Data Fusion: six shifts that are changing the profession of the scientist

At the Data Fusion conference we took part in a discussion of how AI is changing the profession of the scientist, along with the publication crisis, the closing of technology and the new logic of a career in science.

The session identified six shifts that are already changing the rules of the game:

• the flow of papers has become so large that following it is physically impossible; the mark of quality is now the name of the laboratory rather than the journal or the conference;

• frontier companies (OpenAI, Anthropic, DeepMind) publish results but do not disclose architectures and methods — even researchers no longer know what exactly makes the models better;

• mathematics remains an exception: one of the few fields where a scientist can work for the sake of knowledge itself, without an applied commission;

• business now enters research at early stages instead of waiting for a finished result;

• the new survival strategy is to look for areas at the intersection of disciplines that LLMs have not yet reached. As Ilya Makarov noted, the synergy between an AI specialist and an expert from another field delivers far more impact than "squeezing metrics out of one small model";

• the researcher becomes an asset for direct investment — the academic path is no longer the only one.

If you are doing something useful, it should not be about understanding — it should be about the fact that some day, in a year, in three, in ten, it converts into a tangible result.

Full write-up of the session: kod.ru