johnfn

35 ms·

661.
by johnfn 1y ago
> Regarding the memory of reasoning from LLMs, I think the issue is that even if you can solve it in the future, you already have code for which you've lost the artifacts associated with the original generation. Overall I find there
662.
by johnfn 1y ago
The article throws out a lot of potential issues with AI generated code, but doesn't stop for a moment to consider if solutions currently exist or might exist to these problems. - Before LLMs, provided that your team did some of their
663.
by johnfn 1y ago
The article specifically addresses this with the fascinating quote from Dave Berry where he reports that the definition of the word unknowingly changed in his head over the last few decades: > “I always thought jerk meant asshole. At lea
664.
by johnfn 1y ago
Your rebuttal to my oversimplification is exactly my rebuttal to your oversimplification.
665.
by johnfn 1y ago
Maybe negative isn’t exactly the right word here. But I also didn’t enjoy the cited post. One reason is that the article really says nothing at all. You could take the article and replace “LLMs”, mad-lib style, with almost any other hyped p
666.
by johnfn 1y ago
You can solve this yourself with a little elbow grease with Docker + a devcontainer. I did this and I’m very happy with the results - Claude can do anything it wants, but it can’t push to prod.
667.
by johnfn 1y ago
I empathize with the author of this post, but is this really the type of story we want on HN? Does this inspire intellectual curiosity? I just feel a directionless anger after reading this.
668.
by johnfn 1y ago
We know how neurons work on the brain. They just send out impulses once they hit their action potential. That's it. They are no more "conscious" than... er...
669.
by johnfn 1y ago
It's obviously not a cost-saving measure? The article clearly cites that you can just start another conversation.
670.
by johnfn 1y ago
Yes, that's a perfect example of what someone who didn't believe that work can be gratifying would say.
671.
by johnfn 1y ago
Is there evidence of Apple doing this in the past?
672.
by johnfn 1y ago
No one is arguing that people don't understand the concept of being excited. But I think a lot of people would be less convinced that it's possible to be passionate about your work.
673.
by johnfn 1y ago
"we"
674.
by johnfn 1y ago
Ah yes, sorry I elided the rest of the list. I think you could roll all these up into "doomerism" though.
675.
by johnfn 1y ago
Believe it or not, every character was typed with my fingers. I'll take this as a compliment :P
676.
by johnfn 1y ago
I couldn't disagree more. Every anti-AI argument I read has the same tired elements - that AI produces slop (is it?) that is soulless (really?). That the human element is lost (are you sure?). As most arguments of the form "hey ev
677.
by johnfn 1y ago
This article is definitely just a joke. Right? Haha? It's a joke? People definitely aren't really doing this in real life? Ha? ...?
678.
by johnfn 1y ago
I tend to be sus of usePrevious. Not saying it doesn't have a place, but often times it's cleaner to just write the diffing code you want directly in the handler. For instance, say you have a text field, and if the value changes y
679.
by johnfn 1y ago
I said clearly and explicitly in my comment that I am happy to read thoughtful anti-AI material. That you disregarded this to write your comment says more about your ability to "form better opinions" than mine.
680.
by johnfn 1y ago
Then don't. You'll likely outperform those who do. Not making any value judgements; both have their place. Sometimes I want to be locked in and sometimes I want to let Claude spin while I take a walk.
681.
by johnfn 1y ago
You'd take the other side of a reverse Gary Marcus? So you'd take Gary Marcus' side?
682.
by johnfn 1y ago
For some reason AI seems to bring out articles that seem to fundamentally lack curiosity - opting instead for gleeful mockery and scorn. I like AI, but I'll happily read thoughtful articles from people who disagree. But not this. This
683.
by johnfn 1y ago
Can you enumerate some of these "immeasurable number of failure states"? For me it starts and stops at "can't count letters in a word". That hardly seems catastrophic. All I have to do is turn on thinking mode and t
684.
by johnfn 1y ago
Oh come now. You are being no fun.
685.
by johnfn 1y ago
I find it hard to empathize with these sorts of articles. I think in, in the spate of GPT-5 related content recently, I've been seeing a lot of articles that boil down to "I tried GPT-5 on a single hard question, and it gave an an
686.
by johnfn 1y ago
No no, getting your entire workflow local requires solving P=NP.
687.
by johnfn 1y ago
The problem is that an online comment is not a dialogue. If I'm conversing with someone in real life, and they say something I strongly disagree with, I can disagree, and we can discuss and perhaps they tell me that I misunderstood the
688.
by johnfn 1y ago
This is a good one. What's gonna happen if I respond to comments on HN for 1000 days in a row? Uh oh.
689.
by johnfn 1y ago
Yea, now I feel like my comment might be misleading. The title mentions crawling; the article itself is talking about something else.
690.
by johnfn 1y ago
This is too severe. There are reasons to need // @ts-ignore. Probably the biggest one in my experience is the inability to type functions with conditional return statements[1]. Probably a good rule of thumb is that you get to use

More ›