Post

Matt Pocock
@mattpocock.com

Pretty sure that soon you'll need GenAI for your app as often as a database. And the same concerns apply. Security, performance, observability. A lot of devs are going to need to upskill.

Oliver Turner
@oliverturner.cloud

Interesting notion! What scenarios do you see GenAI being as fundamental a need as a db? Recommendations, definitely, but what besides?

0 reposts1 like

Matt Pocock
@mattpocock.com

Recommendations, relational search, autocomplete, will become table stakes. More users will want chat interfaces. AI features are an easy 'pro' tier upsell to any app.

Misha Kaletsky
@mmkal.com

I agree about most of that, but not chat. Chat is a great stopgap for developers/product teams who haven't figured out the right happy path yet, and want to let their users find it. But once you've found it, UI ("book now" button) beats AI (typing "b-o-o-k-n-o-w") every time.

Misha Kaletsky
@mmkal.com

That's what we found empirically when I was working on AI chatbots at Microsoft (admittedly "early" AI, pre GPT). People liked when chatbots enabled new functionality - and AI was great for shipping new stuff fast. But for any given functionality it was just not efficient UX.

Akos
@akoskm.com

“Recommendations” is a huge field in itself. If you build a RAG for a book store that detects fraudulent books, that’s a “recommendation” engine too, that shows moderators books they should keep an eye on. Similarly with comments, spam, etc.