Integration Love Story with Scott Hanselman

Introduktion

Scott Hanselman, thank you for joining us. It’s absolutely my pleasure.

We always like our guests to introduce themselves — but you’re a special guest, so we’d like you to introduce yourself to someone who is not in the tech industry. How would you do that?

I would say I’m a professor. I would say I’m a computer teacher. Teacher implies grade school or high school in the United States, and professor implies university. I have never taught high school, but I have taught at two universities. So professor, to mean getting a degree.

And what’s your main topic when you teach?

I teach computers, computer ethics, AI, programming languages, and the history of computers. I think computer science history is an under-respected space. You can see right here — I’ve got a PDP-11 minicomputer running in the background, 3D printed helmets, an arcade machine. There’s an Apple One around the corner that you can’t see. The last 50 years of computing, basically.

You’ve also done more than a thousand podcast episodes.

I just recorded episode 999 this morning. I’m trying to get Satya for the thousand, but I might have to make it 1,024 because his schedule is quite challenging. I need a famous person for a thousand but I don’t have one on speed dial.

Läraren som förändrade allt

How did the teaching start? Why was it important to you?

When I was about 10 or 11, I was very affected by a teacher who — when my parents met with her in grade five — told them to buy me a computer. They bought me a Commodore 64, and that was extremely meaningful. If that had not happened, I would not be where I am.

This wasn’t just a teacher I remembered fondly. She was at my wedding. I kept in touch with her until she passed away, for 25 or 30 years. And I still have my Commodore 64s right here.

The enthusiasm she gave me was so infectious that it informed my own enthusiasm. A teacher is someone who’s excited about something — so excited that they want to take time out of their day to tell you about that thing.

I’ve never liked the word evangelism. Technology evangelists, developer evangelists — I think that’s always a bad idea. Evangelism implies that I have a religion and your religion is bad, so you should convert to mine. That’s just a bad vibe. A teacher is not trying to convert you to anything. A teacher is trying to awaken what’s going on in your own brain.

So I’m kind of wandering around saying: have you heard about 3D printing? Are you familiar with Raspberry Pi? I made a robot earlier this week that has an arm and stuff. How can you not be excited about these things? I’m trying to be like the person on children’s television who sings and dances and teaches you how to read — except for grown people who like computers.

Scaling yourself — då och nu

About 12 years ago I watched your talk “Scaling Yourself” — I rewatched it yesterday. It’s been a part of my thinking about efficiency versus effectiveness ever since. If you were to give that talk today, what would you rewrite?

I would remind people to stay awake — and by awake I mean present and in the moment. I think we lose days, we lose time. You go to work, maybe you like your job maybe you don’t, you do the thing, you’re tired, you put Teams on your phone, Outlook, Slack — and then you stop working at 5. Then at 7 or 8, after dinner, you tend to check your email again. And the next thing you know it’s 10:30 at night and you’re looking at work. That’s an unhealthy thing.

You have to kind of ask yourself: what am I doing? Staying awake and present and intentional as much as possible. If there’s a big reorg happening at work, sure, focus on that. But if you’re doing it every day, all day, at all times — and then Instagram and Bluesky and everything else, each one another inbox — you forget about what’s in front of you and the people around you that matter.

That sounds like it requires a certain maturity to actually achieve.

There’s a word for what you’re describing — FOMO, fear of missing out. There’s a million things you’re going to learn, but you’re not going to learn them all. Being present and learning something slowly has value. Learning takes intention. I try to set appointments with myself — 90 minutes, read this book. Because if you just try to find time, like “I’ll read it when I have a spare moment,” that’s not planning.

And one thing from the talk that gets quoted the most: airplane mode works on the ground. But the trick is to physically put the phone in another room. The phone is the problem. If you’re alone in a room without your phone, everything’s fine. All it takes is one moment of weakness and you’re looking at Reddit.

AI — verktyg, papegojor och vad vi egentligen optimerar för

Everything right now is AI. You talk about efficiency and effectiveness in your talk. I think about the possibilities to scale with AI — but also to be more effective.

