Thomas Dohmke

So we got to build a stack, an engineering stack, a platform stack that resembles our business.

Another example is GitHub.

We have a product called GitHub Actions, which is our continuous integration deployment service.

And we are looking into how can we get that from 99.9% uptime to 99.99, so-called four nines uptime.

That's not something you can just go and ask the agent, right?

Like, well, this is an engineering problem that you need to solve.

That is very similar, I think, from a complexity perspective, if you're the mayor of a city and you promise everybody you're making public transport better, right?

That sounds simple when you're promising it during the election.

And it's much harder when you're actually elected and now you have to figure out, how do I change the city?

How do I get the permits?

How do I build?

Where's the funding coming from?

And I think that is where you're always going to have humans

humans that use agents and combine different agents, but you got to know at which point in your day-to-day process, in your engineering process, you're using these agents for what purpose?

So we're going to see two trends, right?

If today you're buying software, SaaS services, platforms like Substack, because it's cheaper for you to spend a small amount of dollars every month on those platforms instead of all building it yourself.

If building it yourself with an agent becomes so simple that you can just do that in a couple of minutes, you will stop paying for these SaaS services.

You know, we will see a commoditization of software that is so simple that recreating that on the fly will replace those SaaS services.

And one such example is Manos, one of the tools that became very popular in recent weeks.

And I met with the founders a couple of weeks ago and they showed me how they were looking for an office in Tokyo.