Ville V. VanninenOct 14
Can you believe I'm doing prompting workshops at work? My first advice: "Please don't, you human brains are so much better at all of this and your thoughts matter." But it's already happening anyway so I'm trying my best to defuse any bombs before they go off.
Show threadManav RathiOct 14

@sakamies I feel full embracement is better, instead of pretending that human brains are "better". Human brains are different, yes, but better is not an appropriate axis.

The people I find best at getting LLMs to do their bidding have a intuitive "Theory of Mind", and talk to the LLM as a peer not as a slave.

(just an opinion! not implying any certainity...)

Show threadVille V. VanninenOct 14
@mnvr I mean yeah, there's some hyperbole and a lot of value judgement in post. I've done and seen useful stuff being done with LLMs. Todays particular frustration stems from how much I see AI output being used as is, without thinking much for one self if the output even solved the problem at hand. It's generating a lot of stuff for me to fix.
Show threadManav RathiOct 15

@sakamies I've fully embraced them and I have no regrets (at least so far).

I wrote this a couple of months ago, and my opinion hasn't changed since then - https://mnvr.in/beginners-mind (specifics have changed, e.g. now I find Codex better)

Beginner's mind | mnvr.in

It has been only one day since I first used Claude Code, and I'm convinced that software engineering as I knew it is dead.

Show threadManav Rathi
@sakamies I realized my post might be too philosophical, so e.g. here is a more practical take on it - https://www.youtube.com/watch?v=fHWFF_pnqDk
Vibe coding in prod | Code w/ Claude

YouTube
Oct 15 at 2:17pmWeb