I haven't watched the whole interview. In the clip, a couple of things jump out:
1. He was speaking to a receptive audience. The head nods when he starts to make the comparison between the energy for bringing a human up to speed versus that for training an AI.
2. He is trying to rebut a _specific_ argument against his product, that it takes even more energy to do a task than a human does, once its training is priced in. He thinks that this is a fair comparison. The _fact_ that he thinks that this is a fair comparison is why I think it is too generous to say that this is just an offhand comment. Putting an LLM on an equal footing with a human, as if an LLM should have the same rights to the Earth as we do, is anti-human.
It also contains a rather glaring logical flaw that I would hope someone as intelligent as Altman should see. The human will be here anyway.
> 2. He is trying to rebut a _specific_ argument against his product, that it takes even more energy to do a task than a human does, once its training is priced in. He thinks that this is a fair comparison. The _fact_ that he thinks that this is a fair comparison is why I think it is too generous to say that this is just an offhand comment. Putting an LLM on an equal footing with a human, as if an LLM should have the same rights to the Earth as we do, is anti-human.
> It also contains a rather glaring logical flaw that I would hope someone as intelligent as Altman should see. The human will be here anyway.
Exactly. Perhaps in Altman's world, a human exists specifically to do tasks for him. But in reality, that human was always going to exist and was going to use those 20 years of energy anyway; they only happened to be employed by his rich ass when he wanted them to do a task. It's not equivalent to burning energy on training an LLM to do that task.
> as if an LLM should have the same rights to the Earth as we do,
I don't see him calling for an LLM to have rights. I don't think this is part of how OpenAI considers its work at all. Anthropic is open-minded about the possibility, but OpenAI is basically "this is a thing, not a person, do not mistake it for a person".
> It also contains a rather glaring logical flaw that I would hope someone as intelligent as Altman should see. The human will be here anyway.
His point is flawed in other ways, like the limited competence of the AI and how even an adult human eating food for 20 years has an energy cost on the low end of the estimated energy cost to train a very small and very rubbish LLM, and nowhere near the energy cost of training one that anyone would care about. And even for those fancy models, they're only ok, not great, etc., and there are lots of models being trained rather than this being a one-time thing. Or in the other direction, each human needs to be trained separately and there's 8 billion of us. And what he says in the video doesn't help much either, it's vibes rather than analysis.
But your point here is the wrong thing to call a flaw.
The human is here anyway? First, no: *some* humans are here anyway, but various governments are currently increasing pension ages due to the insufficient number of new humans available to economically support people who are claiming pensions.
Second: so what if it was yes? That argument didn't stop us substituting combustion engines and hydraulics for human muscle.
> He’s clearly saying “lots of important things consume energy” not “let’s replace humans with GPUs” or “humans are wasteful too”.
When people have to interpret what you are saying, assuming that you are too intelligent and empathic to mean what you actually said, I think it says a lot.
"What he said is wrong, illogical and dangerous, but you have to forget it and consider that he probably meant this completely different thing that I will expose to you. Because he cannot be rich and powerful AND capable of expressing basic ideas on his own, what did you expect?"
The people who got offended at the 2011 campaign are not the same people who are offended at this 2025 campaign. In the united states, if you do anything, someone, somewhere, will be offended. That's kind of our whole shtick.
I haven't thought of a word for it yet, but it has something to do with how many people participate in the discourse now. The numbers are large enough that someone somewhere will always have some opinion. Every time.
> this would be the first time that a high core count CCD will have the ability to support a V-Cache die. If AMD sticks to the same ratio of base die cache to V-Cache die cache, then each 32 core CCD would have up to 384MB of L3 cache which equates to 3 Gigabytes of L3 cache across the chip.
Loved Myst/Riven when I was a kid. Played most of it sitting on my grandpas lap.
