Content pfp
Content
@
https://opensea.io/collection/dev-21
0 reply
0 recast
2 reactions

Stephan pfp
Stephan
@stephancill
Feels like the vibe coding narrative has taken over the LLM-assisted coding use case over the past couple of weeks? Super hard to get cursor to make targeted changes to a file without it editing a bunch of other unrelated stuff. Has become less useful in my day to day workflow as a result
6 replies
2 recasts
18 reactions

Rafi pfp
Rafi
@rafi
Same for me. I see all this extraordinary examples on X and people talking about it all the time. I haven’t seen anyone doing proper vibe coding and all examples look alike. It’s a psyop
2 replies
0 recast
1 reaction

Stephan pfp
Stephan
@stephancill
vibe coding is very easy to do but that's not my goal in the case i'm describing in OP. i feel like sota models are being tuned for one-shotting full features and making tons of changes without the dev having to review it - this is frustrating when you just want it to make minimal diffs also this is my experience with vibe coding: https://warpcast.com/stephancill/0xeb8cabf2
2 replies
0 recast
1 reaction

Rafi pfp
Rafi
@rafi
Can't believe I had vibe coder @leopastel literally next door for a week. Would love to see him work. Stream maybe? https://warpcast.com/leopastel/0x4db6ba26
1 reply
0 recast
2 reactions

Rafi pfp
Rafi
@rafi
I see model opportunity Make sense that general purpose models are being tuned for one-shots. Easier to market and stunt people with marvel of stochastic intelligence
0 reply
0 recast
0 reaction