We sat down with Steve Klabnik for a discussion about something that might sound meta but turns out to be incredibly practical: how to actually have useful conversations about AI when your goal is learning rather than winning arguments.
Steve wrote a blog post about his disappointment with the AI discourse that resonated with us, so we wanted to talk to him about it.
Many of us exploring this rapidly evolving space have run into the same frustrating pattern. You want to understand how AI tools work, what they're good at, or how other people are using them. But conversations quickly derail into heated debates about ethics, energy usage, or whether AGI is around the corner.
Steve walks through the strategies he developed after finding himself stuck in this exact situation: wanting to learn about AI but constantly hitting conversational dead ends.
You can watch the full conversation on YouTube or read selected quotes below.
"I think that the core thing...is understanding what the other person that you're talking to is actually interested in and open to...sometimes people aren't even interested in switching their opinion from one side of that to the other and sort of like recognizing that you're not gonna budge someone on that...and then not wasting your own time trying to change their mind." — Steve Klabnik
"Recognizing like, okay, this is not the person I need to be talking to about this thing right now, is kind of like step zero of having these conversations." — Steve Klabnik
"I've had the worst of these conversations on what I would say is like mass social media platforms. Because anyone can read anybody's conversation and jump into the middle...any random person with any random agenda can show up and decide to fight you about whatever it is that you're talking about." — Steve Klabnik
"I've had significantly more productive conversations with people whenever it was in smaller venues and down to even like in person as opposed to the internet. Like some of the most productive conversations I've had on this topic have been like at dinner with a couple of friends who have a variety of opinions." — Steve Klabnik
"I have a long history of playing games. Like I played Magic the Gathering as a kid and I've been a video gamer for a while...A lot of things that you need to do in something like poker is embrace the fact that it's random, but come up with strategies in order to mitigate that randomness so you can get the outcome that you want." — Steve Klabnik
"Like take TDD for example, right? Like you could be like, I'm using TDD in a Rails app, you know, with Cucumber and I liked it and then somebody else is like, it worked horribly at my job and it made it not work really well...We've had these kinds of situations before, but that never turned into the same degree of vitriol or explosion in all cases." — Steve Klabnik
"There is so much change in this space and it's happening so quickly, what skills are relevant and what skills you need to get good at are totally different...Everyone was talking about prompt engineering a year ago, but now it's about context engineering and prompt engineering is not seen as being very important." — Steve Klabnik
"Skills are not easy to develop or free to develop in a time or money sense...when you're trying to acquire those skills, it becomes difficult cause you're like, I wanna learn if I do wanna learn about these things, and why does this matter and why does this not matter? And how is this changing?" — Steve Klabnik
"I think this kind of goes back to how I've built all the skills I've ever built, which is being willing to be bad at something and not getting on tilt about it...to get better at anything you have to try stuff and then see what works and see what doesn't, and then iterate from there." — Steve Klabnik
"Because this world is evolving so quickly, you kind of have to be in the mindset of like, I am trying stuff out. I see what works and I see what doesn't. And I continue doing the things that work for me and I stopped doing the things that don't. And be willing to reevaluate and be open to new things." — Steve Klabnik
Learn more about Agentic Engineering
Working toward genuine, quality software in an era where code production is not the constraint anymore. Learn our take about Agentic Engineering and catch up with all the sessions where we're joined by experts to further develop this topic.
Looking for a better editor?
You can try Zed today on macOS, Windows, or Linux. Download now!
We are hiring!
If you're passionate about the topics we cover on our blog, please consider joining our team to help us ship the future of software development.
Instagram Engineering
Shopify Engineering