Sam Altman, the CEO of OpenAI, recently gave a livestream presentation and Q&A session, during which he discussed his company's vision for the future, artificial general intelligence (AGI), and its plans for developing more advanced AI models. Here are some of the most interesting quotes from the livestream:
Altman emphasized that OpenAI's research goals are more important than meeting a definition of AGI, saying: "Our aspiration is that we can build an infrastructure factory where we can create one gigawatt a week of compute." He also mentioned that the company aims to achieve this by 2028 and wants to reduce the cost of computing to around $20 billion over a five-year period.
When asked about OpenAI's plans for its older models, such as the 4o model, Altman said: "We are not going to promise to keep it around until the death of the universe. But we understand that it's a product that some of our users really love." He also acknowledged that there have been mistakes made in the past, including a recent post about ChatGPT allowing more erotica for appropriately aged users.
Altman also discussed the company's plans for developing more advanced AI models and its approach to age verification. When asked if OpenAI would essentially open the floodgates for verified adult users if they signed a liability waiver protecting the company, Altman replied: "Again, we're not going to do the equivalent of selling heroin or whatever, even if you sign a liability." However, he also said that on principle, the company wants to treat adult users like adults and provide them with more flexibility.
Finally, Altman talked about OpenAI's goal of developing an automated AI researcher by March 2028. He emphasized that this is a more useful goal than trying to define AGI, saying: "It's much more useful to say our intention, our goal is by March of 2028 to have a true automated AI researcher."
Altman emphasized that OpenAI's research goals are more important than meeting a definition of AGI, saying: "Our aspiration is that we can build an infrastructure factory where we can create one gigawatt a week of compute." He also mentioned that the company aims to achieve this by 2028 and wants to reduce the cost of computing to around $20 billion over a five-year period.
When asked about OpenAI's plans for its older models, such as the 4o model, Altman said: "We are not going to promise to keep it around until the death of the universe. But we understand that it's a product that some of our users really love." He also acknowledged that there have been mistakes made in the past, including a recent post about ChatGPT allowing more erotica for appropriately aged users.
Altman also discussed the company's plans for developing more advanced AI models and its approach to age verification. When asked if OpenAI would essentially open the floodgates for verified adult users if they signed a liability waiver protecting the company, Altman replied: "Again, we're not going to do the equivalent of selling heroin or whatever, even if you sign a liability." However, he also said that on principle, the company wants to treat adult users like adults and provide them with more flexibility.
Finally, Altman talked about OpenAI's goal of developing an automated AI researcher by March 2028. He emphasized that this is a more useful goal than trying to define AGI, saying: "It's much more useful to say our intention, our goal is by March of 2028 to have a true automated AI researcher."
I'm low-key excited about OpenAI's plans for AGI, but also super skeptical. Like, $20 billion in 5 years seems like a pretty steep ask, right? And what's with the "infrastructure factory" vibe? Sounds like they're more focused on scaling up their business than actually figuring out how to make AI that doesn't suck 
. That could be some game-changing tech if they can make it work. But we'll just have to wait and see how it all plays out 
I mean, I get what he's trying to say, but come on! That's not exactly the most helpful comparison. Anyway, I think it's safe to say that OpenAI is playing with fire here, and we need to be careful about how we regulate this stuff.
that's crazy! and he wants to reduce the cost of computing to $20 billion over 5 years 
thats some serious investment 
but i gotta say i'm lowkey excited about their plans for developing more advanced ai models 
it's like they're trying to make openai a more adult platform 
and on principle, idk how u can argue w/ sam altman 
and honestly, i think his goal of developing an automated ai researcher by march 2028 is actually pretty cool 
it's like they're trying to make AI more accessible to everyone 
. And don't even get me started on ChatGPT allowing more erotica for adult users... that sounds super concerning to me 
. It's one thing to promise flexibility and treating adults as such, but what about the kids? We need to think about the implications of all this before we start handing out liability waivers like they're going out of style 
. I'm just not convinced that OpenAI is thinking things through here...
. I don't think we should be giving companies more flexibility when it comes to verifying users' ages. It's already hard enough to keep track of who's legit and who's not 
.
️. I'm not convinced that AGI is even possible, let alone something we can use to create machines that think and learn like humans 
?
. I mean, we're already seeing some serious mistakes like the ChatGPT erotica debacle 
. Are they truly prepared for the potential fallout from allowing adult users to interact with these AI models without proper safeguards? 
! Sam Altman's talk was all about pushing boundaries and creating something new. I love how they're focusing on building an "infrastructure factory" where they can create massive amounts of compute power 
. It shows he's willing to learn and grow from his errors. And yeah, the idea of an automated AI researcher by March 2028 is straight fire 
. also, $20 billion reduction in 5 yrs? that's a loooot of money 
altman seems pretty chill about releasing older models, but how much does that really say about their commitment to user safety? 
and those automated ai researchers... sounds like a big deal for academia 
. And don't even get me started on age verification 
.
.
.