Sam Altman says OpenAI wants to sell intelligence like a utility
During a recent appearance at BlackRock in Washington, D.C., OpenAI’s Sam Altman, shared his vision for the future of AI. At one point saying, “We see a future where intelligence is a utility, like electricity or water, and people buy it from us on a meter.”
Altman was describing a world where AI becomes a foundational infrastructure, something woven into everyday life so deeply that consumers and businesses simply “plug into” it the same way they rely on electricity, Wi-Fi or running water.
Every capitalist is a conman.
I am so jealous of this guy’s ability to lie. I’m just so crap at it, I end up just telling the truth.
That’s not going to happen no matter what. Open source models are already catching up to frontier proprietary models. Altman wishes he had a monopoly over that, but he doesn’t. The bubble won’t last, and things like OpenRouter will become the main way people use AI. Google alone is a major reason why intelligence will never be like a utility like electricity or water or internet.
I’m interested to see what kind of hardware will produce “good enough” AI capabilities in a couple years as things are refined and tuned further. The gap between the absolutely massive commercial models and open source models keeps shifting but I don’t have the same fear that I had a few years ago that it might not be possible to get good results from anything less than millions of dollars worth of hardware.
For real. It’s like imagining metered electrical generation becoming the norm if solar was already cheaper and easier to install. Like if it wasn’t for decades of hydro and fossil fuel being the cheaper more flexible option it never would of made sense to focus so much on metered connections in the first place (and increasingly making less sense now…).
They will have to have some enforced mechanism to “franchise the sunshine” as they old saying went to prevent a world where they (big SaaS AI) are not clearly the option of last resort.
Even big corp and government world is increasingly “sovereign AI” focused now, just like hybrid cloud almost always makes more sense at certain level of IT infrastructure maturity.
Another desperate attempt to monetize. AI bubble-burst, here we come!
yeah no.
Literally the future capitalism has always wanted; all common resources seized from the public for the good of private equity.
ever since feudalism fell, dipshits all over the world have had one thing in mind: bring it back. now they’re almost there. and the peasants are all too ready to give it back.
There’s a name for that already: technofeudalism.
Dude outside of the technology difference. We are at a point most surfs in feudalistic societies had less problems to deal with then we do.
Like when you start accounting for betterment in medicine and farming alone. The avg joe would likely have a less stressful life under feudalism. Most people just want to be left alone, work an honest job, and have time to raise a family or at least spend time with them.
The problem is not that they’re trying to turn us back into a feudalistic society. They’re trying to turn us into a corpitocracy or an oligarchy. While, a feudalistic society can have a lot of the same similarities as an oligarchy or corporatocracy. They tend to be far more for the people and fair.
Feudalism would unironically be an absolute ideal outcome if we had to choose between the three.
Is at least in a feudalistic society. The farmers would own the their own land and there’s not much the big corporations would be able to do about that. It would actually elevate a lot of farmers in large landowners onto the same playing field as the big businesses that have a lot of money but not a lot of land.
If anything it would put them at a disadvantage cuz now they would have to fight an uphill battle to gain more land that they need to expand for these data centers.
Ideally we don’t go to any of them lol
The dude changes his pricing model everytime he’s interviewed. Dude has no idea what he wants to do, so long as it’s billable.
He just wants money
If it gets much more expensive then it already is I doubt it
I made a simple project tracker in vite in 30 minutes. It cost me ten bucks in tokens. That’s with a pretty cheap model (sonnet 4-6)
Right now it just makes more sense to not use the API directly, but the subscriptions, they seem to be better priced. Also, Anthropic seems to be more expensive…
How that will look in the future, I don’t know, I’m fairly sure they’re progressively increasing the price, or rather reduce the amount of tokens you can use in the subscription (as they did with gpt 5.5 and claude opus 4.7). The chinese competitors are getting increasingly more interesting. It’s also quite impressive how well small models like qwen 3.6 27B run already on a (not so affordable) 24+GB GPU, unfortunately still far from the quality of say gpt 5.5, but probably comparable to o4 or something like that, certainly usable.
While I don’t begrudge them trying to make money from their service, the deleterious effect on people’s lives, careers cannot be overlooked. I think it’s obvious that most AI companies are incredibly unethical so governments need to impose the ethics onto them and companies tempted to use AI. They should never be considered as powerful as a utility nor invaluable. Never.
I do begrudge them. Their service was created through mass copyright infringement
It’s a billionaire. I begrudge it for its mere existence.
I was heartened by college graduates booing these assholes.
They scraped the internet’s knowledge and want to sell it back to those who actually created it.
Rent-seeking has entirely replaced innovation in modern capitalism.
Innovation was only ever a small part. Rent-seeking is more fundamental to capitalism than “innovation” is. It’s just more brazen than we remember from our childhoods.
Paving Paradise in 3,2,1…
I’m not anti-AI, but anti whatever fresh hell they are unloading unto the masses. This is something that requires careful planning to ensure we don’t devastate resources or stall critical think skills and knowledge.
This is one of the hardest points for me to articulate, trying to convince everyday folks including families and friends that these Technologies are actively making us dumber.
