Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
On Thursday, I had dinner with Sam TamanA few other executives Openai, and a small group of correspondents in San Francisco. Altman answered our questions for hours. There was no topic outside the border, and everything, except for what was said on the desert, was in the record.
It is uncommon for such a wide range of interview with the main CEO of technology for a meal. But there is nothing common about the situation that altman finds itself. Chatgpt has become one of the most widely used products on Earth. Now, Altman is planning aggressive expansion in consumer systems, computers, and social media. He is interested in buying Chrome if the United States government forces Google to sell it. Oh, he wants to collect trillion dollars to build data centers.
But first, focus on responding to the GPT-5 program last week. About an hour before the start of dinner, Openai pushed an update to re -“warm” to 4O, its former virtual model of ChatgPT. It was Altman who called for a speedy return of 4O as an option to pay the subscribers after some protested against its disappearance on Reddit and X.
He said, “I think we completely fill out some things.” “On the other hand, our API traffic has doubled within 48 hours and grows. We are outside the graphics processing units.
He linked the percentage of Chatgpt users who have unhealthy relationships with the product in the “road less than 1 percent”, but he admitted that Openai’s employees are “a lot of meetings” on this topic. “There are people who actually felt that they had a relationship with Chatgpt, and those people we were aware of and think. Then there are hundreds of millions of other people who have no relationship with Parasocial with Chatgpt, but they are used to the fact that he responded to them in a certain way, and they will bear the health of some things, and they will be supported in a certain way.”
“You will definitely see that some companies go to make Japanese sexual robots because they believe they have identified something here to work.” “You will not see us. We will continue to work hard to create a useful application, and we will try to allow users to use it in the way they want, but not to the extent that people who have really fragile mental cases are wrongly exploited.”
Altman wants to feel ChatGPT as personal as possible, but not necessarily plays with a specific ideology or a political point of view. “I don’t think our products should wake up. I don’t think they should be on the contrary, either. I think our products should have somewhat center on the way, the middle position, and then you should be able to push it to a large extent. If you love,” I want you to reflect you. “
Chatgpt has almost interconnected the user base within a year and now reaches more than 700 million people every week. “Soon, billions of people will speak daily to Chatgpt,” Altman said. “We are the fifth largest website in the world at the present time. I think we are on a clear path to the third.” (This means overcoming Instagram and Facebook.) Then it becomes more difficult. For Chatgpt be larger than Google, this is really difficult. “
In order to run it to keep the scaling, Openai needs more graphics processing units. This is one of Altman’s higher priorities. “You should expect Openai trillion dollars to build the data center in the very not far -not,” he said with confidence in the room.
“We have to make these terrible bites at the present time,” he said. “We have better models, and we cannot provide them because we do not have the ability. We have other types of new products and services that we love to offer.”
It is also believed to be in the bubble of artificial intelligence. “When bubbles occur, smart people are excessive in getting the nucleus of the truth,” he explained. “If you look at most bubbles in history, like the technology bubble, there was a real thing. Technology was really important. The Internet was really big. People got excessive. Are we at a stage where investors as a whole are excessive about artificial intelligence?
Assure Modern reports Openai plans to finance the startup of the computer interface to compete Elon Musk Neuralink. “I think the nerve facades are great ideas for exploring them. I would like to be able to think about something and respond to it.”
Do Simo religions Ceivilt Ope To run “applications” means that there will be other independent applications besides ChatGPT? “Yes, you should expect this from us.” Glimpse Social media ambitions: “I am interested in whether it is possible to build a cooler type of social experience with artificial intelligence.” He also said: “If Chrome will really sell, then we must take a look at it.”
Although Altman has a lot of interests, it is not actually clear that the operation of Openai in the long run is one of them. He said at some point: “I am not a naturally appropriate person to be the CEO of a public company,” he said at some point. “Can you imagine me in a profit call?”
Then I asked whether he would be the CEO within a few years. “I mean, artificial intelligence may be in three years. This is a long time.”
Here are some other things that altman said:
- GPT-5: “We have had a large GPU crisis. We can go to another giant style. We can go to it, and many people want to use it, disappoint you. And so we said, let’s be a really smart and useful model, but also try to improve the cost of inference. I think we have done a great job with that.”
- Openai’s AI with Johnnyf: “It will take some time, but I think you think it is very weight. I think it’s incredible. Don’t get a new model often. There have been like just two years in the past fifty years. So let yourself be happy and surprised. It’s really worth waiting.”
- The future of the web and publishers: “I think people will go to lower web sites. I think people will care more about the human content more than ever. My direction’s bets will be that the content that was created for a person, and is coordinated to the human being, rises significantly.”
- What does AGI mean: “The teacher may be the most important for us is when most of our research set is allocated to the artificial intelligence researcher instead of human researchers. But I don’t think this will be very bilateral, because I think it will feel that people get more help and a little more help.”
- “If we don’t pay for training, we will be a very profitable company.”
- “I no longer use Google.
Interesting professional moves this week:
- I assume I should start asking, “Are you about to leave your job?” In interviews. A week after my country Decoder Episode It was published, CEO of GitHub Thomas Duhamki He announced that he will leave For a starting life. The real story here is that GitHub is less independent than the rest of Microsoft, which indicates that commercial interests Jay Barrick A new team of artificial intelligence that has absorbed it has officially exceeded the Gaytap Ethics Open in Swiss, Switzerland for coding.
- Igor BabushkinCo -founder and actual chairman Elon Musk a job, He announced that he is leaving Two years later to launch an investment company that focuses on the safety of artificial intelligence. (I’m sure Modern GROK addresses king nothing To do the timing of this news.)
- Alexander Wang Add More researchers Openai to the new artificial intelligence laboratory in Meta: Hyung Chung wonand Jason WeiAnd Zhiqing Sun.
- Antarbur added Dave OrHe is a Google old warrior who recently managed safety in Gemini, as head of safety. It also rented Jordan Burgis And his team is to start Humanloop, which works on LLM assessments of companies.
- Joel PinoThe former head of Meta joined artificial intelligence research, as COHERE as head of the artificial intelligence employees.
If you don’t already do, Do not forget to subscribe to freedomIncluding Unlimited access to Command line And all our reports.
As always, welcome your comments. You can respond here or Ping to me safely to sign.
Follow the topics and authors From this story to see more like this in your main briefing on the main page and receive email updates.