I think OpenAI will IPO at 1T. I don’t want to say bubble but it could be one of these stocks super hyped that never goes anywhere after the IPO(I.e airbnb during Covid)
Brendan can do whatever he wants. Hes that good. If anybody seriously needed to interview him 20+ times to figure it out, then the burden is now on them to not fuck it up.
> ...it's not just about saving costs – it's about saving the planet
There's something that doesn't sit right with me about this statement, and I'm not sure what it is. Are you sure you didn't just join for the money? (edit: cool problems, too)
Right? Like what an incredibly naive thing to think, that BG is going to contain power consumption lmao. OpenAI is always going to run their hardware hot. If BG frees up compute, a new workload will just fill it.
Sure you might argue "well if they can do more with less they won't need as many data centers." But who is going to believe that a company that can squeeze more money from their investment won't grow?
Tangentially, I am looking forward to learn the new innovations that come from this problem space. [Self-rightous] BG certainly is exceptional at presenting hard topics in an approachable and digestible manner. And now it seems he has an unlimited fund to get creative.
Reminds me of when I was younger and thought of companies like Google and Tesla as a force for good that will create and use technology to make people's lives better. Surely OpenAI and these LLM companies will change the world for the better, right? They wouldn't burn down our planet for short-term monetary gain, right?
I've learned over the years that I was naive and it's a coincidence if the tech giants make people's lives better. That's not their goal.
I saw this video where this gen z girl was saying she preferred working for boomers who just wanted her to show up on time, get her work done, and maybe stay a little late from time to time. She said it was exhausting working for millenials who wanted her to think her job was "saving the world."
Probably because "making the world a better place" has been a trope used so much in the industry that it's made it to a TV show [1]. It's fine to be passionate about your job. It's fine to be paid well. You don't need to make us believe that you're mother Theresa on top of it.
The greatness of human accomplishment has always been measured by size. The bigger, the better. Until now. Nanotech. Smart cars. Small is the new big. In the coming months, Hooli will deliver Nucleus, the most sophisticated compression software platform the world has ever seen. Because if we can make your audio and video files smaller, we can make cancer smaller. And hunger. And AIDS.
If you trust what the executives of OpenAI and Anthropic say about their respective projects, its a die roll as to whether or not they will totally destroy the world. A theme of the last 5-10 years has been tech dropping the whitewashing of their reputation and embracing the idea that they what they are doing is incredibly sociopathic and still somehow cool (to them, I guess). Guess not everyone got the memo.
The HBO “Silicon Valley” series’ version of “making the world a better place by” nonsense. The blog article has fallen for the marketing of OpenAI. OpenAI is making the world a worse place by inflating the cost of RAM and even getting rid of RAM chip providers from the consumer space. Not to mention all the wasted power on compute for all sorts of meaningless tasks. At least with something like Claude I am saving months if not years of engineering effort and resources in a few hours.
Sam Altman is pro-extinctionist like most of the surveillance capitalist ghouls. He literally invests in mind uploading companies and believes only the rich deserve to "survive" the singularity he believes it is his job to bring about.
Sure, humans going extinct is good for the planet, I guess, but be up front about what you are really supporting.
If you’re going to hold datacenter operators to blame for the waste associated with non-optimized computation, then it would seem to follow that they get some credit for optimizing.
For me, it just sounds like a ChatGPT-generated sentence. Especially, it likes to write sentences like "it's not just about... - it's about ..." and it first sounds legit, but it doesn't really make much sense when you start to think about it.
If it's in your power, make sure user prompts and llm responses are never read, never analyzed and never used for training - not anonymized, not derived, not at all.
> She was worried about a friend who was travelling in a far-away city, with little timezone overlap when they could chat, but she could talk to ChatGPT anytime about what the city was like and what tourist activities her friend might be doing, which helped her feel connected. She liked the memory feature too, saying it was like talking to a person who was living there.
This seems rather sad. Is this really what AI is for?
And we do not need gigawatts and gigawatts for this use case anyway. A small local model or batched inference of a small model should do just fine.
It’s super dope, and you can have it talk to people for you in the local language when you go there. I’ve busted it out to explain what I’m thinking for me. Watching travel shows on TV or reading travel magazines is sadder.
It's sad that we aren't all rich enough to have a personal assistant to tend to our sides 24/7? mean, it seems more useful than, say, cruise ships, but they get to exist.
To answer a few people at once: I did mention compensation as a factor in the post, but I didn't elaborate details, so easy to miss. Comp is important of course, but so are the other factors. It feels like I can't go for a day without reading about the cost of AI datacenters in the news, and I can do something about it.
Apparently, there's this guy who's really good at optimizing computer performance and makes a lot of money doing it. At the same time, he writes mediocre school essays that are actually a bit embarrassing. Guys, if you have the opportunity to land a very well-paid job, then do it. Take the money. Live your life. But please spare us the public self-castration.
Brendan, I'm a big fan of your book, and work.
I don't have a problem with you joining OpenAI; best of luck there!
However, I'm not sure your analysis is quite correct, in this case.
If OpenAI can mobilize X (giga)dollars to buy Y amounts of energy, your work there will not reduce X or Y, it will simply help them produce more "tokens" (or whatever "unit of AI") for a given amount of energy.
So in a sense you're helping make OpenAI tools better, more effective, but it's not helping reduce resource usage.
Did the article intentionally start with a LLM cliche to filter out all the people who hate reading obviously generated content? I would say it worked.
79 comments
[ 4.2 ms ] story [ 73.0 ms ] threadThere's something that doesn't sit right with me about this statement, and I'm not sure what it is. Are you sure you didn't just join for the money? (edit: cool problems, too)
Sure you might argue "well if they can do more with less they won't need as many data centers." But who is going to believe that a company that can squeeze more money from their investment won't grow?
Tangentially, I am looking forward to learn the new innovations that come from this problem space. [Self-rightous] BG certainly is exceptional at presenting hard topics in an approachable and digestible manner. And now it seems he has an unlimited fund to get creative.
I've learned over the years that I was naive and it's a coincidence if the tech giants make people's lives better. That's not their goal.
[1]: https://www.youtube.com/watch?v=B8C5sjjhsso
Gavin Belson
Sure, humans going extinct is good for the planet, I guess, but be up front about what you are really supporting.
This seems rather sad. Is this really what AI is for?
And we do not need gigawatts and gigawatts for this use case anyway. A small local model or batched inference of a small model should do just fine.
I guess I'm a dinosaur but I think emailing the friend to ask what they are actually up to would be even better than involving an LLM to imagine it.
Asynchronous human to human communication is a pretty solved problem.
Or, you know, Signal/Matrix/WhatsApp/{your_preferred_chat_app}. If you're already texting things, might as well do that.
How could she not know?
However, I'm not sure your analysis is quite correct, in this case.
If OpenAI can mobilize X (giga)dollars to buy Y amounts of energy, your work there will not reduce X or Y, it will simply help them produce more "tokens" (or whatever "unit of AI") for a given amount of energy.
So in a sense you're helping make OpenAI tools better, more effective, but it's not helping reduce resource usage.
https://en.wikipedia.org/wiki/Jevons_paradox
You're in for a surprise buddy.