Trying to become an expert on every new development in artificial intelligence may be the wrong race to run.

Jeff Dean, the longtime Google researcher who recently left the company after 27 years, has a different prescription for students and young professionals entering technology: cover more ground, learn what is possible across different areas and become good at spotting connections that others have yet to make.

Speaking at the Asian American Scholar Forum’s 2026 Frontier & Pioneer Symposium in his first public appearance since leaving Google, Dean argued that breadth can sometimes be more useful than immediately going deep into a single piece of research.

His advice was unusually specific. Rather than spending all available time studying one paper in detail, Dean said students could gain more from skimming 10 papers. He went further, suggesting that even working through 100 abstracts can help build a broader picture of ideas, methods and possibilities.

The point is not to understand less. It is to create a bigger mental map before deciding where deeper attention is worth spending.

Why Jeff Dean thinks breadth matters in the AI era

Artificial intelligence is developing across several scientific and technical disciplines at once. For someone starting a career, keeping up with every model, paper or new technique can quickly become unrealistic.

Dean’s approach turns that problem on its head. Instead of attempting to master everything, young technologists can expose themselves to a wider range of ideas and then look for combinations that have not been explored.

That ability to connect seemingly separate concepts, he suggested, can open routes into problems that previously looked difficult or even unsolvable.

There is also a practical lesson about choosing which problems deserve years of work.

Dean cautioned against projects that may require 20 years without a convincing idea for how to solve them. At the opposite end, a problem likely to be finished within two years could be too straightforward to produce the kind of breakthrough a researcher is hoping for.

His preferred territory sits somewhere in between.

Dean described roughly five years as an attractive horizon for a serious long-term problem, giving researchers enough room to experiment, fail and try different approaches without committing themselves to something with no visible path forward.

That five-year window may be the more important part of his advice. In a field often obsessed with what will change next month, Dean is encouraging younger researchers to think on a longer clock.

From Google chief scientist to DiscoveryLoop

Dean’s comments come at a major transition in his own career.

He left Google earlier this month after spending 27 years at the company. He led Google AI from 2018 until 2023 and then served as Google’s chief scientist from 2023 to 2026.

He is now cofounder and CEO of DiscoveryLoop, an artificial intelligence company focused on speeding up scientific and engineering discovery.

That next chapter closely matches the philosophy he outlined to students. Dean sees AI not simply as a technology to be learned, but as a tool that can help people tackle problems they previously lacked the expertise or resources to address.

AI could make specialist knowledge far more accessible

Dean remains strongly optimistic about what AI can do despite concerns surrounding employment, inequality and the wider effects of increasingly capable systems.

He pointed to healthcare and education among the areas where AI could help people do more than they could on their own.

One reason for that optimism is the possibility of putting highly specialized knowledge inside models that can work across many disciplines.

Dean said models capable of understanding numerous areas of science and engineering could provide what he described as “Ph.D.-level expertise” across multiple fields.

For researchers, engineers and students, that could change what it means to enter an unfamiliar discipline. A person may no longer need years of formal specialization before gaining useful access to advanced knowledge in another field.

Dean did not present AI as a guaranteed route to success. He acknowledged that he regularly runs into failure himself and does not have a “magic answer” for every problem.

His answer instead appears to be experimentation: learn widely, identify worthwhile problems, try several approaches and accept that many will not work.

A different career strategy for Gen Z

The larger message for Gen Z is less about reading papers quickly and more about deciding what kind of knowledge will matter when information is abundant.

If AI increasingly makes technical expertise easier to access, memorizing every development may offer less of an advantage. Knowing which ideas belong together, which questions deserve attention and where to invest several years of effort could become more valuable.

That makes Dean’s advice a notable departure from the pressure many young professionals feel to constantly keep pace with AI.

His formula is simpler: see more ideas, connect more dots and choose problems with enough difficulty to matter, but enough visibility to make progress possible.

In an industry changing almost daily, Dean is effectively telling the next generation not to chase every update.

He wants them to widen the field of view first.

Also Read: Orange Health revenue jumps 65% in FY26, but it spent Rs 1.73 to earn every rupee