Wiring up a solar and battery array, and then wiring up an entire miniature rack mount full of tech myself using ‘AI’ was absolutely critical in understanding the Nuance between different products and between different wiring schemes, but I realized after about 3 months that I was spending at least 15 times a day asking about the ampacity of different wire gauges (“how much current can this gauge of wire carry safely? What about that gauge of wire?”) Before I finally just made a table of common wire gauges in both aluminum and copper, and then printed it out and tacked it onto my wall like it was still 1997.
I reduced my net time spent querying by at least 20% in the past month by looking at my patterns.
This isn’t a brag. This is me admitting that I got stupid and then I’m forgetting the power isn’t knowing stuff but in having that knowledge at our fingertips, and that asking some mega Data Center two states away to boil half their freshwater and brown out half their town so that I can be told that I really do have to up my wiring material, makes me feel gross.
Fundamentally there’s no difference between a chart you made and asking ai. The problem is you’re not attempting to retain the information.
The upside to the chart is that it doesn’t rely on someone else providing you the means to your informational resources.
The problem isn’t AI makes us dumber. It makes it too easy to be lazy. If you actively try to retain the information that you are gaining and putting into practice from AI and not just letting it do everything for you. Then it’s no different than any other resource. Be it a shit you made on the wall or some shitty ass Reddit thread from 9 years ago. That has one dude with that answer.
Informational resources are only as valuable as your ability to have access to them and your willingness to retain the information so you don’t have to keep going back to your informational resources.
What AI does make you worse a is learning how to quickly and efficiently reference material. You become beholden to the AI to provide you information. Which is slow tedious and needs to be double checked half the time.
Honestly, the worst part of AI is the fact that it’s removing and delaying access to informational resources. It’s actually the same reason I personally hate discord. It’s unsearchable it absorbs information and hides it away and makes everything tedious and less useful.
AI is functionally just a walled garden of information. Instead of letting information be freely shared, you are putting public knowledge behind a paywall.
But stupid people never once attempted to retain information from a book chart or anything else. Anyways. So functionally they are as stupid now with AI as they were before without AI.
I mean, I used an LLM to generate the ampacty chart when I couldn’t find one I liked with web searches, and then just cleaned it up in a spreadsheet for asthetics, before printing it out. 😅
At first I tried using one of the image generation functions but I noticed that it was making up imaginary values AFTER if printed it. And therein lies my core criticism of relying on LLM’s, that they will lie to me with impunity and with absolute confidence every fourth answer they give.
You and I agree that readily being able to drill down to information is the superpower that we have to leverage in our human world, and that pattern recognition as well as memorization are key tools on that path.
I also feel that delay you’re talking about, injecting latency into my day, as I type out the same question for the eighth time on the sixth day, and realizing this is something that could just be on my wall and that I should probably try to flashcard.
I have definitely seen smart people get dumb with AI though, because it coopts and changes behaviors on some weird and fundamental level. Not just dumb people like me. 😅
Seriously. I’ve used it a couple times at work (in education) solely because i needed to fit a rubric in our LMS and the fucking UI to do it means I would’ve spent an hour, whereas the API just filled shit in as placeholders and editing was faster than creating.
Otherwise, I do my work by hand. I even set up excel sheets to do stuff for me like flag grade patterns or grade exams with a typed-in key. It’s almost fun, but I work with so many people who insist that “claude can do it” but then can’t have a followup conversation about what we supposedly discussed via email.
stall critical thinking skills and knowledge
That’s the point. They want people dumb so they can sell “intelligence” to people.
Only profits and power matters to these people. They will happily destroy peoples lives. Sam Altman is an obvious psychopath. Zuckerberg is less open with his real thoughts after he said “dumb fucks”, but equally ruthless. Musk lacks empathy and is autistic and doesn’t understand human emotions. Trump is a narcissist.
These are the people who drives Ai.
Or people who will need it will use local only, purpose specific, models to do the job… Things you don’t need a datacenter fed by a $100 bills fueled power plant for.
No, we wont
LLMs are not intelligent lol
I feel like I spend months trying to find the right phrase the people will understand when I mention llms, since I refuse to call these things artificial intelligence.
My favorite thus far, are “spicy autocorrect” and “next words calculator.” The fact that it has all of the compendium of human knowledge on physics or last millennium economics, means it is an excellent research assistant and engineering consultant, as long as I can keep in mind that it’s going to lie to me with impunity and with utmost confidence every fourth question I ask.
That’s the problem, a public archive with all it’s training data that you can query would be infinitely more valuable.
I mean, mostly yes, but also I did benefit from an LLM arguing with me and telling me to not be so cavalier with 15kwh of batteries. It was helpful to have some safety guard rails. It was useful for giving more expansive considerations and playing devil’s advocate against my designs.
I only shocked myself twice, from the solar panels. 😅
The Google AI overview is so often wrong when I ask it basic questions like who is in a TV show or movie, it’s unfucking believable. I’m very anti-Ai. It only creates bad, not good
Not to mention that a lot of kids are functionally illiterate because they just ask ChatGPT basic shit and it just spurts out bullshit to them.
It has actually only created a single good. Better language translators. Every major translator has been “ai” based since about 2016. The recent speed up of llms have massively improved their accurate ability to retain context and pick up on slang and colloquialisms.
Its seriously night and day how fucking good translators got entirely thanks to ai.
But that’s literally the only thing I’ve seen come out of llms that was worth two fucking shits.