We need to remind ourselves of the distinction. Efficiency is doing things right. Effectiveness is doing the right thing. We are focused too much on AI as an efficiency tool. AI will help you be efficient. AI will not help you be effective. It doesn’t have good judgment.

And to what end? If you make insulin more efficiently — is the goal to make more money, or to make sure everyone who needs insulin has it? We need to remind ourselves what we’re optimizing for. When you die and there’s a gravestone, does it say: “He was 10% more efficient than the rest of us through his excellent use of large language models”? Is that what goes on your tombstone?

What I want is to reduce toil. I want people to do the fun things and not the tedious things. If AI can do the toil, then I can do the fun stuff. But then the question becomes — why am I doing 40 hours of fun stuff at work? The fun stuff is at home too. Let me get in, do my job, and by Thursday at lunchtime, go home. I’ll be equally as productive. We’ll still have growth. Everyone wins. But no one is having that conversation.

The 40-hour week is a thing the Western world made up. It’s easy to count and easy to pay for. It doesn’t necessarily have a deeper purpose than that.

Vad AI faktiskt är — och vad det inte är

Will AI be able to write code the way a great novelist writes a great novel?

If I AI-generated you a novel in Swedish and said “this is amazing, you’re going to love it,” you’d read it as native speakers and go: something doesn’t feel right about this. Did you use AI? You know it’s not art. It’s shovelware.

We keep saying it’s going to get better, it’s going to get better. Will it? Can next-word prediction ever be truly intelligent? What will it be? It’ll be a parrot. You have parrots in Sweden — why do they speak the language of the household they grew up in? Because there’s someone next to them talking all the time, and after 10 years the parrot repeats the words. You don’t think the parrot is going to develop intelligence just because we talk to it long enough.

So AI is not a fact-maker. Large language models are only as good as the data they were trained on — and they were trained on Reddit and Stack Overflow, not all of which is correct. Trust but verify. A human needs to go over whatever it generates.

I very much believe in human-in-the-loop. I believe in AIs that make plans, then I review the plan, and then I push the button. When Microsoft’s software engineer agent runs a task, the pull request is in my name. You’re in a self-driving car, you let go of the wheel, it hits somebody — you are the driver. You have to understand that.

Vad vi ska lära nästa generation

What would your advice be to developers coming into this community now, in 2025, with all this uncertainty?

People should learn about their history and learn about the full stack.

If AI can do one thing well, it can teach. But using it just to generate boilerplate code is a very simplistic use. The more powerful use is to say: I don’t understand this codebase, can we sit down and talk about it? Use it as a rubber duck. Help me walk through this so I might better understand it. That’s genuinely helpful for an early-career developer.

But the challenge is teaching people what questions to ask.

And we also need to remind people that deciding what you’re going to do with a tool is more important than the tool itself. If you give a monkey a hammer, it might learn to build houses for other monkeys — or it might just hit the other monkeys with the hammer. It might think the hammer is a weapon. We need to be thinking about these tools and why we were given them and what they’re actually for.

Integration Love Story

Do you have a love story regarding tech — a moment where you fell in love with it?

The Commodore 64 for me. That was the first computer I ever saw, and I continue to tinker with 40-year-old computers.

Here — I’ll show you something I built. This is from Smartykit.io, an Apple One kit. They give you all the tools and components and you build it from scratch. It comes as a pile of wires and you place the CPU, the memory bus, all of it. My dad made the board it’s mounted on. Learning how a computer works at that low level — that’s where the love is.

Avslutning och näste gäst

You have a suggested next guest for us?

A good person to talk to would be Taylor Poindexter. She works at Spotify, she’s a developer, and she also loves whiskey — she has an account called Woman with Whiskey and does tastings and things like that. She and I gave the closing keynote of the Strange Loop conference together a couple of years ago. I think she would be a great interview.

And a challenging question for her?

I would want to know how she got started. Not her first computer specifically — just how she started.

Thank you and we’ll reach out to her.

Thank you so much. It was a true pleasure.