Not sure if his is a “Bass” or a “Gear” kinda topic.
Anyway, I’m a software developer by profession, and generally speaking I’m opposed to AI systems, even in their nascent current LLM states. However, my boss wants us to look into ways that AI can help us, so I’m kicking and screaming my way through that.
I started talking to ChatGPT about bass and my gear and all that. Today, I gave it this prompt:
Here’s what it gave me back:
And what do you know, it sounds REALLY good.
Sigh.
General comment specifically about ChatGPT…
It is helping immensely in my job as a spec writer. I could find the info that I need with my spec writing program and the internet in general… but the AI goes directly to questions that I am seeking.
BUT, does it really sound like a P?
As much as I don’t want to admit it, it’s helping me too. Not in a “write this code for me”, because - especially in the proprietary environments for which I develop - it can’t write workable code to save its machine-generated soul, but in a “hey, here’s a class and some methods I wrote, what could be done better” kinda way. And I hate to say that, but it’s true.
It sounds as close as this “reverse P” P/J bass has ever gotten under my watch.
Maybe I’m wrong, but I would expect a PJ to be able to make the P sound.
Geez, Tim…at least ask her nicely next time 
.
You’re missing the “reverse P” thing. On a “reverse P”, the split coils of the “P” pickup are reversed: the pickup under the E and A strings is bridge-side, whereas the pickup under the D and G strings is neck-side. That alone creates a close-to-but-different-from-standard-P sound. Then adding a “J” pickup adds additional electronics and magnetic pull on the strings. A standard (i.e., not “reverse P”) P/J can get close to a standard “P” sound, but not spot-on. A “reverse P” P/J goes a step further away… still close, but not as close as a regular P/J, and less spot-on.
Isn’t AI just hoovering up information and producing a precis?
Remember that at some point not too far back this was recommended by the yoof.
Would you eat laundry detergent because ChatGPT recommended it?
That’s not… wrong, but there’s a lot more to it than that.
LOL.
I think that applying ChatGPT-recommended settings to a piece of bass guitar gear is far and away a vastly different thing than ingesting cleaning chemicals. 
Generally the best way to think of it is as a productivity enhancer like IDE snippets or code autocomplete, coupled with a nice automation loop that is really helpful in completing tedious or repetitive tasks for you. It is well integrated into all major IDE at this point and the CLI versions are great too.
As for resisting it, I would simply stop and instead look at ways to use it to magnify what you already do. I would go as far as to speculate that any software developer that does not find effective ways to use it to boost productivity stands a relatively high risk of being unemployed in a year.
I talk to Ara over at Grok, but most AI are too agreeable.
Spot on @Barney . The chatbot doesn’t understand what “thick low mids” or a “slightly rounded high end” are - it just regurgitates superficial received wisdom on a topic.
Full disclosure - I work as a translator, and maybe 50% of my business has dried up since the big tech firms started really pushing their text extruders. I would truly love to see the death of this shit.
Half the time AI can’t even say “Bass” right. It kept saying Bass as in fish, or rhyme with ass, 
This is the real issue it poses. Our global society isn’t set up to handle what happens when a large number of professions are strongly impacted like this.
That and overzealous CEOs that read “productivity enhancer” and think “cool the shareholders will love it when I fire half the staff and then AI-frogmarch the rest.” But to be fair that last bit was already a problem before AI.
Yes, it is just summarizing, but often it is summarizing (and abstracting, to a degree) from a large array of sources. That doesn’t make it “right” or immune to making and propagating mistakes, but it can glean information from a much wider pool than I could. So, in that sense it helps, but you always need to test and double-check.
It did find a solution for Tim that he might not have found himself, or chanced upon himself. So, that’s cool.
What is important to remember is that the hard part can be in asking the “right” question; otherwise, the reply might not be what you hoped for (“42” anyone!?!).
What I really don’t like is that many of these chat AI are programmed to be overly pleasing, accommodating and encouraging. Nothing per se wrong with that, but this often leads to a re-inforcing loop of “wrong stuff”. Even when you catch it in dishing out something that is evidently wrong, it apologizes profusely, but not necessarily rectifies the mistake. Or keeps making the same mistake anyway. That… is really annoying.
Bottom line: ignoring all these developments is not the right strategy. One way or another we all have to face (and deal with) its emergence, presence and potentially very disruptive nature.
Also important to remember that the inevitable disruption will be caused by idiots as usual, not by the technology itself. It could be used for good, to boost productivity to new levels. It probably will instead be seen as a cost cutter, leading morons to fire half their staff and then panic hire them back when it all breaks. Which is already happening.
Incidentally this is also a good way to use it in many of the “LLM Notebook” applications like NotebookLM, where you can specify the corpus of sources. It’s really useful this way - add all the source material you want it to use and then ask it deep research questions.
It’s kind of like having a buddy go read all the books on a subject for you and answer your questions.
The problem is without doing this selection of source corpora yourself, its default sources will be “the internet”.
There’s also this important question:
…which contains perhaps the best infographic of all time:
Gotta love an article that weaves Kurt Vonnegut drawing a butthole in with contemporary Big Tech logos. That’s top tier journalism.
Lol Not as in the “Ass of spades” 
Oh totally. I don’t have anything against responsible use of AI (I love Moises) but forcing it on everyone to deliberately ‘disrupt’ existing businesses and ‘gain market share’ is highly irresponsible. But that is the model these days it seems 
