ok not me but this youtuber wheezywaiter:

I used to watch this guy religiously, but recently got back into watching his content. His channel is full of 30/60/90-day challenges where he does something outside of his comfort zone and reflects at the end of the challenge.

This one caught my attention big-time though because it's something that I actually could see myself using AI for and really opened my eyes to the possibilities.

anyone here using AI like this? And if so, please tell me more because I figured AI would be a useful tool in this way but it's almost surreal seeing it being used like this.

Imma be honest, yo. I'm working on a full redesign of my site. And, I'm no webdev, and sometimes I get really fuggin stuck. Whether with my static site generator (Hugo), or programming (Go, Javascript), or CSS, and I try and I search and I fail. And I go over to the fuggin Chatbot, and as long as I know what I'm talking about, and I know how to communicate it, usually, at the very least, if it doesn't give me the answer, it points me in a better direction. It's been a lot of work, and I honestly can't tell you if I'd have made it this far without the bot, but I'm pretty happy its around for some use cases. I think the bot makes it easier to not give up on some things, and that's really what progressing and moving forward and improvement is ever really about, keeping on, ya' know?

Sorry, maybe you were going for the more irl ideas of ai, but I think this is pretty related. In the video's case: Someone who wants to get healthier / get better at cooking and really has no idea how to go about it. Just being told a flat out general direction, being pointed somewhere, that can do a lot.

I think you can ask people with more experience in web development to help you with your website, before using an AI chatbot. Sure, it can be helpful for some small things, but for a whole website? I wouldn't do that.

Oh thanks Geti, that means a lot. Part of it is, really, not wanting to bother anybody. Everyone's busy with their own stuff. You can't bother the chatbot. But no, don't worry it's not making my site. It's more like when I want to create a feature and I get stuck, and I ask for help from the bot. I know every line of code I'm putting in and I'm understanding and learning it. There's definitely a fine line here though: "Oh this bot gave me code and I'm gonna copy and paste" vs. "this bot gave me this approach, let's study it, see how it works, figure out why it works, decide if I like it." It really is helping me learn, I feel.

I get your point. I think that's the correct way to use a service like ChatGPT, although I still think that having good documentation for whatever you use is invaluable. By the way that reminds me of the only time I (indirectly) used it to fix an issue with a service I host.

Again, I feel bad for slightly side-tracking this thread, but I found this video of essentially some normie chick talking about making a website, and about halfway through it the way she brings up the chatbot, I thought was pretty funny:

totally fine i posted this in general for a reason!

i really wonder what percentage of people making personal sites are doing this exact thing because learning CSS early on, for example, can cause extreme imposter syndrome, at least for me.

im still begrudgingly not using AI for coding, and i probably should because it would save me a ton of time, but my head is just not there yet.

I haven't been using LLMs at all. I don't trust them, I don't like that the data used to train them wasn't necessarily ethically acquired, and I didn't spend decades learning so that I could become dependent on software that is little better than the ELIZA-style chatbot I can access in my text editor by typing M-x doctor.

And, quite frankly, I don't trust anything coming out of Silicon Valley nowadays. I don't think any of you should, either.

This is my opinion as well.

It drives me up a wall that everyone is calling it "AI" too, because there's actually no intelligence involved. Yet here we are, everyone is saying "AI" this, or "AI" that, because to the layman it looks like the computer is being very smart, when it's been curated, and for every really convincing thing generated by it, someone had to swipe left or right on it several times until it spit out something convincing enough to post it online.

Sorry to be that guy but that is kind of a silly take. It's called AI because it is AI. It isn't sentient, it's not alive, but it possesses the capacity to respond intelligently and do some forms of basic reasoning. It's also the case that none of the leading products out there are just the ML-derived neural model. An AI system will be comprised, as are our organic intelligences, of systems of systems. As far as the baseline frontier LLMs, the capabilities line continues to go up, most notable recent example being today's release of LLaMa 3.

AI assistants can save huge amounts of time and bolster productivity for non-expert users a tremendous amount. For the next decade we'll basically be distilling the 15% of human intelligence not already captured by the machine as it turns us all into data sausage [I am not gleeful about this or in support of it, just noting the inevitable and correcting the record as far as semantics]. Currently they're not expert level but there will be umpteen approaches to structuralizing the monetization of expert distillation.