My people. The only icon I keep on my desktop is the recycle bin, and I only keep that around to throw away icons for programs that place a desktop icon without asking.
My people. The only icon I keep on my desktop is the recycle bin, and I only keep that around to throw away icons for programs that place a desktop icon without asking.
That was insightful. To the oblivious user desktop icons are something plain that should just work. But there's always so much hidden complexity behind all things computing. I'm only a novice and...
That was insightful. To the oblivious user desktop icons are something plain that should just work. But there's always so much hidden complexity behind all things computing.
I'm only a novice and occasional coder but this absolves me of some contrition:
Reading code is the hardest part
When going through old code, especially some that has barely any comments, it can take a very long time to understand what is actually going on
Writing code seems so much easier than reading it. I can't be the only one spending copious time on thinking how minor variables should be named. I do this because I might have to readjust a script I wrote months or years later. And at that point I want it to read as autological as possible.
Ironically I like to keep my desktop clean. Just like a physical desktop. If I place something on the desktop it's because it needs attention, but I can only resolve it at a later time.
Reading code is hard but AI is helping with a lot of that. I've started using AI to help make my comments on my own code because what I write and what it writes are totally different. The AI is...
Reading code is hard but AI is helping with a lot of that. I've started using AI to help make my comments on my own code because what I write and what it writes are totally different. The AI is writing functional and logical comments, I'm writing things that make sense at the time but later or to a stranger would be complete nonsense out of context to what I meant at the time.
AI is both helping and hindering. It can explain an isolated chunk of code. However, as someone who has learned coding in school recently, AI has hindered new coder's ability to read code if...
AI is both helping and hindering. It can explain an isolated chunk of code. However, as someone who has learned coding in school recently, AI has hindered new coder's ability to read code if relied on too heavily. Before chatGPT , when a bug is encountered, it may require looking at similar stackoverflow discussions, and understanding the code to be able to adapt to your situation. However, an over reliance on AI would tailor solutions for your exact situation, but it is limiting to smaller chunks of code. This smaller limitation makes it excel in school assignments when you need to write various small independent functions, but may not work well when working as a programmer with larger codebases. AI in my limited experience cannot explain more complex systems as the code is too interdependent on various parts of code, so for bigger systems people need to read the code to understand. However, some are just going to AI to explain small chunks of code, rather than learning to read it.
This is very true. The larger the codebase the less AI can help you with it. I can write basic and functional code just fine. Having AI available to double check ways to achieve something or to...
This is very true.
The larger the codebase the less AI can help you with it.
I can write basic and functional code just fine. Having AI available to double check ways to achieve something or to walk through a debug because you're not getting the output you expect is pretty handy. However, I'd have no issues tackling this on my own and have done for years, it just speeds up the process. The biggest thing is not to trust it at all! You have to understand what it writes and thoroughly check it before any usage as I've had so much useless garbage generated, it's not even funny.
Building LLMs into VsCode has been a nice edition and it has sped up development.
Yeah, there is definitely an advantage to AI into your workflow if you are experienced and do not have moral concerns with AI. Especially when it comes to writing boilerplate code. My concern is...
Yeah, there is definitely an advantage to AI into your workflow if you are experienced and do not have moral concerns with AI. Especially when it comes to writing boilerplate code. My concern is more that AI being accessible to beginners can be problematic as they don't get the practice on the smaller learning problems and then AI hits the limitation when they start working with bigger code bases. I think some of my opinions is also based on going back to school to learn programming in my mid 20s, when I knew how to learn but my classmates took the easy route of AI instead of learning how to learn and therefore did not learn some things like rewriting code or reading code well.
Claude, at the moment. It seems the most understanding with all the scripting that I write (Bash and PowerShell). I know I need to give o1 Mini a go. I'm happy to take suggestions for any that are...
Claude, at the moment. It seems the most understanding with all the scripting that I write (Bash and PowerShell).