I can’t explain why, maybe it’s my current lack of sleep, but I can’t imagine a life where someone has the time to dig so deep into something so recreational. I mean no spite and no insult! This is cool! Maybe I’m just jealous - I spent my day choring and looking after kids - and I will collapse in 2 hours when I’m done with my obligations.
If you find yourself with free time - use it! Create works like this article or like Riven itself. Life is SO short. Make something people want. Apparently I wanted to spend 10 minutes comparing renders from 30 years ago instead of picking up toys!
Not to be "that guy" but Anduril is Aragorn's sword and is the most good-guy good-thing that could ever be fantasized about. It's used to defeat Sauron. And the Palantir stones are not "the bad guys tool", they were made by the Elves in ancient history and a few of them wound up in the bad guys hands. Misread LOTR indeed!
I specifically referred to the witch kings surveillance artefacts with misreading.
I don’t think their creation story is mentioned in LOTR, other than that they are extremely powerful and dangerous.
But you are right of course about Anduril and if you take the whole silmarillion as background. I never really liked that part though
The Palantiri were created by the Elves in Valinor and given to the high race of Men.
The witch-king could in theory have used a Palantir, but there’s no suggestion he did.
The seven stars in Gondor’s crest are the Palantari, and in the War of the Ring, Aragon specifically requested they be added to his banner. They represent the highest level of the civilization of Men.
Yes, but the elf who created them is quite a tragic character himself. To the extent that his own mother chose to die after giving birth because she knew how much sorrow he would eventually bring. So I'd be careful to not paint them as a good thing either.
you're right, and definitely Palantir is a harder sell here. But to say "they named their weapons company Anduril, what are they, bad guys?" frustrates the nerd in me quite a lot.
Well sure you still have 2 or 3 infra people but now you don’t need 15. Comparing to modern Hetzner is also not fair to “cloud” in the sense that click-and-get-server didn’t exist until cloud providers popped up. That was initially the whole point. If bare metal behind an API existed in 2009 the whole industry would look very different. Contingencies Rule Everything Around Me.
While this is tragic, undeniably so, it’s worth knowing that a head on collision in Malibu two days ago killed a 50 year old man when a 20 year old crossed over the double yellow line. It was obvious seeing the car that the young person was racing and driving dangerously. It barely made the news. I only know it happened because I drove past the wreck.
Tragic about the cat - and Waymo must improve - but we cannot lose sight of the greater good.
No, We should be fighting tooth and nail against these companies. They're not here to save us from ourselves. They're using public streets to Alpha (beta if you want to be generous) test autonomous lethal weapons, and then profit off of it when it works.
I can't find anything saying waymo has a thermal camera. They aren't expensive- certainly not compared to the LIDAR- and provide extremely discriminated input on "am I about to kill something?" They're not perfect as foul weather and fog are likely to blind thermal- but they shouldn't be driving in suboptimal conditions until they have a track record of safety in optimal ones.
What criteria would you consider sufficient for deployment on public streets? My experience is that people opposed to AV technology usually aren't familiar with the level of validation that's been done and tend to have expectations that are either impossible or are already met.
Waymo has experimented with thermal imaging in the past. I've never seen experiments indicating it's a particularly valuable modality for AVs, and high resolution thermal cameras exceed the price of decent LIDAR these days. You can easily spend $10k+ on a FLIR sensor with a pixel count higher than 4 digits.
Waymo was started partly to save lives by Sebastian Thrun who lost a friend to a car accident when he was 18. They have about 1/3 the accident rate of human drivers. Calling this stuff evil is kind of sad.
In some countries, drivers are expected to prove their ability to operate heavy machinery safely, held that promise, and governments prioritize zero deaths in their spending and policy making.
In the U.S., billions of dollars that could be spent on proven ways of solving the problem are instead spend on speculative robotic car development.
