ChatMIT: Sam Altman on Start-ups, GPT, and the AI Revolution
Altman: “This is probably the most exciting time to be launching your career in many decades.”
Photo Courtesy of Jared Chaney
On May 2, over 1,000 members of the MIT community sat in Kresge’s Swraj Paul Theater to hear from Sam Altman, co-founder and CEO of OpenAI in an event called ChatMIT hosted by President Sally Kornbluth.
Kornbluth posed many questions surrounding AI, including mitigating bias in training models, protecting user privacy, and optimizing energy usage. Altman also discussed the potential of AI to reshape the workforce and revolutionize scientific discovery.
OpenAI has made “surprisingly good progress,” Altman said, in learning how to make a system behave according to a certain set of values. He added, “Humans are clearly very biased creatures,” whereas GPT-4 does not share “our same psychological flaws.”
Regarding data usage, Altman said, “I can imagine this future in which, if you want, you have a personalized AI that knows that has read every email, every text message you've ever sent or received, has access to a full recording of your life.”
“You can imagine that would be super helpful. You can also imagine the privacy concerns that that would present,” he added. “That's a new thing for society to navigate. I don't know what the answer will be.”
Altman stated that soon after GPT-4’s release, everyone wanted to know, “‘Where’s GPT-5?’” and the expectation for perpetual progress was “legitimately great.” A major improvement to the current GPT model, Altman envisions, is to separate the storage of data from the training done on that data. To memorize the data it trains on is, according to Altman, “a weird waste of resources.” A more efficient GPT would partition reasoning from memory.
Particularly resonant with the students in the audience was Altman’s advice about taking part in this technology revolution.
“This is probably the most exciting time to be launching your career in many decades,” Altman stated. He described the internet boom and subsequent smaller waves in mobile and cloud technologies to emphasize how the world has been waiting for almost a decade. Now, as Altman sees it, AI is next up.
“The ground is shaking,” Altman said, encouraging creation of start-ups, which have a number of advantages over big companies. “You can take more risks than usual, and you can have more impact than normal,” he added. “You can move much faster. You can live in the future... And that's how you win.”
Altman is under no illusions about how AI will change the workforce, including eliminating, altering, and creating jobs. “That always happens with technology,” he stated, though, “It's probably never happened this fast.” Naturally, Altman is biased towards a career involving AI.
“The most important lesson to learn early on in your career is that you can figure anything out,” he stated. “You stumble your way through it, have a fast iteration speed, try to drift towards the most interesting problems to be around, the most impressive people and have this trust that you will successfully iterate to the right thing.”
Altman took a strong stance against those who believe progress in AI should be halted because of issues such as energy usage or bias. “It is true that AI needs a huge amount of energy, but not huge relative to what the rest of the world needs,” Altman said. He provided the example of Google, which consumes an incredible amount of energy but also saves it by improving access to information, thereby reducing extraneous costs that would be expended in the lack of such knowledge. It is actually a “net savings in energy,” he stated.
Altman plans to continue figuring out ways to make AI more energetically efficient and argues that it could help yield insight into harnessing non-carbon based energy, which would be a “massive win.”
“The way we are teaching our young people that the world is totally screwed, that it's hopeless to try to solve problems, that all we can do is stay in our bedrooms in the dark, is a really deeply unproductive streak,” Altman said. “You all need to make it part of your life mission to fight against this,” he added.
Using GPT for scientific discovery and personalized learning experiences are among the benefits that excite Altman the most, but he believes it is just the beginning. “Babies born today will never know a world in which the products and services that you use are not intelligent,” and cognition will be omnipresent in day-to-day life.
“We have a ton of work in front of us,” Altman stated. “We have a new tool in the tech tree of humanity, and people are using it to create amazing things. They will continue to get way more capable and way more prominence over time. It's going to integrate into society in an important and transformative way.”