I'm reminded of the scifi paradox of interstellar travel: when it becomes possible to send a ship to Alpha Centauri, it also becomes pointless to do so, because in the time it would take for the ship we sent today to reach its destination, we would've come up with a faster ship which would then overtake the previous one.
It’s a provocative question but the analysis seems to conflate a product with a business. A startup is an organization that serves a demand in a market.
> Imagine what you could build if you just wait a year.
The models will be more capable but you’ll have given up a year of experience solving the problem and competing in the market. You might even have less generative ability than if you’d managed to get some traction and draw in other smart people to the organization (who all multiply what can be achieved with the models.)
Of course it is possible to start too early: If you reach the end of runway before takeoff.
And no matter when you start, if the business isn’t defensible there’s little or negative ROI for the effort.
Yeah, I think that's mostly true. It's certainly true that software alone isn't what make most big businesses big (eg, all the twitter clones that never became twitter, and that's probably even more true with b2b companies).
That said, I'm not sure the experience of understanding the market is all that helpful, or at least, is a very deep defense. it's basically this other comment to me (https://benn.substack.com/p/will-there-ever-be-a-worse-time-to/comment/185787147): If you build a successful thing, it still takes someone very talented to beat you. But that's increasingly all they need, whereas before they had to be talented *and* had to put in a ton of work.
what happens when the creation cost goes to 0 - only the distribution part becomes the one having value. It has happened with content, which you referred to in a prev. article, right?
Sorta? On one hand, I don't think it being fast to create stuff necessarily means it's easy to create good stuff, or that anyone can do it. On the other hand, you can mostly just copy people. Like:
- Someone makes a good thing
- Someone else has a good idea about how to do the good thing, but a little better.
Before, the second person had to build all the stuff the first person built before they extend it with their improvement. But now, you can do it pretty quickly. So it's not exactly that anyone can create anything, but I think it's something like "anyone can start from almost the frontier." Which, yeah, makes the moats that got dug by people just doing a bunch of work suddenly seem a lot shallower.
Ok, written like this, the first part made me think of the clothing industry with fast fashion. Not sure to which extent the analogy stands. That’d be an interesting one for a future article 👀👻
I was actually talking to someone recently and they made that exact point. They had worked for a clothing brand, and they said there was a point not that long ago when "anyone" could make a clothing brand. It used to be that you had to have access to manufacturers and distributors and places to put stores and all of that. And now, all of that is pretty easy to set up or outsource (eg, you can find manufacturers; you can buy blanks easily; you can have a shopify store, etc). I couldn't do all that very easily, but if you know your way around the industry, it's not that hard. And so all you really need to do to be successful is have a product people think is cool, and way to make people aware of it.
I'm reminded of the scifi paradox of interstellar travel: when it becomes possible to send a ship to Alpha Centauri, it also becomes pointless to do so, because in the time it would take for the ship we sent today to reach its destination, we would've come up with a faster ship which would then overtake the previous one.
I've never heard that and it's amazing.
Can we make benn.chat a thing
The core conceit of it needs some adjusting, probably. (In building it, I learned it's probably more suited for (benn.)email than (benn.)chat)
It’s a provocative question but the analysis seems to conflate a product with a business. A startup is an organization that serves a demand in a market.
> Imagine what you could build if you just wait a year.
The models will be more capable but you’ll have given up a year of experience solving the problem and competing in the market. You might even have less generative ability than if you’d managed to get some traction and draw in other smart people to the organization (who all multiply what can be achieved with the models.)
Of course it is possible to start too early: If you reach the end of runway before takeoff.
And no matter when you start, if the business isn’t defensible there’s little or negative ROI for the effort.
Yeah, I think that's mostly true. It's certainly true that software alone isn't what make most big businesses big (eg, all the twitter clones that never became twitter, and that's probably even more true with b2b companies).
That said, I'm not sure the experience of understanding the market is all that helpful, or at least, is a very deep defense. it's basically this other comment to me (https://benn.substack.com/p/will-there-ever-be-a-worse-time-to/comment/185787147): If you build a successful thing, it still takes someone very talented to beat you. But that's increasingly all they need, whereas before they had to be talented *and* had to put in a ton of work.
Tired: npm install
Wired: claude build my app
It's still not exactly there, for a bunch of reasons, but man, it's hard to imagine it won't be pretty close in a few years.
what happens when the creation cost goes to 0 - only the distribution part becomes the one having value. It has happened with content, which you referred to in a prev. article, right?
Sorta? On one hand, I don't think it being fast to create stuff necessarily means it's easy to create good stuff, or that anyone can do it. On the other hand, you can mostly just copy people. Like:
- Someone makes a good thing
- Someone else has a good idea about how to do the good thing, but a little better.
Before, the second person had to build all the stuff the first person built before they extend it with their improvement. But now, you can do it pretty quickly. So it's not exactly that anyone can create anything, but I think it's something like "anyone can start from almost the frontier." Which, yeah, makes the moats that got dug by people just doing a bunch of work suddenly seem a lot shallower.
Ok, written like this, the first part made me think of the clothing industry with fast fashion. Not sure to which extent the analogy stands. That’d be an interesting one for a future article 👀👻
I was actually talking to someone recently and they made that exact point. They had worked for a clothing brand, and they said there was a point not that long ago when "anyone" could make a clothing brand. It used to be that you had to have access to manufacturers and distributors and places to put stores and all of that. And now, all of that is pretty easy to set up or outsource (eg, you can find manufacturers; you can buy blanks easily; you can have a shopify store, etc). I couldn't do all that very easily, but if you know your way around the industry, it's not that hard. And so all you really need to do to be successful is have a product people think is cool, and way to make people aware of it.