maddmann

6 ms·

31.
by maddmann 10mo ago
Great point, interesting how agents somehow pick up the same bias.
32.
by maddmann 10mo ago
Really good idea!
33.
by maddmann 10mo ago
lol 5000 tests. Agentic code tools have a significant bias to add versus remove/condense. This leads to a lot of bloat and orphaned code. Definitely something that still needs to be solved for by agentic tools.
34.
by maddmann 10mo ago
I think there is going to be 2-3 year lag in understanding how llms actually impact developer productivity. There are way too many balls in the air, and anyone claiming specific numbers on productivity increase is likely very very wrong. Fo
35.
by maddmann 10mo ago
I strongly agree. It may be even more than 90%. For example yesterday I was able to use lovable (and Claude code web) on my phone to build out an almost 1:1 replacement (for my purposes) for an expensive subscription based app for working o
36.
Show HN: I replaced my premium workout app with vibecode(strengthquest.lovable.app)
9 points by maddmann 10mo ago|1 comments
37.
by maddmann 10mo ago
I’m not going to generalize, but I already told you, I met people who had a profound dislike of elon/trump and I don’t think I saw a single Tesla during that trip. Maybe Italy is a different case since they have a passion for high qual
38.
by maddmann 10mo ago
I was in Italy over the summer and they seemed to universally detest musk, and trump there. So much, in fact, that it is embarrassing to be a US citizen in Europe with these fools “representing” the US.
39.
by maddmann 11mo ago
How did capitalism create overpopulation? Isn’t that more related to agriculture and better medical tech?
40.
by maddmann 11mo ago
Love it! Interesting to read these, and you must have a good prompt underneath
41.
by maddmann 11mo ago
“ AI is going to utterly cripple people intellectually and motivationally. If you can't even do the above ten second process, you may want to make more of an effort before you find yourself utterly fucked and starving.” This does not s
42.
by maddmann 11mo ago
[flagged]
43.
by maddmann 11mo ago
Cool, nice work. I’ll check it out at some point
44.
by maddmann 11mo ago
I’m curious how this would work for production level apps. My concern is that it would go off and do a lot of extra stuff and grow your codebase with orphaned code or non mvp features.
45.
by maddmann 11mo ago
You lint the file paths inside Claude.md?
46.
by maddmann 11mo ago
Yeah it is a mystery to me how folks could also maintain context in more than two sessions. The code review would be brutal. You’ll also end up dealing with merge conflicts if you haven’t carefully split the work or modularized the code.
47.
by maddmann 11mo ago
I really enjoyed reading this. One thought I had on the issue of paths in Claude.md My concern with hardcoding paths inside a doc, it will likely become outdated as the codebase evolves. One solution would be to script it and have it run pr
48.
by maddmann 11mo ago
“The purpose of DEI is to enforce what is idealized as the equal outcome, assuming that all observed differences are the result of discrimination“ I don’t think a single proponent of DEI has ever said this, and it is telling to me that you
49.
by maddmann 1y ago
Did you read the article? I think this case study shows why gm is likely to be key to avoiding mass starvation as climate change becomes a bigger issue.
50.
by maddmann 1y ago
I’m curious what you are expecting when you say progress has stagnated?
51.
by maddmann 1y ago
It seems like a very backwards approach to testing a new technology: instead of “let me figure out how to use this tool”, the approach I see a lot is “let me continually use this tool in an known incorrect way” Of course there is valuable k
52.
by maddmann 1y ago
Good point: perhaps op should have had the llm output a script to compare the two lists.
53.
by maddmann 1y ago
“Defaults are shit” — is that really true though?! Just because it shits the bed on some tasks does not mean it is shit. For people integrating llms into any workflow that requires a modicum of precision or determinism, one must always eval
54.
by maddmann 1y ago
Thanks for giving it a try! I am also a novice with Strudel so had a similar motivation when building this, and also just a curiosity if llms could output interesting compositions. One thing you might find interesting is the search tool for
55.
Experimental platform using LLMs to generate algorithmic music(vibecompose.vercel.app)
2 points by maddmann 1y ago|3 comments
56.
by maddmann 1y ago
I have been experimenting this past week with a small web platform that mixes JavaScript, music pattern generation, and large language models to create algorithmic compositions. The idea is to have a language model suggest Strudel style mus
57.
by maddmann 1y ago
Yeah definitely feels vibe coded. I am so sick of the overuse of icons.
58.
by maddmann 1y ago
Haha it sounds you applied the opposite of the YAGNI principle building this.
59.
by maddmann 1y ago
There is something creepily reminiscent to multi level marketing with posts that discuss how to build with ai and then push very narrow subscription based saas “products”. I.e you can get rich too (sadly though rich in this case is not eno
60.
by maddmann 1y ago
I wonder how many big tech lobbyists got there hands on this and shaped it?

More ›