Happy New Year. I think you’re allowed to keep saying that until Valentine’s Day. Let me know how you go with that…
Anyway, so far this year, I’ve been doing my usual thing, listening to long-form interviews while half-working and half-thinking, then trying to pull out what actually matters before the hype cycle eats it.
This week, it was Lenny’s Podcast on YouTube, and his interview with Zevi Arnovitz has been stuck in my head for one reason: it’s one of the clearest examples I’ve seen of what “AI changes work” actually looks like in practice.
Zevi’s a PM at Meta, previously at Wix, and he’s very openly non-technical. He says he has “zero technical background”, he did music in high school, he wasn’t in a tech unit, he’s the opposite of the typical “I’ve been coding since I was 12” story. But, he’s building real products anyway, using Cursor with Claude Code, with a workflow that’s surprisingly disciplined for someone who describes code as terrifying.
The quote that anchors the whole thing is the one we’ve all heard, but most people still treat it like a motivational poster.
“It’s not that you’ll be replaced by AI. You’ll be replaced by someone who’s better at using AI than you.”
I don’t love fear-based takes, but this one is useful because I think it reframes the competition. It’s not AI vs humans, it’s humans-with-AI vs humans-without-AI.
The part worth paying attention to for me isn’t “AI can code.” It’s that building is becoming a workflow you can learn, even if you’re non-technical, as long as you’re willing to be structured and ruthless about review.
Zevi describes watching a YouTube video where someone was building apps with AI tools like Bolt or Lovable, and it felt like someone walked up and said, “you have superpowers now.” That’s the feeling a lot of people are having, not because they’ve suddenly become engineers, but because the bottleneck has shifted from “can I write code” to “can I think clearly, plan properly, and check what’s been built.”
His workflow is basically a compressed version of how strong teams already operate. Capture an idea quickly, explore it properly, create a plan, execute the plan, review the code, run peer review, update documentation. It’s not sexy. That’s why it works. Most people skip the explore and plan steps, then act surprised when the output is sloppy or the build collapses under its own weight.
The most CTO-relevant point in the interview is this: writing code is now the easy part, reviewing the code AI wrote is the hard part. Zevi’s solution is simple. He manually QA’s what he can, then has the model review its own work, then has other models review it, then uses a “peer review” command that forces the primary model to either fix issues or explain why they’re not issues. He’s basically creating a mini panel of reviewers and making them argue, then he decides what to do next. You can use the exact same idea for strategy docs, comms, hiring plans, anything where first drafts are cheap and quality comes from judgement.
Lenny pushes on the fear that a lot of people have, that using AI is “outsourcing your thinking” and your skills will atrophy.
Zevi’s response is interesting to me. AI only makes you worse if you use it to avoid owning the output. If you show up to a product review and blame the machine, that’s not modern, it’s just abdication. The responsibility still sits with you, and in some ways it’s heavier now because AI lets you ship faster, which means you can ship mistakes faster too.
The most hopeful angle is how many doors this opens for people who aren’t engineers. Zevi talks about starting with ChatGPT/Claude projects because it’s less intimidating than jumping straight into an IDE, he calls it “exposure therapy.” He also talks about using AI to prep for interviews, building a “coach” project, pulling in frameworks, running mock interviews, even building small practice tools for weak spots. You can argue junior roles are changing, and they are, but there’s also a new path forming that didn’t really exist in the same way even a couple of years ago. You can learn faster, get reps faster, build something real, and show evidence of ability instead of just claiming you have it.
If you take one thing into the year, make it this. Try picking one part of your work where you’re currently slow or blocked, planning, writing, analysis, prototyping, interview prep, whatever, and make it AI-native in a way that still forces you to own the result.
Now, if this sort of thing is what you are remotely interested in, then you may like the sound of the next meetup that I have planned.
The Product Talk Meetup isn’t just a Meetup for product managers. It’s for people who are working in product-based businesses and whose work touches a customer or a customer experience in some way.
The first Meetup is next month, I’d love to see you there and meet you IRL if we haven’t already. This is what you can expect.
“AI talk is everywhere right now. This session focuses on real-world applications, workflows, constraints, and outcomes. So yes, Product Talk is doing an AI night too, but with Actual Intelligence, and real operators who’ve built real things.
Two speakers, two different worlds, both with outcomes.
Jared Korinko (Product @ Google, Google Photos) is shipping for 1B+ users, and will share how he built a revenue-generating side hustle purely with AI, the approach, the tools, and the parts that didn’t work as advertised.
Joel Kenyon (FMR Head of Product, Amused Group) has led the design and delivery of a wagering platform powering 10 brands, and has embedded GenAI into product workflows to accelerate validation and roadmap intelligence. He’ll talk about what AI looks like when it’s part of a real product organisation, with constraints, compliance, and customers who do not care about your tooling.
If you’re building SaaS as a PM, designer, engineer or even founder, and you want practical signal over AI noise, this one will be worth your time.”
- When: Wed 18th
- Doors: 17:30
- First speaker: 18:15
- Wrap: ~19:30
- Where: Ippon Technologies office, 8/607 Bourke St. Cnr Bourke and King St’s.
And yes, this edition has basically been written so I can promote this Meetup 🥴
Before you go, can you help me out?
Now over 1300 people have subscribed to this newsletter, which makes me feel that it’s resonating. So thank you to all of you 🙏🏻
If you found this issue useful, here are four ways you could support me.
- Share it – Forward this to a founder, leader. hiring manager, or job seeker who you think might find it useful and/or interesting.
- Subscribe to this newsletter
- Follow and/or connect with me on LinkedIn if you’re not already – https://www.linkedin.com/in/simonfromcrew/
- Send feedback – What topics would help you most? What’s working? What isn’t? Just hit reply.
These small actions make a massive difference in helping me reach more people who need this content.
Thanks for being here. Have a brilliant week ahead.
Weekly newsletter drop: Wednesday, 1:30 pm
Need hiring help?
If you’re struggling with a specific hiring challenge, I still offer 1:1 strategy calls through my consultancy, Crew. Book a 15-minute clarity call here.
Simon
