A former engineer at Meta and OpenAI says there's danger in specializing too early

Business Insider

25-05-2025

Early-career engineers should beware of specializing too soon, said Philip Su, a former engineer at Microsoft, Meta, and OpenAI, and founder of podcast player Superphonic.
"That, I think, is a tricky decision depending on how well the person knows themselves," Su said on a recent episode of " The Developing Dev" podcast when asked if it was better to be a generalist or to pick a niche and stick to it.
"So there's the occasional exceptional person — like these prodigies in chess, for instance, right?" he added. "They will have been a prodigy by the time they're eight or nine years old, and so they're obviously fit to play chess. That person should specialize, because that's an unnaturally unique talent, right?"
For most, there are dangers to singular focus, Su said — especially in the " age of AI." There's always the possibility, he said, that your specialization is rendered obsolete.
"If you join some company and you're diehard committed to like, Technology A, right?" Su said. "What if in three years that thing becomes irrelevant, and that's all you know? You know, you're like the COBOL person hoping Y2K happens again, right? Because COBOL's not used anywhere, but that's your specialty."
Before you choose to completely dedicate yourself to any one area, Su suggested taking a few years to develop a range of skills and to determine what best suits you.
"If you are 22, 23, starting your career, I would, in general, encourage at least dabble in a few things before you like diehard commit," he said.
Figuring out what's right for you is easiest when you're sure of what you want, Su said, not just in work, but in life.
"Decisions, for me, a lot of times were hard because I didn't have clear values," he said. "If you know exactly where you're going, decisions toward getting there become a lot easier."
If Su could give advice to his younger self, he added, he'd tell him to take more time to really pinpoint his desires, rather than forging ahead toward an idealized goal.
"I think another thing is, I often feel like I was the dog that caught the car," he said of becoming a development manager at Microsoft, where he worked prior to OpenAI and Meta.
"The problem with peaking early, you know — because I hit that level when I was probably, I don't know, 30 years old or something like this — the problem is, you're like a child actor," Su added. "The question is, what are you going to do with the rest of your life?"
In addition to making sure you truly want what you're chasing, if you're particularly focused on your career to the exclusion of all else, Su said you should be prepared to make sacrifices.
"So A: be sure that's really what you want. And part B is, be sure you're comfortable with other things breaking, you know?" he said. "Because that is what it will take to get there, if that's truly what you want."
Some reeds "bend," Su said, while others "break" completely — so it's worth evaluating your priorities with great care. In the grand scheme of a career, he added, there's ultimately not that much of a difference between becoming a senior engineer at 30 versus 38.
"So it's like, how fast do I want to be at my terminal level? Like, what's the real plan there?" Su said. "Versus, can I keep a healthy relationship with my spouse, with my kids, right? That's important."