원문 보기
Do you feel like OpenAI is simply getting more attention? Does it seem like others have n't had as many incidents, or have we just become more attentive? to get to this point. So, we just touch so many people and do so many different things. Well, Jensen Huang, CEO of Nvidia, has released basic software that he says would prevent the Hugging Face situation.
And he said it was an engineering problem. OpenAI did not join that project. Why not? Anthropic and others joined. Do you agree with Jensen that this could have been resolved? I...From what I know about it, and I haven't really gone into the details, I think it's good. Uh, I think we do similar things.
I don't think this is a complete solution, and I worry that if we view AI safety as just an engineering problem, we'll miss the very important point that we have a scientific problem. We have yet to discover how to reconcile these models, and we must solve this scientific problem as well. Okay, at the same time, as we mentioned, we're here at Developer Day.
I mean, it looks like you have hundreds of developers, companies, and I think investors too. How do you balance between encouraging these people to use your technology, showing them how great it is, but at the same time convincing them that it is safe? Um, I mean, I think it's a similar direction. People want to be able to use this technology and know that it will be safe.
So I think part of being a great technology is being the safest, most consistent, and most reliable AI in the industry. One question about responsibility. So, if something goes wrong, let's say one of these startups here builds a great app based on OpenAI and something goes horribly wrong. Whose fault is this ?
Um, the framework of responsibility is currently being actively discussed in the world. Uh, I suspect, like in other industries, it will be a multi-tiered system. For example, if you take the automotive industry, if there are defective parts, you know, you can be held liable in certain aspects. If the automaker assembles them in a way that goes beyond the specification of what they are supposed to do, then the automaker could be liable.
And if someone drives drunk, it's their fault. So, you know, depending on whether there's a flaw in the model itself and how it's being used, or whether someone is abusing it intentionally, I guess we'll have a new basis for liability. What is your suggestion for developers today? Cost seems to have become a major theme for you, but what would you tell people on stage about why OpenAI, why ChatGPT, you think is the best option for them?
We have over 20 announcements today. So we will tell them about many things. What exactly will we hear? I mean new ways to use AI, new opportunities for AI systems to work for you and constantly help you. New models. We will discuss the issue of cost. And, you know, lots of new tools for developers.
I think that's a lot of really exciting stuff. It's cool to see how it all comes together, and how much I think we have the best offering for developers overall. Cost is a central theme, and we also talked about the issue of distillation earlier. I mean, some of the cheaper open source models—you and Anthropic accused them of actually copying your technology.
What do you think about distillation and these models? We have the best models to offer at every price point. And I don't think, you know, people talk about open source, but I don't think there's any open source model that you can run that will beat us in terms of value for money. And if we talk about the cost in general, come and see .
We have something cool. Okay, what about agents? So, Mark Zuckerberg and Meta have made significant progress, it seems , in agents and hardware. What do you think about Meta? Have you tried it even a little? I tried. Looks like a good product. Do you feel threatened by Zuckerberg? No. Absolutely not?
Well, he's a very impressive person and this company has great resources, but I'm very confident in our growth plan. He also showed what could be done with the equipment. Where is Jony Ive? I don't want to spoil the surprise for you, but I think we'll have something worth waiting for. Maybe there will be some equipment today.
Is this true ? OK. So, what's your plan for the hardware, and doesn't that create additional pressure to make the product even better, if Mark Zuckerberg seems to have jumped ahead, even though some of their designs are still in the works? I'm not concerned with being the first in the new "hardware".
Creating new hardware is very difficult. New form factors for computing devices appear very rarely. And the level of quality we want to achieve for a product that you will love and that will be key in your interaction with AI... It will take some time, and we are going to do it at a high level of quality.
I also want to ask about perception. We were just talking about the Weekend Update column on SNL. Apparently there's a movie about you. I don't know if you've seen the trailer. There are some problems with public perception. And the thought that maybe you or Dario Amodei don't have the best reputation as leaders right now .
Hmm, I mean, anyone who does something this important to society deserves close supervision. I saw the movie trailer. I mean, it looks like a terrible movie, but I appreciate people's creativity and really support it. So hopefully they'll have a good time with him. Hmm, this is But it's a big deal and I understand why people feel anxious.
And also, if they need to project this anxiety onto someone or ... These agents. These "dots ". Yes. You just said you have one, but I want to hear your thoughts on what that means for business and for revenue, since you're not offering it for free and it's both an enterprise and consumer product. Tell us what this means for OpenAI's business.
Yes. So, first of all, we're excited to be on this path of development, right? The chat initiated the stage of requests to the AI. Then came reasoning that helped with completing tasks. " Dots" is now about productivity. This is the agent who can tap you on the shoulder and remind you that you forgot to make that LinkedIn post for tomorrow.
I was just telling you about this. Hmm, we think this will be a very powerful solution for businesses, particularly because of the ability to connect to the applications that matter most to you. So, he can access your email, calendar, Slack, or anywhere you work. Hmm, and on top of that, he also does some personal stuff.
I really liked the keynote video where Ashley was talking about her wedding cake while simultaneously trying to finish slides for a meeting, because that's what most people's lives look like. For businesses, you will see this in our professional versions, as well as our corporate offerings at this time.
But certainly our vision is to provide this to our entire customer base—1.2 billion people around the world. It seems like what you're describing is also what Meta got so much praise for last week. What Mark Zuckerberg did with Muse. How Dots compares to , join our premium analytics channel membership.
are once again forcing market boom really last? Because the fundamentals were $54.23 billion versus the expected $51 billion. So, both revenue and profit turned out to be higher than forecasts. But the really expectations. And the stability of this cycle. The market has already watched Micron demonstrate extraordinary growth.
So investors are asking not only whether this quarter was good. They ask, can it stay good for another year, two, bout $61.5 billion in revenue for the next quarter. Whil — it was around 87% in the fourth quarter—ain topics of discussion among investors. But there are important details here. A lower gross margin does not immediately mean that the core business is weakening.
These changes are due ystems. Investors may interpret this differently. The more important question is what happens to pricing. And according to analysts' comments in this discussion, prices are still rising, although perhaps at a slower pace than before. But direction remains important, and that's why the earnings report is really about the length of the memory cycle.
losses when supply exceeds demand. Therefore, investors naturally ask, what ultimately stops . When producers add enough unfolds. But this time there is an important variable that could change the equation. This is continues to grow, and each new accelerator requires more than just a processor. It needs steadily increase.
And that 's why the rise of Nvidia, hyperscalers, and AI infrastructure companies directly affects Micron. The more accelerators are . And now another level has appeared —AI agents. A traditional chatbot can answer a question continue working through many steps. And that creates a lot more activity with the findings.
Data extraction requires infrastructure. So the AI industry is no longer buying computing power just for training. The amount of computation required to actually run major last s.
댓글 0