I know I need to give o1 Mini a go. I'm happy to take suggestions for any that are better for my use case though.
I had several of the problems mentioned in the article under KDE Plasma, I'm so really looking forward to these fixes. Under windows, there's a third party utility called "DesktopOK" which brings...
I had several of the problems mentioned in the article under KDE Plasma, I'm so really looking forward to these fixes.
Under windows, there's a third party utility called "DesktopOK" which brings some sane management options (either manual or automatic) for desktop icons positioning and saving their states, but under KDE I was unable to find anything similar nor script some backup strategy.
A kindred spirit...
My people. The only icon I keep on my desktop is the recycle bin, and I only keep that around to throw away icons for programs that place a desktop icon without asking.
I use i3, which saves the problem of having anything on my desktop. :D
Launchers for life.
That was insightful. To the oblivious user desktop icons are something plain that should just work. But there's always so much hidden complexity behind all things computing.
I'm only a novice and occasional coder but this absolves me of some contrition:
Writing code seems so much easier than reading it. I can't be the only one spending copious time on thinking how minor variables should be named. I do this because I might have to readjust a script I wrote months or years later. And at that point I want it to read as autological as possible.
Ironically I like to keep my desktop clean. Just like a physical desktop. If I place something on the desktop it's because it needs attention, but I can only resolve it at a later time.
Reading code is hard but AI is helping with a lot of that. I've started using AI to help make my comments on my own code because what I write and what it writes are totally different. The AI is writing functional and logical comments, I'm writing things that make sense at the time but later or to a stranger would be complete nonsense out of context to what I meant at the time.
AI is both helping and hindering. It can explain an isolated chunk of code. However, as someone who has learned coding in school recently, AI has hindered new coder's ability to read code if relied on too heavily. Before chatGPT , when a bug is encountered, it may require looking at similar stackoverflow discussions, and understanding the code to be able to adapt to your situation. However, an over reliance on AI would tailor solutions for your exact situation, but it is limiting to smaller chunks of code. This smaller limitation makes it excel in school assignments when you need to write various small independent functions, but may not work well when working as a programmer with larger codebases. AI in my limited experience cannot explain more complex systems as the code is too interdependent on various parts of code, so for bigger systems people need to read the code to understand. However, some are just going to AI to explain small chunks of code, rather than learning to read it.
This is very true.
The larger the codebase the less AI can help you with it.
I can write basic and functional code just fine. Having AI available to double check ways to achieve something or to walk through a debug because you're not getting the output you expect is pretty handy. However, I'd have no issues tackling this on my own and have done for years, it just speeds up the process. The biggest thing is not to trust it at all! You have to understand what it writes and thoroughly check it before any usage as I've had so much useless garbage generated, it's not even funny.
Building LLMs into VsCode has been a nice edition and it has sped up development.
Yeah, there is definitely an advantage to AI into your workflow if you are experienced and do not have moral concerns with AI. Especially when it comes to writing boilerplate code. My concern is more that AI being accessible to beginners can be problematic as they don't get the practice on the smaller learning problems and then AI hits the limitation when they start working with bigger code bases. I think some of my opinions is also based on going back to school to learn programming in my mid 20s, when I knew how to learn but my classmates took the easy route of AI instead of learning how to learn and therefore did not learn some things like rewriting code or reading code well.
good call, which assistant are you using?
Claude, at the moment. It seems the most understanding with all the scripting that I write (Bash and PowerShell).
I know I need to give o1 Mini a go. I'm happy to take suggestions for any that are better for my use case though.
Oh, been wanting to try Claude but it's not available in my country, likely due to GDPR.
I had several of the problems mentioned in the article under KDE Plasma, I'm so really looking forward to these fixes.
Under windows, there's a third party utility called "DesktopOK" which brings some sane management options (either manual or automatic) for desktop icons positioning and saving their states, but under KDE I was unable to find anything similar nor script some backup strategy.