Re:Cloudlflare builds OAuth with Claude and publishes all th...
by DonHopkins on Tuesday June 03, 2025 @04:33AM
(#44167758)
Did you save any time though by not reading the article or you just tried to prove a point that if you already think you know what the article says by the title you can criticize it without reading it?
Do you comment before reading the article because it takes less time than reading the article and commenting on what it actually says? Not just the article: You didn't even read the post you're vibe-replying to. It's all there.
As he wrote, and I repeat: the pre-conceived point he was TRYING to prove was that "LLMs were glorified Markov chain generators that didn't actually understand code and couldn't produce anything novel". And much to his shocked amazement, he was proven wrong: "And then, uh... the code actually looked pretty good."
Of COURSE people who are not willing to read have terrible experiences with LLMs. The entire point is that this guy actually DID read, and because of that, he got good results. You should try reading yourself some time! You might be as shocked as he was. (Not that I expect you to read this far...)
Do you comment before reading the article because it takes less time than reading the article and commenting on what it actually says? Not just the article: You didn't even read the post you're vibe-replying to. It's all there.
As he wrote, and I repeat: the pre-conceived point he was TRYING to prove was that "LLMs were glorified Markov chain generators that didn't actually understand code and couldn't produce anything novel". And much to his shocked amazement, he was proven wrong: "And then, uh... the code actually looked pretty good."
Of COURSE people who are not willing to read have terrible experiences with LLMs. The entire point is that this guy actually DID read, and because of that, he got good results. You should try reading yourself some time! You might be as shocked as he was. (Not that I expect you to read this far...)