"Everybody remembers the first half of the jack-of-all-trades video. Everybody know the second half? And master? No, that's right." Idan Gazit from GitHub Next dropped this line at the Berkeley Haas AI Conference, and it crystallized something I've been thinking about for a while.
For years, the career advice was: specialize. Pick a lane. Go deep. Become the world's foremost expert in one narrow thing. That advice made sense in a world where execution was expensive and expertise was the bottleneck.
That world is ending.
The Specialist Advantage Is Shrinking
Here's the uncomfortable math: AI tools are getting better at deep, specialized work every month. Code generation, data analysis, design iteration, legal research, financial modeling — these were all specialist domains that required years of deep practice. Now an AI agent can do passable-to-good work in any of them.
This doesn't mean specialists are irrelevant. The best specialist will always outperform the best generalist in their specific domain. But the gap is shrinking, and the cost of accessing specialist-level work is collapsing.
What's NOT getting automated: the ability to see across domains. The judgment to know which specialist tool to apply to which problem. The taste to recognize when an AI's output is 90% there and what the missing 10% is. The communication skills to translate between engineering, design, business, and customers.
Rocket Boots for Generalists
Idan put it perfectly: "The ability to say I am self-capable of going from zero to one — today with AI, it's like wearing rocket boots. Hybridity was undervalued in the past. Now it's actually a superpower."
Before AI tools, being a hybrid was a disadvantage in hiring. You'd interview against specialists who could go deeper in their one area. Now, the hybrid who can use AI to execute at a competent level across multiple domains — and who has the judgment to orchestrate those domains into a coherent product — is more valuable than the specialist who only knows one thing really well.
GitHub Next explicitly hires for this: "I'm hiring makers of a lot of different skill sets. Every one of those people tends to be a hybrid. Somebody who, when I ask them 'what are you good at?' they're like, 'well, I'm kind of good at this, but also that.'"
The Webflow CPO Agrees
Rachel Wolin at Webflow echoed this from a completely different angle. She started her career as a software engineer, transitioned to product management at Haas, and now runs product at a company with hundreds of employees. She still writes code. She still builds side projects. She calls herself "a builder first and an executive second."
Her framework: "Can I scale myself up? Can I scale myself down?" The market rewards people who can do both — who can think strategically at the executive level AND get hands-on with the actual work. That's the hybrid advantage.
At their recent offsite, 86 people at Webflow published repos and prototypes. The message wasn't "engineers should build" — it was "everyone should build." Product managers, designers, marketers — all shipping code.
What Makes a Valuable Hybrid
Based on what I heard across the conference, the most valuable hybrids share a few characteristics:
1. They can go from zero to one alone.
Not zero to scale — just zero to one. A working prototype. A functional MVP. Something people can touch and react to. AI tools make this dramatically easier, but you still need the instinct to know what to build.
2. They speak multiple languages.
Not programming languages — organizational languages. They can talk to engineers about trade-offs, to designers about user experience, to executives about business impact, and to customers about their problems. Each of those conversations requires a different vocabulary and different empathy.
3. They have taste across domains.
They know when a design is almost right. They know when code architecture will cause problems later. They know when a business model has a hole in it. This cross-domain taste only comes from spending time in each domain — even if you never become the best at any one of them.
4. They learn fast.
The Snowflake and Handshake panel emphasized curiosity as a durable skill. The hybrid advantage multiplied by AI tools only works if you're constantly learning new tools, new domains, new ways of working. The people who downloaded Cursor the weekend it launched have a different skill set than people who are still thinking about trying it.
The Personal Angle
This resonates because it describes my own path. I'm not the best engineer, the best designer, or the best strategist. But I can build a full product — from concept to code to deployment — because I've spent time in each of those domains. Every project in my portfolio exists because I could see across the boundaries between disciplines and build something coherent.
In the pre-AI era, that made me a generalist in a world that rewarded specialists. In the AI era, it makes me the kind of person these companies say they're hiring: someone who can go from zero to one, who can orchestrate across domains, and who treats building as a daily practice rather than someone else's job.
The age of "pick a lane" is over. Pick all the lanes. AI will handle the depth. Your job is the breadth.