Some of the posts in the blog contribution avalanche on Terry Tao’s blog are so out of touch with the mathematics spirit that it is almost funny. An example is “egoless mathematics in the AI era”. As comments have pointed out, it is most likely written by a machine. Nice try, openAI or Antropic. The title looks good because “ego” is tainted in a negative way like “egoistic” or “egocentric”. Having therefore an “ego free mathematics” can be seen at first as a good thing. Take that post and replace “ego” with “human” and you see the plot. Why is taking away “ego” important. I myself feel pride also if an other human achieves something great. A beautiful proof done by a fellow human is a reason for celebration. Having it done by a machine is not the same. The article talk about mathematical discoveries and states “its worth is not diminished by the absence of our authorship.” I (as a human) strongly disagree with that. Human less mathematics is by nature soul less (but this is hard to define and quantify) If I read a novel and learn that it is machine generated, I myself lose interest. Having a machine as an author is playing in a different category. It is psychological maybe but I need to have a book, music or mathematical paper tied to a human to be valuable.
My takeaway is (and that might sound insulting) is that anybody who applauds that machines take over creative tasks, has not been a creative person after all. They had envied that ability and now are happy to see it becoming less valuable. A writer, musician, engineer or mathematician who has generated something and fought for it over a long period of time laments naturally that there are now tools which generate things almost effortlessly. It is maybe similar to a mountain climber who liked to climb a route just to see it filled with ladders and having thousands of tourists jamming it. It is not “ego” or “exclusivity” but “human pride” and “accomplishment”. The “ego-less mathematics” blog post mentioned above completely misses the point. I’m not the only person who thinks this article is a a clever hack by a machine to infiltrate our creative domain. Unfortunately the process we are in is irreversible. We have to live with it. I myself try to keep parts of my own most dear work shielded away from the machines and if a project needs machine help, acknowledge the help. An example were I myself got assisted was this paper on natural groups, where I needed help to getting ironed out an earlier mistake in a paper from 2022. The machine was perfect in finding the right references to fix the issue in the definition. There is no question that I could not have done that in such a short time myself. But I have more joy with struggling through a problem on my own, even if it is much slower and even if I make tons of mistakes along the way.
An other argument by the machine is that “now everybody can be creative”. We have taken away the exclusivity and made it for everybody. Isn’t this nice? The machines have produced “access”, where very few could go before. Its like building a street to a remote beach that had been exclusive for a few and now the tourists can come in with cars. Many tourist attractions have seen this dilemma. It is nice to be accessible and cheap, but it also diminishes the value of the place. Because by nature, it is no more so exclusive. An other allegory is climbing the Mount Everest. There had been a famous picture in the Natural Geographic showing how overrun this mountain has become. Is it an achievement to have gone up? No more. It is a matter of money to buy all the equipment and guides and permits and bit of luck with the weather. Going up the Everest is still a bit of a risk. People die. But nobody risks the life by entering a prompt to a machine. A monkey can pass through an online education system these days. One just has to train it to click one button starting the agent which takes any assignment posted and gets a solution and submits the solution. Similarly with writing a 100 papers a day: start an agent that looks online for open problems, submits it to a swarm of other agents and collects all the papers written in perfect latex format, then submit them to the ArXiv (or github if the ArXiv is not an option), as Open AI did on October 6th. Of course that needs money but there are folks who have the cash to do it. Should they get the fields medal for it? Of course not.
One can argue that the new tools can accelerate progress. But cheap progress is not really progress. It can have a destructive effect. Motivation is an important force of power. It is like generating plastic roses. They might look good, even indistinguishable from real roses, but it one can use other senses to see that there is no soul behind it. They do not smell, they do not sting and even if this would be generated by machines too, they have not been seen growing. There is something amazing of growing a plant by your self. I myself made some experiments with that. An apple tree which lived for a few years in our garden died last year. I planted a new one. Whether it will survive is not clear. This uncertainty adds value. We grow a lemon tree since more than 10 years. It needs to be taken care of, it needs to be watered, it needs to be replanted, it needs to be taken in during winters etc. It produces maybe a handful of lemons every year. Buying lemons in the shop is not the same. A lemon you have seen grow from a small flower to a full grown part is infinitely more valuable. The following picture is done by a human. This fact makes it valuable. It is not a matter of “ego”. I did not paint it. It is valuable because a human in 1908 (John William Waterhouse) has created it.
I now, a month after the October 5 event actually, applaud open AI to post the dump. We were actually asked by the advisory group on mathematics what to do. I wrote in that feedback form that any result that has been found by Open AI should actually be posted. it would be devastating for the math community to know that “the labs” have solutions to problems which are not published. It would demotivate work on problems even more because what’s the point to work on a problem which the machine has already solved and hidden away? The dump has showed the world what is possible. It would have been more painful, if some rogue people would have slowly done the proofs with the machine and cheated their way to fame claiming that they have done it on their own. Now we all know what is possible. Still in the summer, there had been “experts” claiming that research mathematics can not be done by the machine. Things go so fast that there are simply no “experts” in the matter. I myself had not been worried in the spring of 2026. The summer 2026 came with some shocks but they could be shelved away, like the solution of the Jacobian conjecture (well it is just “search” for an example) or the Navier Stokes event (well it is just “search for an example). The October 6 dump was completely different as there were many proofs which can not be just put down as “search”. They involved some creativity. We have been invaded by an other species that rivals our creativity. And that is of concern for humans. It has nothing to do with “egoism”. The new species attacks domains which have been considered safe places for humanity.