Robotic cars are not the only solution. They may eventually be as effective as proven solutions that are offensive to U.S. car supremacists, but as of today, robotic cars have proven only to be better than untrained, inattentive U.S. drivers and the life-threatening domestic policies that enable them. Robotic cars aren’t trying to solve the problem; they’re trying to capture spending on the problem. If transportation policy magically changed overnight to force immediate, funded implementation of proven safety processes from other countries, the excuses given for Waymo and others to beta-test their “these fatalities are a necessary accident in service of zero deaths” robotic vehicles would no longer hold water.
Autonomous vehicles don't impede transit advocacy, and they genuinely can be massive accessibility improvements to disabled and disadvantaged populations. Unless you have a magic wand to make those changes, it seems like AVs are an improvement over the current situation?
Transit changes are not required to implement the safety changes made by other countries. The cause and effect is reversed here: safety changes make transit more appealing because safety changes tend to decrease peak vehicle capacity, but transit does not make safety more appealing to untrained and overconfident (or willfully unsafe) drivers. You can’t just focus on transit while ignoring drivers and expect people to stop dying.
I remember during the first days of Covid lockdown how 99% of the cars on the busy hill outside my apartment were replaced by transit, with a commute distance of zero miles. The people who liked to do downhill racing on that hill during the day sped up from their usual brake-screech limits of 40mph to as high as 70mph, in a 35mph residential with an unsignaled busy crosswalk. And they continued doing this until the end of the lockdown when other cars got in their way again. Transit might reduce total car volume but it would increase the mean kill rate per roadway vehicle without safety culture and spending shifts.
Regardless, people who are on transit or in autonomous vehicles are people who won't be increasing roadway risk. AVs can adopt new driving rules without the typical years of political struggle over license points. AVs also have capabilities for better traffic shaping in cities. Responding to SF's market street closure was very easy with AVs. It was the uber drivers and tourists who struggled.
Meanwhile, she persisted, we could have zero deaths tomorrow if it was important to our culture. As we each recognized, the culture clearly isn’t changing anytime soon — but that lack of cultural concern invalidates “reduces deaths” as a relevant marketing claim for robotic cars. Why is it the preferred talking point for advocates when it’s demonstrably irrelevant?
If we lowered the speed limit to 20 miles per hour country wide we would be a lot closer to zero deaths. But at what cost?
We live in an imperfect world of trade offs, not perfect solutions. Even requiring people to wear seatbelts has a cost
I genuinely don't understand what policy changes you think would lead to zero deaths tomorrow. We'd still have deaths even if no one left their driveway without a valid CDL and a resolve to never exceed 10km/h.
Yeah, I get that; and! some of it is particularly nonintuitive outcomes from human psych/soci that look ghastly through a rational behavior lens. There’s a lot of reading that one can do on the subject if independently curious.
So tell me where to find a not-profit-seeking human.
I’d say a government employee just seeks profit by doing as little as possible for the fixed paycheck they get. _Everyone_ has a profit motive. The question is how their profit aligned with that of others.
> tell me where to find a not-profit-seeking human.
Easy, most people fit that description actually. And that's fortunate because otherwise the world would collapse pretty quickly from lack of midwifes and gynecologists.
In fact, the neoliberal cult that neglects the human nature and pretends everything is shaped by monetary incentives is slowly destroying our societies…
Everyone need to pay their bills but that's a completely different thing from being “profit seeking”.
I'm pretty positive that very few of the women you know do prostitute themselves for a living despite it being the most profitable activity imaginable. Turns out most women aren't profit-seeking after all.
Reporting. I try to solve problems just to get them solved. I don't seek to enrich myself. By the way, living in your world of profit-seeking-at-all-costs maximalism is the cruelest fate imaginable, but assholes normalizing maximized greed are a dime a dozen, ruining things for everyone else.
He’s clearly saying “lots of important things consume energy” not “let’s replace humans with GPUs” or “humans are wasteful too”.
If Altman is to blame for anything, it’s that AI is a scissor-generator extraordinaire.
reply