Who will lose their job to artificial intelligence

Summary

Viktor Blaskov opens his podcast Po sredata with the question an ordinary listener asks: will we all lose our jobs to artificial intelligence? His guest, introduced as an engineer, entrepreneur and admin of dev.bg, answers yes: a large share of people will. How fast, and with what effect on society, is hard to predict, and he thinks it unlikely that politicians will stop it with heavy regulation. His company has worked with language models for eight years, since the BERT class, long before ChatGPT.

His picture is a pyramid that turns over. The most valuable people will be those who work with both mind and hands; office work in front of a computer will be hit hardest. First to go are schedules, reception and administration. Legal researchers will first be made faster by the models, then replaced by the next generation of them. Doctors are not immune: imaging diagnostics loses to a system that compares a scan with 100 million others in a second, and pharmacists are at high risk; the surgeon stays. Earlier revolutions closed jobs and opened others: the steam engine took the grooms' work and made work in the mines and on the railways. Artificial intelligence, he says, only closes them.

The second half is about who wins the race. Europe has no AI unicorn and no large AI innovation, he says, only Mistral with mid-quality models, while the US and China compete and Europe works out how to regulate. China can set a doctrine with a 120-year horizon; Western democracies keep turning the political compass, and politics, unlike business, does not dismiss you for repeated wrong decisions. Bulgaria, in his view, has all of Europe's disadvantages and no advantage of its own; its better chance is agriculture and tourism rather than high technology.

Key moments

1. 00:50Yes: a large share of people will lose their jobs, and heavy regulation is unlikely to stop it.
2. 02:01The pyramid turns over: work that joins mind and hands becomes the most valuable.
3. 04:21Legal research: first made faster by the models, then taken over by the next generation.
4. 06:58First to go: schedules, reception, administration. Imaging diagnostics and pharmacists are at risk; the surgeon is not.
5. 08:50The steam engine closed jobs and opened others; artificial intelligence only closes them.
6. 22:06China can plan with a 120-year horizon; Western democracies keep turning the political compass.
7. 23:21Europe has no AI unicorn and no large AI innovation, while the US and China compete.
8. 48:46Bulgaria's better chance: agriculture and tourism rather than high technology.