Execution still matters. Using AI in the background or not, how the idea is implemented and delivered still matters just as much as it always did. People use the term “AI slop” because someone thought execution didn’t matter and AI could do it all for them… and it was terrible. This goes back to what you said about trying to delight the user, this is execution… AI or not.
As for the rest, I think this line is key:
> But I was thinking more like a private user who already trusts what I use by default (because I built it)
Of course you trust what you wrote, but do you trust what everyone else writes? Put yourself in the shoes of your potential customers. Most people don’t know you, your values, your intent… all they have to go by is what you tell them. Also remember, people lie. So don’t just tell them, prove it.
And to your point with AI, the ability for someone to make something that was seemingly done with care and to delight, just so they can steal data, has never been easier. Your take away was execution is cheap, but maybe the take away should be data harvesting is now cheap, and that data is more valuable than ever, so people are right to be wary of anything that is accessing their data.
Thanks. I’m not saying execution doesn’t matter. I’m saying it’s cheap. It’s fast. Which means it’s no longer the moat. Ideas are.
On the data side, your data is not valuable to me. Your money is. I’ve never made a product that sells user data or uses it for ads.
I enjoy making good technology that people want to use and pay for. Maybe one day we will make some data- or ad-supported business. I don’t know. It never seemed like the right idea, and I wouldn’t know where to start.
But I believe for that kind of thing, to make the kind of money I make, you have to have a gigantic scale, and the things I make have not really had a gigantic scale. They’re more specialist, technological products rather than mass-market, viral whatever.
It’s the same dynamics and economics I think that led me to learn that open source is not really a fit for the products I enjoy making; shareware-style trial and license is more of a fit.
So when I release a paid product in that style and people come at me like, “oh my God, what are you doing with my data?” I kind of think, what the actual fuck. What am I doing with your data? It’s nothing. I don’t care about it. I’m merely the Postal Service for it or whatever.
I feel like, don’t people know that I’m not making data products? And I think importantly, if I were, then I would establish trustworthy expectations with clarity about that. I would do it cleanly: yes, this product is free because we’re mining your data to sell it. If I was going to do that, I would just say it, you know. Because I think the people you lose through that honesty is not bad. There’s plenty of apparently data products that are free but making money through selling user data, but really I find it odd how much people think about that stuff when looking at the kind of products I make that are always paid, no-data products — especially when I don’t think about user data at all except “how can I make it secure so people can’t hack it.”
So I guess my learning here has been more like: there’s a lot of customers out there with, like, technology PTSD. Which means, even though I’ve kind of missed it in the news, there must’ve been a lot of shady data practices and rip-off scamming kind of things going on, and people are scared about that. But the weird thing to me is, like, do people really believe that kind of stuff is so lucrative that you could make more money doing that than selling paid products? I mean, what the actual fuck?
So there’s a part of this that still doesn’t make sense to me logically, but I accept that people have this kind of PTSD about data brokers and shady practices. And I think it’s important to be really clear about how you’re protecting them from that, essentially. And I didn’t give any thought to that before, because of course I would have no interest in doing that bad stuff, so why would I think I would need to say it? But it turns out the environment we live in, you kind of have to assure people
I'm don't blame you for the mistake, but if AI can't write a correct HN comment maybe execution (ie code review, testing, reading what you're posting) matters more than you think.
No not more than I think because speed of execution is paramount. These things like checklists, they don't really matter. They are important, but they're not the main thing.
Execution still matters. Using AI in the background or not, how the idea is implemented and delivered still matters just as much as it always did. People use the term “AI slop” because someone thought execution didn’t matter and AI could do it all for them… and it was terrible. This goes back to what you said about trying to delight the user, this is execution… AI or not.
As for the rest, I think this line is key:
> But I was thinking more like a private user who already trusts what I use by default (because I built it)
Of course you trust what you wrote, but do you trust what everyone else writes? Put yourself in the shoes of your potential customers. Most people don’t know you, your values, your intent… all they have to go by is what you tell them. Also remember, people lie. So don’t just tell them, prove it.
And to your point with AI, the ability for someone to make something that was seemingly done with care and to delight, just so they can steal data, has never been easier. Your take away was execution is cheap, but maybe the take away should be data harvesting is now cheap, and that data is more valuable than ever, so people are right to be wary of anything that is accessing their data.
Thanks. I’m not saying execution doesn’t matter. I’m saying it’s cheap. It’s fast. Which means it’s no longer the moat. Ideas are.
On the data side, your data is not valuable to me. Your money is. I’ve never made a product that sells user data or uses it for ads.
I enjoy making good technology that people want to use and pay for. Maybe one day we will make some data- or ad-supported business. I don’t know. It never seemed like the right idea, and I wouldn’t know where to start.
But I believe for that kind of thing, to make the kind of money I make, you have to have a gigantic scale, and the things I make have not really had a gigantic scale. They’re more specialist, technological products rather than mass-market, viral whatever.
It’s the same dynamics and economics I think that led me to learn that open source is not really a fit for the products I enjoy making; shareware-style trial and license is more of a fit.
So when I release a paid product in that style and people come at me like, “oh my God, what are you doing with my data?” I kind of think, what the actual fuck. What am I doing with your data? It’s nothing. I don’t care about it. I’m merely the Postal Service for it or whatever.
I feel like, don’t people know that I’m not making data products? And I think importantly, if I were, then I would establish trustworthy expectations with clarity about that. I would do it cleanly: yes, this product is free because we’re mining your data to sell it. If I was going to do that, I would just say it, you know. Because I think the people you lose through that honesty is not bad. There’s plenty of apparently data products that are free but making money through selling user data, but really I find it odd how much people think about that stuff when looking at the kind of products I make that are always paid, no-data products — especially when I don’t think about user data at all except “how can I make it secure so people can’t hack it.”
So I guess my learning here has been more like: there’s a lot of customers out there with, like, technology PTSD. Which means, even though I’ve kind of missed it in the news, there must’ve been a lot of shady data practices and rip-off scamming kind of things going on, and people are scared about that. But the weird thing to me is, like, do people really believe that kind of stuff is so lucrative that you could make more money doing that than selling paid products? I mean, what the actual fuck?
So there’s a part of this that still doesn’t make sense to me logically, but I accept that people have this kind of PTSD about data brokers and shady practices. And I think it’s important to be really clear about how you’re protecting them from that, essentially. And I didn’t give any thought to that before, because of course I would have no interest in doing that bad stuff, so why would I think I would need to say it? But it turns out the environment we live in, you kind of have to assure people
2026! Blasted AI copy edit to reduce spiciness also took it back in time
I'm don't blame you for the mistake, but if AI can't write a correct HN comment maybe execution (ie code review, testing, reading what you're posting) matters more than you think.
No not more than I think because speed of execution is paramount. These things like checklists, they don't really matter. They are important, but they're not the main thing.
"They aren't being "ninnies"; they are being wise." damn the AI writing is so blatant