Something I’m finding myself doing regularly: rather than reaching for an existing tool, I ask an agent to build me a custom one.
When times tables became part of the agenda for my son, I asked Sol 6.1 inside Anvil to make us a times tables app.
It didn’t take many prompts to get started. We tried it, added features and improved it as we went. Using it has now become part of his regular practice.
It reminded me of this app from @stefano.menci, back in the old days, when you had to write the code yourself
I’m curious what other custom apps people are building. I’m increasingly finding that, for these small everyday needs, making something that fits is easier than hunting for an existing tool.
My problem is I do this too much, because it’s so easy. I now have over a dozen ways to process the same format stats file in very subtly different ways with very similar results. And I’ve no idea which one is which
Seriously though, I think (monolithic?) admin panels are, if not dead, then certainly reduced in value as a business. It’s now so easy to create personal apps that do exactly what you want that it gets hard to justify expensive subscriptions for something you can probably reproduce for a token spend that’s less than 6 months of subs.
You probably couldn’t sell them, and I think long term we might see a pull towards well written apps by actual developers, but a lot of damage is going to be done to those businesses in the mean time, mine included.
I think the recent Adobe cloning incident shows that “good enough” or “it will do for this project” type software - especially tools rather than end products in themselves - isn’t going away in the near future.
For general little apps, I think I may start asking my LLMs to find a suitable app, instead of creating one.
It’s a matter of personal time.
Entire afternoons can easily go by while I tinker with “my” app via LLM. But with an existing app, someone else has already spent that time, so maybe I don’t have to.
