" or me. If you have studied AI in the pre-LLM times you will have an understanding of the problems that need to be solved for general reasoning intelligence. We haven't solved any of those problems.
LLMs are a large step forward in the field of NLP, but that's it.
I'm not here to argue that point. But my crappy NUC based LLM can answer the Beatles question. What it can't answer is "what was the name of the guy who famously rolled his eyes while Yoko Ono screwed up his performance?" Right off the top of my head, neither can I. But OpenAI's model would have no problem with that question.
So when I say "smarter than you or I" I don't mean AGI and I'm not trying to get into an AGI conversation. I mean something that runs locally, doesn't rely on the Internet and can answer almost any general trivia question accurately. And LLMs are already there, but not yet on your phone. When they are on your phone, people won't need OpenAI or Sam Altman's models anymore. That's all I'm trying to say. I'm refuting Altman's claim that we will pay for intelligence like a utility.
Yes, and I agreed with every word of your previous comment with the exception of
which I think is an extremely prevalent misconception about the state of AI that's worth debunking every time it is mentioned, and this misconception is due to the exact same causes as the point of this post: The CEOs are misrepresenting what this technology will scale to, in order to extract the maximum investment from unsophisticated investors. In doing so they're causing enormous turmoil in the labour markets that is affecting the lives of every white-collar worker.
Their narratives are harmful and they need to be called out.
I run a Qwen 3.8 27b at home on gaming hardware. Got it scraping and posting news articles and general stupid questions. Takes 5 minutes instead of seconds because it has a gemma model doing RAG in the back end.
Other than some coding, I really don't have a reason to use a big model.
Uh about that not needing Internet part...
The greatest leap forward since the release of gpt3 and ChatGPT is the ability to look stuff up. Well that and tool use in general... Basically restricting internet access handicaps it real bad, especially with smaller models
Doesn't detract much from your overall point but it's important to keep in mind
LLMs will never be AGI or whatever we end up calling that next, but it will get more efficient at being what it is. I do fear for the future of the entertainment industry, but at least it won't be too long before we can start building our own PCs again.
You may be correct from a technology perspective but you are forgetting how people work. These days people won’t even run their own cable modem or Wi-Fi access point. They would rather rent them from the ISP.
Right. This is because the ISPs actively obfuscate or hide the knowledge necessary to do this. In other words, they expect us to offload the intelligence behind this from them. It's already happening. It's what they done with cooking and pretty much every other skill we humans have quite amazingly developed. The news doesn't report about Local Language Models. They promote hyperscale data center Large Language Model stuff, making us associate "AI" with that and thus mystifying other ideas (that are less beholden to large corporations) to the point that "normies" are made to feel as though they are not intelligent enough to run this stuff themselves.
Just look at Linux. It's still treated as a kind of niche, "computer-smart" thing. I'm a GenX/Millennial "normie." I do have certain computer skills that both Boomers and GenZ lack, but I'm probably about at pace with most people my age. I switched over to Linux about a year and a half ago (from MacOS, where all the computing knowledge has been offloaded to a corporation) and have not looked back. These things are learnable. The learning just needs to be encouraged and made available. But the Sam Altman's of the world have such utter contempt for the rest of us they'd rather us be the mouth-breathing, drool-faced sheep they see us rather than actually do anything to better the human condition.
I'm talking about something that runs locally. On your phone, even. You wouldn't need a cable modem, wifi or an ISP. You wouldn't need Sam Altman or OpenAI. That's all I'm trying to say. I'm refuting the idea that we will need to buy intelligence like Sam Altman is claiming in the OP.
That's not at all correct. Or more appropriately, it's a true number used in a completely incorrect context.
It actually takes about 100 kilowatts to 1 megawatt to maintain a human brain, depending on your assumptions. You can't power your brain with electricity; you have to power it with food. And you can't power just your brain, you have to power your whole body. And producing food takes a ton of energy, way way more than people realize. It's just that most of it comes free from sunlight with crops grown in open fields. But if we're really talking about the energy needs of humans vs AI, the competition is between using those fields to grow food for humans to eat vs using those fields to hold solar panels to power AI.
I get this 0.1-1.0 MW figure from the YouTube and Nebula channel, "Science and Futurism with Isaac Arthur." There it was used in the context of deep space exploration and colonization, not terrestrial AI deployment. If you want to estimate the power requirements for life support for a hypothetical interstellar spaceship (just life support, not propulsion), you multiply the assumed crew number by 0.1-1.0 MW. If you're traveling between stars you have to bring your own sunlight with you. The vast majority of that energy is just for the huge amount of light needed to grow the food people eat. Interestingly, this life support requirement would still be a rounding error compared to the energy needs for propulsion. But if you want to estimate how big of a reactor you would need to sustain all of a person's needs without free sunlight, 0.1-1.0 MW is a decent figure.
For you as an individual deciding whether to pay to power an AI vs doing something yourself, sure, the 20 watt figure has some merit. You're going to have to eat enough to live no matter what you do. Maybe you burn a few extra calories while thinking hard on a problem, but your calorie needs are mostly fixed.
But when we're talking about whether it's cheaper to hire large numbers of humans or build large numbers of data centers to serve the needs of the ownership class, you have to look at the actual raw power requirements needed to keep a human being alive. Plant growth is an incredibly inefficient process. But computers can be powered extremely efficiently by just plopping some solar panels down in a field.
Well yes and if you wish to make an apple pie from scratch, you must first invent the universe. I think it can reasonably be said we're kinda spitballing here.
While I love Carl Sagan, you're doing him a disservice by using that quote here. When talking about whether human beings will be replaced by AI, you have to consider all the resources that go into maintaining both. You can't just hand wave that away by saying, "well, yeah, I guess, nothing means anything if you analyze it too deeply. Ultimately everything required the energy of the Big Bang, so we can't draw any real conclusions on the energy needs of anything. The world's just vibes bro."
We're not trying to get an exact figure. But when trying to predict whether AI will replace human labor, you have to make a reasonably fair comparison, otherwise your conclusions are completely worthless. Your entire point was that AI requires more energy to run than humans for a given task. If you think energy limitations are going to be the bottleneck for AI deployment, you need to make a fair comparison for the energy needs for both human and AI work.
Look at it this way. Let's say I have an acre of land. I want to turn that land into as much mediocre writing as possible. I could do two things:
Inevitably, I would be able to produce thousands of times more mediocre writing from that same acre of land with option (2) than (1). Biology is stupidly inefficient from an energy efficiency perspective.