The End of the Specialist

A consistent theme across tech leaders is the dawn of the "Generalist Renaissance". AI is accelerating a massive shift from specialist value to generalist value, where the most highly sought-after talent is no longer "T-shaped" (deep in a single area), but "block-shaped"—deeply capable across multiple domains.

Because AI dramatically compresses the skill acquisition curve, wearing many different hats is now the default expectation in flat teams, a sentiment explicitly echoed by Grant Lee of Gamma. Startup founders are increasingly obsessed with packing as many skill sets as possible into a single person. For example, Dan Shipper of Every runs five products with a 15-person team using 100% AI-written code. Shipper compares his team to the citizens of Ancient Athens, who were expected to be fighters, judges, and generals—acting as generalists by necessity. Similarly, Anton Osika of Lovable, who achieved $10M ARR in 60 days with roughly 15 people, notes that if he is putting together a team today, he obsesses "about getting as many skill sets as possible per person".

In flatter organizations, the surface area of responsibility for every role is expanding, and AI agents are the mechanism compressing the time required to manage that scope.

• Product Managers: In his highly popular Make PM fun again with AI agents newsletter, Tal Raviv points out that as tech orgs flatten, more repetitive tasks are falling on fewer PMs. To survive, PMs are leveraging proactive AI agents to scan calendars, aggregate user feedback across support tickets and Slack, and monitor competitor websites on autopilot. Consequently, PMs are now expected to own research synthesis directly, relying on agents for data aggregation while the PM handles human interpretation and judgment. However, as noted in a podcast with Aishwarya Reganti and Kiriti Badam, PMs must also develop enough technical fluency to actively direct and set constraints for these agentic systems.
• Designers: Design leader Jenny Wen recently shared that she is actively hiring for "strong generalists" who fit this "block-shaped" archetype. She notes that while tools like Claude aren't replacing strong designers yet, they are incredibly useful for rapid first passes and generating divergent options. Moving forward, designers who can execute light PM work, write effective prompts, and synthesize their own user research will be far more valuable.
• Developers: Writing code is no longer the sole mandate. Boris Cherny of Anthropic advises that engineers who want to succeed in this new era must be curious, cross multiple disciplines, and develop strong product and design sensibilities rather than just focusing on the engineering piece of the puzzle.

Perhaps the most heavily impacted specialty is user research. Dedicated user researcher roles are finding themselves under intense pressure, with their responsibilities heavily dispersed among PMs and designers.

Teams are increasingly utilizing synthetic data and AI-generated user personas as a "first filter" to pressure-test hypotheses before spending time and money on real human interviews. PMs and designers can now use tools like DataDisco as a research workforce to get usability feedback, without the time and cost of recruiting real participants.