What Senior Engineers Do Differently: Lindsey Simon of Vercel on Craft in the Agent Era

What actually separates a great software engineer from a good one — especially now that AI agents can write the code? For Lindsey Simon, VP of Engineering at Vercel, the answer has surprisingly little to do with raw coding speed. In a conversation on the Beyond Coding podcast, he lays out what excellence looks like today: unblocking yourself and others, systematizing your help, staying skeptical, and never stepping away from the craft of building.

This post condenses the episode. The views and examples are Simon's.

The One Thing the Best Engineers Do

Ask Simon to picture the strongest engineers in his orbit, and the pattern is immediate: "They unblock themselves, and they're really good at unblocking other people, proactively." The best engineers stay plugged into what's happening across the company — reading, writing, and helping — and they don't stop at solving a problem once. They systematize the fix, leaving a breadcrumb that unblocks the next ten people who hit the same wall.

His favourite example is a Vercel engineer, Shu Ding, who took years of hard-won React expertise — debugging async fetches, handling instant mutations across client and server — and codified it into a reusable "skill." Once distributed, that single artifact scaled his impact to an industry level. This, Simon argues, is the real leverage of the moment: an ecosystem of shareable skills and plugins that lets one person's experience help everyone.

Everyone Is a Junior Engineer Again

Simon is candid that the disruption unsettles even veterans. "We're all junior engineers again," he says — these tools haven't been out longer than any of us have been in the industry, so the FOMO is universal. The temptation is to hide out of fear, but "the only way out is through." What pulled people into building in the first place — the art of crafting things you'd want to use and share — hasn't gone anywhere, and he says he's more motivated to build now than before.

He doesn't see roles disappearing so much as consolidating. Agents make it easy to work in areas where you're less trained, which dissolves the rigid lanes between engineering manager, product manager, and designer. He points to the valley-wide trend of everyone becoming a "member of the technical staff" — partly anti-poaching cover, but also a genuine flattening. You don't have to be a designer by trade, he notes, to do the act of design.

Level Doesn't Matter — Impact Does

That flattening extends to hierarchy, and Simon sees it as one of tech's most attractive qualities. At Vercel, interns present at Friday demo days in front of the founder, the CTO, and the entire engineering team — and in that room, everyone is equal. "It doesn't actually matter what your level is," he says. "You could impact the business if you figure out the thing, if you spend the time."

Far from worrying about junior engineers, he'd happily hire them. They've grown up builder-minded, unburdened by the belief that you need specific expertise before attempting anything — bold enough to think, "I don't know anything, so I may as well do everything." A new grad with four relevant years of hackathons and internships, he argues, can be better prepared than someone with six years of less-relevant experience.

Ask the Agent First — Then Ask a Human Better

One of the sharpest shifts Simon describes is in how people ask for help. Traditional education trains you to figure things out alone; the workplace rewards knowing when to ask. Now there's a middle step: ask an agent first. When its answer looks fuzzy or generic, you take a much better question to an experienced colleague — one that's grounded because you've already engaged with the problem.

That reframes asking for help entirely. Instead of "I'm stuck," it becomes "I tried to unblock myself and I'm not satisfied with the answer, and here's why." People on the receiving end appreciate it, Simon says, because it signals you did the work first. And the deeper skill underneath is questioning: repeatedly asking "why is that the answer?" and digging backwards until you genuinely understand the problem space. He likens it to evaluating search sources — a skill his ten-year-old daughter is now taught in school — and insists that trusting AI output has to be paired with staying skeptical and validating with peers.

Careers, Fulfillment, and Feedback

Simon is refreshingly honest about his own tensions. Vercel deliberately keeps the individual-contributor and management tracks parallel, with the top IC level as high as the top management level, so no one hits a ceiling by staying hands-on. He never wanted to manage — he calls himself "unmanageable" — and notes that the people who lead and innovate tend to be the ones still building, connected to the iteration cycle and talking to customers.

He admits his biggest FOMO now is being pulled away from daily practice, since for most of his career his lever was pure PR throughput. Take that away, and the question becomes where you find fulfillment. For him it's the company's mission and the feeling of enablement — the pride of watching five people feel good about work you helped them do, or seeing a company build something great on the platform and thinking "we helped them do this." He even muses that he wouldn't mind, one day, becoming a teacher who helps people learn how to learn.

That instinct shapes how he treats feedback. Building for famously fickle engineers, he trusts both social signal and telemetry, because people "vote with their feet and with what saves them time." His attitude toward harsh feedback is generous: "if they're not wrong, nobody's wrong here — if we think they're wrong, then we didn't communicate it well." And he describes watching a user struggle silently through a product as "knuckle-whitening" — and the single most motivating thing there is.

The Takeaway: Fundamentals, Writing, and Getting Out of the Cave

Simon's advice for anyone trying to level up is almost old-fashioned. Go back and close the knowledge gaps you skated past — the "85 on a test" that quietly grows into splinters — so you can build on stronger foundations and move faster. He calls for a resurgence of liberal-arts education: "if English is the most important programming language, you better learn how to write and write effectively," because distilling information is a practiced art that matters even more when you can burn a lot of tokens getting nowhere.

His single biggest takeaway comes back to feedback. Don't disappear "into the cave," Socratic-style, hoping to emerge with an earth-changing thing. Great practitioners develop by constantly getting feedback from more people and learning to distill the signal. And, he adds, learn to write about what you're doing — it forces you to turn experiments into concise descriptions, and that turns out to be genuinely interesting to everyone else.

---

Originally published on Beyond Coding. Watch the full episode: https://www.youtube.com/watch?v=mlvbbqYtUz0