Also in my experience LLM can often propose solutions which are working but way too complex.
Story time: just yesterday, in VueJS I was trying to iterate over a list of items and render .text of reach item as HTML, but I needed to process it first. Note that in VueJS this is done by adding eg. <span v-html="item.text"></span> where content of the attribute is the JavaScript expression needed to get the text.
First I asked ChatGPT to write the function for processing the text. That worked pretty well and even used part of the JavaScript API which I was not aware about.
Next, I had a "dumb moment" when I did not realize that as I'm iterating through items I can just say <span v-html="processHtml(item.text)"></span>, that's all I really needed. Somehow I thought (or should I say, "hallucinated", ba dum tsss) for a moment that v-html is special or something (it is used differently than the most abundant type of syntax). So I went ahead and asked ChatGPT how to render processed texts while iterating.
It came with a rather contrived solution which involved creating another computed property containing a list of processed texts. I started to integrate it into the existing loop: I would have to add index and use that index to pull the code from the computed property, which already felt a little bit weird.
That's when it struck me: no, no, no, I can just f*ing use the function.
TL; DR: The point is, while ChatGPT was helpful I still needed to babysit it. And if I didn't snap from my lazy moment, or if I simply didn't know better, I would end up with code which is more complex, more surprising, which means harder to reason about for both humans and LLM's. (For humans because now it forces you to speculate about coder's intent, and for LLM's because it's less likely to be reminiscent of surrounding code in its learning data.)
You don't know that houses can't move. Absence of a proof does not imply impossibility.
Sounds ridiculous (esp. for windows / houses) but I think it actually shows where Occam's Razor comes to the rescue: When deciding what to believe, you should consider how many assumptions either model of the world would have to include in order to explain your observations.
Turns you don't need to look for indisputable mathematically rigorous proofs, you just need to find the best model.
Well, I was thinking of a quote that was much more similar to what I wrote (and it's not in the video you linked).
I had such a trouble finding it that I'm starting to feel like it might be one of those "quotes" where the credited author never really said that, but I haven't completely given up :D
While Linux pretty much dominates almost every walk of our lives, even on the consumer devices like smartphones and smart TVs, it has not had the same success on the desktop. What does Torvalds think about it? Is Linux a failure on the desktop? Not really. “The desktop hasn’t really taken over the world like Linux has in many other areas, but just looking at my own use, my desktop looks so much better than I ever could have imagined. Despite the fact that I’m known for sometimes not being very polite to some of the desktop UI people, because I want to get my work done. Pretty is not my primary thing. I actually am very happy with the Linux desktop, and I started the project for my own needs, and my needs are very much fulfilled. That’s why, to me, it’s not a failure. I would obviously love for Linux to take over that world too, but it turns out it’s a really hard area to enter. I’m still working on it. It’s been 25 years. I can do this for another 25. I’ll wear them down.”
Funny how he made it basically for his desktop computer.
33 years later, and Linux is dominating in every part of the OS world except ... the desktop.
(I'm paraphrasing his quote -- he said something like this years ago, can't find it, though.)
(Edit: to be more fair with quotes, it might be the case that I "hallucinated" the quote. he might not have said that, or he might have just said part of it and other part would be someone else's comment. This cio.com article is probably a better source on his position )
Yeah. And I like how even from the message it shows that it's been already well recognized by then.
If I recall correctly from some RMS' talks I've seen many years ago, they've been working on it for years before, it's just the kernel that was missing. As I see it, GNU and Linux was the breakthrough for FLOSS, since at that time you would still have to use a proprietary kernel. (Well, there's GNU Hurd, but I'm not sure if it existed at that time, and even if it did, it was not ready.)
That's a noble goal but does adding more people help the (long-term only, please) effectiveness? At what point does it start hindering it?
I would assume that someone like a pharmacist has to be focused all the time, stakes is high..
Do we have precise data about how physiological state of a pharmacist is changing through the shift? Do we know whether or not the pauses between people -- which we might or might not have considered a wasted time -- are actually essential for their ability to stay focused and reliable? (Is the answer the same for all of them?) Or maybe they could actually still use part of that time in a productive way, right? Also, why is there lack of people in the first place?
Focusing solely on adding more people to the equation seems to neglect factors like this. This tells me that whoever this factoid is trying to impress is not someone who I would want to trust with managing a pharmacy (or anything except maybe some production line) in the first place.
Is "pharmacists seeing more patients" really a measure of something good? I'm a non-native English speaker so cut me some slack but all I can imagine is just longer queues in the pharmacy and more tired pharmacists (and people who now need to wait in the queue now).
I mean, you got a point, but as citizen of Czech Republic, they don't seem to do enough, and U.S. politics really feels like "right" and "more right".
You guys need to fix your voting system. (I know it's easier said than done, my country is about 5th of the age of U.S. Advantage of starting later is a real thing.)
Fun fact: With those 4 games it looks like a tie, but weighed by Steam scores, Linux wins. (Warhammer has like 3/5).
(Disclaimer: I have never played any of those 4 games and don't plan to in a forseeable future. I also realize full well how ridiculously insignificant a sample of 4 is.)
netvor @ netvor @lemmy.world Posts 12Comments 276Joined 2 yr. ago
Also in my experience LLM can often propose solutions which are working but way too complex.
Story time: just yesterday, in VueJS I was trying to iterate over a list of items and render
.textof reach item as HTML, but I needed to process it first. Note that in VueJS this is done by adding eg.<span v-html="item.text"></span>where content of the attribute is the JavaScript expression needed to get the text.First I asked ChatGPT to write the function for processing the text. That worked pretty well and even used part of the JavaScript API which I was not aware about.
Next, I had a "dumb moment" when I did not realize that as I'm iterating through items I can just say
<span v-html="processHtml(item.text)"></span>, that's all I really needed. Somehow I thought (or should I say, "hallucinated", ba dum tsss) for a moment that v-html is special or something (it is used differently than the most abundant type of syntax). So I went ahead and asked ChatGPT how to render processed texts while iterating.It came with a rather contrived solution which involved creating another computed property containing a list of processed texts. I started to integrate it into the existing loop: I would have to add index and use that index to pull the code from the computed property, which already felt a little bit weird.
That's when it struck me: no, no, no, I can just f*ing use the function.
TL; DR: The point is, while ChatGPT was helpful I still needed to babysit it. And if I didn't snap from my lazy moment, or if I simply didn't know better, I would end up with code which is more complex, more surprising, which means harder to reason about for both humans and LLM's. (For humans because now it forces you to speculate about coder's intent, and for LLM's because it's less likely to be reminiscent of surrounding code in its learning data.)