←back to thread

Claude Sonnet will ship in Xcode

(developer.apple.com)
485 points zora_goron | 1 comments | 29 Aug 25 00:44 UTC | HN request time: 0s | source
Show context
throwawa14223 ◴[29 Aug 25 02:08 UTC] No.45059251[source]▶
It's getting harder to find IDEs that properly boycott LLMs.
replies(12): >>45059290 #>>45059291 #>>45059362 #>>45059390 #>>45059429 #>>45059473 #>>45059478 #>>45059506 #>>45059650 #>>45060008 #>>45061090 #>>45061442 #
drusepth ◴[29 Aug 25 02:15 UTC] No.45059290[source]▶
Ironically, you could probably vibe code your own.
replies(3): >>45059424 #>>45059487 #>>45060689 #
rs186 ◴[29 Aug 25 02:41 UTC] No.45059487[source]▶
Good luck getting just scroll bar right with vibe coding. You'll be surprised how much engineering is done to get that part work smoothly.
replies(1): >>45060126 #
CamperBob2 ◴[29 Aug 25 04:17 UTC] No.45060126[source]▶
If enough examples are in-distribution, the model's scroll bar implementation will work just fine. (Eventually, after the human learns what to ask for and how to ask for it.)

Why wouldn't it?

replies(1): >>45060572 #
shakna ◴[29 Aug 25 05:32 UTC] No.45060572[source]▶
Most programs today regularly have bugs with scrolling. Thus, an LLM will produce for you... A buggy piece of code.
replies(1): >>45061180 #
adastra22 ◴[29 Aug 25 07:21 UTC] No.45061180[source]▶
LLMs are not Xerox machines. They can, in fact, produce better code than is in their training set.
replies(1): >>45061430 #
mirkodrummer ◴[29 Aug 25 08:01 UTC] No.45061430[source]▶
That is funny for how much is wrong. Ask the LLMs to vibe code a text editor and you'll get a React app using Supabase. Engineering !== Token prediction
replies(3): >>45062123 #>>45066613 #>>45068661 #
1. CamperBob2 ◴[29 Aug 25 17:01 UTC] No.45066613{6}[source]▶
Ask the LLMs to vibe code a text editor, and you'll get pretty much what you deserve in return for zero effort of your own.

Ask the best available models -- emphasis on models -- for help designing the text editor at a structural rather than functional level first, being specific about what you want and emphasizing component-level test whenever possible, and only then follow up with actual code generation, and you'll get much better results.