A Caltech Professor Says AI Cannot Exchange This One Talent

Certainly one of AI’s main researchers has a easy piece of profession recommendation for younger folks anxious about future-proof abilities within the ChatGPT period: be curious.

“I feel one job that won’t get replaced by AI is the flexibility to be curious and go after arduous issues,” Anima Anandkumar, a professor on the California Institute of Expertise, stated in an interview with EO Studio that aired on Monday.

“So for younger folks, my recommendation is to not be afraid of AI or fear what abilities to be taught that AI could change them with, however actually be in that path of curiosity,” Anandkumar added.

Anandkumar, a former senior director of Nvidia’s AI analysis and principal scientist at Amazon Net Providers, left the non-public sector in 2023 to return full time to academia. She has served because the Bren Professor within the pc science and arithmetic division at Caltech since 2017.

“I am unable to think about a world the place scientists might be out of jobs,” Anandkumar, who beforehand helped construct an AI-based climate mannequin, added. “A scientist tackles open issues — from subatomic matter to galaxies — and there is an limitless listing of these.”

She additionally stated that whereas labs like Google’s DeepMind are exploring “AI scientist” fashions, she believes the true limitation is sensible validation, not an absence of concepts.

Nonetheless, she’s skeptical of the hype round absolutely autonomous AI scientists.

“The bottleneck goes to the lab or going to the true world and testing them. That’s sluggish, that’s costly,” she stated.

Coding is altering, however nice programmers nonetheless win

Anandkumar additionally shared profession recommendation for these in software program improvement, which is being considerably disrupted by AI.

“A foul programmer who will not be higher than AI might be changed,” she stated. “However a fantastic programmer who can assess what AI is doing, make fixes, [and] guarantee these packages are written properly might be in additional demand than ever.”

Her level echoes what OpenAI CEO Sam Altman stated in March: college students ought to “get actually good at utilizing AI instruments” as fashions more and more take over routine code technology.

New graduates are feeling the stress, although. A 2025 Handshake survey of over 3,000 school seniors discovered that 62% of these accustomed to AI instruments stated they had been anxious about how these instruments may have an effect on their careers, up from 44% the 12 months earlier than. Amongst pc science college students, 28% described themselves as “very pessimistic” about their job prospects, citing shrinking openings and fiercer competitors. Job postings fell 15%, whereas purposes jumped 21%.

In the meantime, some tech leaders are overtly sounding the alarm. Victor Lazarte, a companion at funding agency Benchmark, not too long ago warned that AI is already changing employees, and stated attorneys and recruiters needs to be particularly involved.

Anandkumar, in contrast, stresses that the important thing benefit nonetheless lies with people who information the programs.

“You could have the company as a human to determine what duties AI does, and then you definately’re evaluating and also you’re in cost,” she stated.

“Do not be afraid of AI,” she added. “Use it as a software to drive that curiosity, be taught new abilities, new information — and do it in a way more interactive approach.”