*(grabbing popcorn)*

I suspect I will agree with many elements of this take.

Okay, so....

How do I start this?

The epiphany happened today when I was looking at a post by someone I know who organises "NoCode meetup" events.

He's a really nice guy. I have attended a couple of the talks he's organised and plan to attend more in the future.

And well, I really like these "nocode" meetups I've been to so far.

Deploying interfaces for non-techy people to develop and edit websites themselves (without having to rely on me all the time) is 99% of what I strive to do in tech roles.

These people are my people, or at least they should be...

The last "nocode" event I went to was very AI (LLM) focused and, well, it failed to mention any of the negative things about AI.

I could have jumped in during the Q and A session to condemn the speaker and list off all the negative things about LLMs, and I probably should have.

However, I was mainly there that day to find potential clients and eat free pizza.

I didn't want to scare off potential clients and I didn't pay much attention to the talk anyway so couldn't really be very specific with my criticism.

Since this event, I've felt uncomfortable about this group. It's been weighing on me as to whether I should continue to attend these events and engage with these spaces or not.

I think I have a better idea today though...

*next slide*

So... well... today, I've refined my views on all this stuff.

I'm sure I've posted on here before about how LLMs can be used as a, sort of, ultra-high level programming language.

I'll link to this previous thread here (if and when I can find it). Until then, I'll summarise what I remember about it...

In that thread I wrote about some detailed videos I watched on Youtube, following individuals developing a game/application/whatever from scratch by only using an LLM to provide them with the code.

Now, I can watch these videos and see, very quickly, that I personally would gain very little benefit from the LLMs that were being used in these videos.

For me, it's much quicker to write "if (x > 0)" etc. than it would be to write out a paragraph describing the same statements.

But.. I have over 20 years of programming experience and seem to (for whatever reason) be attuned to how programming usually tends to be done.

The "no code" people I've met, however, simply don't vibe with how programming is conventially done and (you know what?) that's totally fine.

Programming, the way it is conventially done, does totally suck.

I can easily scoff at the headache of a process these youtubers went through when developing their software. But, the headache of a process I regularly go through when I develop software isn't exactly fantastic either.

For me, being adept at maths, enjoying problem solving, and having a fuckton of patience have allowed me to keep on going during those darkest hours of troubleshooting at 3am.

If other people don't have the patience for programming, in ye olden way, I don't blame them.

And so, what I'm saying here, is that I think it IS GOOD that people who previously havn't embraced developing their own software, have now found a way to embrace it - a way that, dare I say, "vibes" with them.

But, here's the nuanced part...

Ok, coming back to this now.

(There might be more large breaks between posts, so just, i dunno, bear that in mind)

Where was I? Nuance. Okay...

The nuance is, basically, that all the bad things people say about the current AI boom are correct.

- It's using up a shit ton of resources and damaging the environment
- The results it churns out are often innacurate, misleading or just shitty
- The companies using it are stealing other people's work to train their AIs without consent or compensation
- Many people are being fired and replaced with AI (despite the shit results AI churns out)
-It reinforces existing prejudices (e.g. racism) and is not actually "impartial and objective" as is often claimed
- Other reasons also, but I think that's enough to get the gist of it...

And so, I know that the AI boom right now is bad, very bad.

So, where am I going with this...

I guess the next point to bring up here and admit is....

*acts nervous*

I have been dabbling in using LLMs a little bit recently. Firstly, out of curiousity to test all the claims. But nowadays, I have to admit to myself, that I have found it useful. I do use them very sparingly though (don't hate me).

There are lots of people I see online who are saying we should not use LLMs at all, not one bit whatsoever. So, well, I am obviously concerned about riling up all the people who have that strong opinion.

So, I guess, next up I will talk about why I have been using LLMs and how I justify it to myself...

(whenever I get around to posting my next post in this thread...)

When it comes to the "three Rs" (Reading wRiting, aRithmetic), I'm pretty good at the math part. The other two, I'm not so confident with.

At college and university there was a big gap in my grades between tasks that involved programming / doing algebra and communication tasks.

I have such a lack of confidence in writing that I submitted not a single word in my final year dissertation at uni, but still managed to get my degree and a decent enough grade despite that.

As of recent, however, I have started to become far more confident in my writing ability and why is that? (I think you might know why...)

Like the "no code" folk I spoke about earlier, I started using an LLM to help me with my writing.

In this process I think I've been starting to feel a bit of a kinship towards what they've been experiencing.

I have been having to do a lot of writing recently, while applying for jobs, and I hate it and I have no idea what employers are looking for.

And so, at the suggestion of an employment advisor, I have been using LLMs to help with my cover letters and such and such.

Apart from once, as an experiment, I don't ever copy/paste precisely what the LLM suggests. And often the LLM will spit out some complete garbage.

It has helped me to, at least, have a starting point to work from, though.

And if all is lost, and I am really stuck, I can just send the LLM result - instead of the alternative... like what I submitted for my dissertation... which was absolutely nothing.

I'm no longer starting at nothing and stuck in dread, going nowhere. I'm starting at something that is passable, at least.

And I am pretty sure that this experience I have been having is similar to the experiences that the nocode folk have been going through.

But, yeah, these LLMs are bad. These companies are bad. How do I justify using these tools? Am I not just contributing to all the bad things that these LLMs represent?

*next slide*

*raises hand*

more of a comment than a question

Neurodivergent me finds LLMs very helpful to find the *shape* of neurotypical communication, though I rarely if ever trust them for the content. So I use it to remind myself what a cover letter to a particular org for a particular role might look like.

But the nondeterministic aspect makes it so frustrating as a code / no-code programming solution.

I am defo sure that being neurodivergent is a contributing reason towards why I have trouble with communication also.

I was openly told that I lost out on a job once because of my communication style (which happened to be very adhd like even if the interviewers didn't realise that).

And in terms of using LLMs for writing software, I sure don't get it no. But for people who find written communication a breeze, and maths/programming a wheeze, I can imagine their experiences being a mirror of mine.

As in, the nocoder folk probably also understand that the code results they get back aren't perfect, but the process of getting that code and altering the code (using natural language sentences and paragraphs) is comfortable to them.