Altman on mathematicians' five-day mindset reversal; Tao says the era of heroes is ending

We're finished, our field is over, we have to get the graduate students to switch majors.

On October 6, after announcing that it had conquered the Navier–Stokes problem, OpenAI released more than 700 mathematical manuscripts in one go.

Some people couldn't believe it: "This is fake."

Others were furious, believing OpenAI shouldn't have done this: if this is the price of gaining more mathematical knowledge, they would rather not have it.

But Altman found that, among the feedback he heard, some people's thinking had changed by about the fifth day.

From fearing that an entire profession would disappear to rediscovering their own value—what happened in between?

From "We're finished" to reunderstanding their own work

Altman reviewed the reactions he observed in the mathematics community over the past 10 days, calling the process "very interesting."

At first it was disbelief: "This is fake."

By the first day, panic surged: "We're finished, this field is over." Some even began urging graduate students to change majors.

By the third day, fury: "It's terrible for OpenAI to do this."

But around the fifth day, he began to hear other voices.

Some mathematicians came to realize again that their work had always included understanding knowledge, exchanging ideas, and explaining complex results in a way people can understand.

Most people have never solved a Millennium Prize Problem, yet they can still make important contributions.

Once they grasped this, they began to believe:

My career as a mathematician will be different, but hopefully better.

Of course, the "five-day" turnaround period is only Altman's summary of feedback from some mathematicians; it does not mean the entire mathematics community has accepted the change.

He also admitted that models are improving faster than expected.

In narrow fields such as mathematical reasoning, he believes "superintelligence" has already arrived.

But for people who bury themselves in research every day, the feelings brought by this change are far more complicated than the excitement and anticipation of "superintelligence is here" and "the singularity has arrived."

Terence Tao: The "romantic era" is fading

Altman talked about new career possibilities, while Terence Tao spoke of what may be lost in this change.

A person thinks hard for years and finally solves a difficult problem.

In a Numberphile interview released on October 15, Tao, a UCLA mathematics professor who earned his doctorate at age 20, said mathematics is moving from individual heroic assaults to an era of large-scale research.

He brought up the story of Wiles spending seven years conquering Fermat's Last Theorem.

People become enamored with a field of study, and also with the process of exploration within it, and with the satisfaction of finally understanding something after hard thinking.

Beyond finding answers themselves, researchers must also spend time understanding the results machines hand over.

In his "Math 2.0" talk, Tao pointed out that the new understanding and new methods gained during exploration may be more important than the final answer.

Why didn't the original path work? What limitations did the new method break through? What other problems can it be used to solve?

These gains often influence an entire subsequent area of research.

Once a problem is solved, the exploration surrounding it is not over.

If people fixate only on "who hands in the answer first," it is easy to overlook the brilliant ideas and valuable gains in the process of exploration.

If automatic doors are installed, does the doorman lose his value?

In an interview at UCLA, Tao told a hotel metaphor.

A hotel wanted to cut costs and brought in a management consultant.

The consultant believed the doorman's job was simply to open the door. So the hotel installed automatic doors and laid off the doorman.

But the consultant failed to see that the doorman also helped guests, called cabs, watched the entrance, and made people feel at ease. These services also affect guests' experience.

For mathematicians, beyond "solving problems," there is also connecting different fields, training the next generation, and teaching others how to think and solve problems.

Other professions are the same.

When output becomes easy, judging what is worth doing and whether the results are useful still requires human follow-up work.

These follow-up tasks are precisely what Tao bluntly expressed dissatisfaction about in the interview.

In the Numberphile interview, he mentioned that one or two of OpenAI's results, at least as they currently appear, contain quite clever new ideas.

But much of the content is difficult to understand, has not yet been carefully checked, and there have also been cases of errors and retractions.

He said bluntly that there would be good things and bad things in it, but we don't yet know the proportions.

He also complained that at the time there were no OpenAI people helping them understand these materials. In his view, the other side seemed to think the work was done.

After results are released, who explains the key steps? Who answers colleagues' questions? After problems are found, who follows up?

When results pour in in batches and subsequent understanding and verification fail to keep up in time, mathematics has a bit of "indigestion."

On October 15, the Simons Foundation announced a joint commitment with OpenAI and Anthropic totaling $15 million in funding and resources to support mathematicians in checking, understanding, and extending AI results.

Tao also sees new research opportunities.

In the UCLA interview, he envisioned applying one method to tens of thousands of problems and systematically observing where it works and where it fails.

In the past, it was difficult to organize enough people to complete such exploration.

The fascinating way of doing research is changing, and new possibilities are also emerging.

When talking about what students should learn, Altman brought up his own childhood experience.

A teacher asked him to learn writing, and he replied that he didn't want to become a writer, he wanted to be a mathematician.

The teacher told him that writing is also a tool for learning to think and organize thoughts.

Even if you don't grow up to write things for others to read, this training still has value.

Altman used this to emphasize that skills that help people think are still worth learning.

For students about to begin their careers, Altman remains optimistic.

He believes that the most dramatic moments of change can also provide more opportunities for people with ideas, motivation, and new skills.

His advice is not to chase only the directions that were popular five years ago, but to do new things.

But he also admits that a large number of work tasks and some positions will undergo drastic adjustments.

From "We're finished" to searching for direction again, the shift in mindset may take only a few days.

How new work should be done, how contributions should be evaluated, and how newcomers should be trained, however, will take longer to figure out.

The questions facing mathematicians have now been placed before more people:

References:

This article comes from the WeChat public account "Xinzhiyuan" (ID: AI_era), author: Yuanyu, published by 36Kr with authorization.