- OpenAI CEO Sam Altman is the face of the AI boom. Here's what he said at his flagship conference.
- Tokens are "terrible," and the "Industrial Revolution" is a bad way to understand AI, Altman said.
- When it comes to AI safety, Altman said the tech industry should take a "centrist path."
If OpenAI CEO Sam Altman stands at the center of the AI boom, what he says about the technology can blaze a trail for the rest of the industry.
At OpenAI's DevDay conference on Tuesday in San Francisco, Altman announced the launch of a new AI agent product called Dots and laid out the company's grand ambitions.
He also set out some of his ideals, quibbled with how AI is understood, and talked about how OpenAI's new product is changing his morning routine.
Of course, everything Altman says should be viewed through a certain frame — he owes financial success to his investors, is competing with Anthropic and other AI companies, and wants to be seen as a capable steward of the powerful tech.
Here are some of the most surprising points he made on Tuesday.
Altman said AI shouldn't only 'turn the economic crank'
President Donald Trump loves to refer to AI as an "Industrial Revolution," only bigger. Altman is not so much a fan of the term.
Altman said he believes AI will have a greater impact on the world than the Industrial Revolution — a dramatic claim, given that it was one of the defining transitions in human history. He said that AI should focus more on empowering people.
"I don't think the purpose of this technology is to just sort of like, turn the economic crank faster and faster for the sake of AIs," Altman said.
"I think we should learn from the mistakes of the Industrial Revolution," he added, "and I think we should be more like the Renaissance."
Altman said tokens are a 'terrible' measure of AI
A token is a cluster of letters and the basic unit of AI's input and output. Companies like OpenAI use tokens to price their models, and corporate customers have grown accustomed to paying for AI access this way. Some developers have even been tokenmaxxing — using as many tokens as possible to appear more productive at work.
Altman said Tuesday that tokens are a "terrible metric for a bunch of reasons," including that different models can churn through different numbers of tokens to complete the same task. In other words, they aren't a good proxy for the actual value of a task, in his view.
The CEO said that his team has considered eliminating tokens entirely as a metric, though customers pushed back. They want some kind of "underlying basis" for what they pay, Altman said, and for now, tokens are it.
"We have not figured out a better way to do that," he said.
OpenAI's Dots help Altman escape brainrot
OpenAI is building a device, and Altman said on Tuesday it would be "reasonable to assume" that the company's new personal AI agent tool, Dots, would run inside it.
For now, the agent helps him get less attached to his smartphone, he said. Because it handles some tasks for him and processes his notifications, he feels calmer about his to-do list when he wakes up.
"I feel like I've finally gotten some of my attention back," Altman said. "I no longer feel quite addicted to my phone in the same way."
He later said that going through everything from overnight in the morning would make his "brain feel polluted, and now my Dot just deals with the stuff."
Altman is figuring out his lane in AI safety
Altman faced several questions about AI safety in the press Q&A, amid an ongoing surge in anxiety over newly capable AI models and a rise in cyber incidents. OpenAI has previously disclosed incidents where its AI agents have gone rogue.
Altman framed the ongoing debate as flawed on either side. On one side, he said, are people who want to charge ahead in the industry and keep quickly ramping upAI's capabilities— basically, the people who downplay safety concerns. On the other hand, many people think AI development should be fully paused or shut down.
Altman said he can imagine things going wrong in either direction: loss-of-control incidents amid an AI capability push, or a concentration of power if AI development stops.
He said safety needs to stay ahead of capabilities, and the push forward should continue.
"I think as an industry, we need to navigate this centrist path," Altman said.
Altman says it's the 'best time' to enter the job market
Concerns about the job market are top of mind for young people graduating into the midst of the AI boom. The unemployment rate for recent college graduates in the US was 5.7% in June, according to the Census Bureau and the Bureau of Labor Statistics,
In response to a young attendee's ask for advice on Tuesday, Altman said, "I think this is the best time to be entering the job market, except hopefully all the time in the future."
He said that gloom-and-doom predictions about job loss largely underestimate how much the economy will grow. The argument stands in stark contrast with what Altman previously said.
"There will be some classes of jobs that will totally go away, but if you are an ambitious person graduating college right now, I think it's probably the best time thus far in history," Altman said.
Have a tip? Contact this reporter via email at scouncil@businessinsider.com, or over text, Signal, Telegram, or WhatsApp at 415-757-8198. Use a personal email address, a nonwork WiFi network, and a nonwork device;here's our guide to sharing information securely.
The post These are the 5 most surprising takeaways from Sam Altman at OpenAI's big developer conference appeared first on Business Insider
