- OpenAI will keep pushing AI’s scientific discoveries, Sam Altman says
- Some researchers are excited by the ‘fantastic’ breakthroughs
- But others warn that the pace of change is ‘psychologically taxing’
One of the promises that huge AI companies like OpenAI and Anthropic have made is that their machine learning models will help to solve some of the oldest problems in math and science. The pace of their discoveries has left many experts struggling to keep up, but OpenAI’s Sam Altman just has just said there’s much more in store — and not everyone is happy about that.
Speaking on the Cheeky Pint podcast, Altman outlined his belief that “Humanity deserves fundamental knowledge, and we should not be sitting on a major fundamental science discovery.”
That rationale is why OpenAI has published breakthrough after breakthrough in these fields, knocking down some of the oldest and toughest problems that have troubled great thinkers for years. But the way in which the company has set out its results has drawn criticism for often inadequate attributions and questionable verification processes.
While Altman acknowledged that, saying on the podcast that “I do have empathy for the method” and saying OpenAI’s quest for mathematical challenges did “get out of control” at times, he suggested that there’s much more to come. Asked if the firm has anymore unpublished breakthroughs in its vaults, he replied “We do. Of course.”
That’s led to accusations that, far from giving people “fundamental knowledge,” OpenAI is gatekeeping important discoveries. And there’s another problem: the sheer scale of the new research is swamping academics. For instance, Wired has noted that the company recently “dumped 722 manuscripts on 400 problems on GitHub,” leaving the scientific community to pick through the enormous mass of information.
And the impact of having the most important work in your life suddenly undercut by a machine is “psychologically taxing for many people,” said Thore Husfeldt of the IT University of Copenhagen. Husfeldt added that there is “a lot of doom and gloom, because every week, or every second week, the results announced by AI just annihilate the ongoing research projects of many people.”
Obliterating entire fields of research
AI models’ ability to solve thorny problems was keenly demonstrated when it apparently disproved the exponential time hypothesis (ETH), which argued that the effort needed to solve a problem increases with its complexity. That’s upended countless theories that assumed that ETH was correct and left many experts scrambling to catch up.
As Linköping University professor Peter Jonsson told Wired, although AI’s tendency to undermine years-long research can be irksome, “at the same time, we learn a lot from it. And this is what science is about.”
UC Santa Barbara professor Daniel Lokshtanov aptly described the mood among many scientists and mathematicians, saying “It’s just a giant bomb dropped into the middle of the field,” he says. “It’s really cool and terrible at the same time.”
For Husfeldt, meanwhile, the discovery is “fantastic.” He added: “Being a mathematician is a lot like stumbling about in a completely dark room. It’s like somebody just illuminated an entire subfield in floodlights of knowledge, and now we just run around like overexcited bunnies and look at all the nooks and crannies that we can suddenly explore. So it’s a super exciting time.”
In the case of the ETH theory, OpenAI’s work has not been peer reviewed or independently verified. But it’s clear that AI research is creating generational changes in many areas — and overturning entire fields of study in an instant, with all the excitement and chaos that comes with that.
Jonsson fittingly summed the situation up, saying: “I don’t care if it’s a human-written proof, or it’s written by AI, or if an alien transmitted this proof to us. If it’s true, then we must simply use it in the future.”
