The Divergence of Common Awareness of Ai and Musicians

  It's been obvious that there is a growing divergence between people who know, partially understand ai - neural network based technology, its current state and people who are unaware.  

 Or, worse, think it's capabilities are "science fiction".

 Likewise, I'm seeing this happen in musician circles.  Particularly with people referencing currently available real-neural network based music software, but also with the tertiary uses of LLMs and retrieval based voice replacement technology.

 There isn't even any cognitive dissonance.  There are simply a lot of people who think:

- Neural net based guitar software is "nothing new, it's been done before".  They think it's another variation on either digital eq, or one shot impulse response convolution;

- Voice replacement, "deep fake" software is either doing something with replaying samples, or is again, a technology that requires Impossible Technology that won't exist for many years.

 There may be a limitation on some people, as far as what they know of computer technology itself, that prevents comprehension of the revolution that is going on.  There may also be subconscious pushback happening, against what seems frightening and alienating.

 Regardless, there is a gap forming that is larger than what happened when digital recording took over.  Back when that happened, there were a lot of people that insisted "it will never be as good as a professional tape machine", that it was too complicated and limited to become a mainstream thing.  I am quite sure there are people who have effectively given up on a profession of recording, not because the music industry has fallen off, but because they literally got left behind because of their beliefs.

  As an aside, there is something similar happening in the computer repair/otaku world.  There seems to be many "tech advisors" that are totally unaware of the shift of graphic cards going from gaming to ai.  Some say things in response to hearing/reading press releases from NVidia (and others), that makes no sense - they're literally not accepting that GPU technology is being steered towards ai, and not video games. 

 There will be many disenfranchised musicians. 

 

That Time They Made Dark Side of the Moon New Coke

(... a warning to my guitar lesson students: if you bring this subject up be prepared for a whorling vortex of invective against this release.  I have not listened beyond the first 4 bars of "Breathe"; there is no point.  My vexation by this has ruined my day, and is probably only matched by my opinions about the abominations known as "Star Wars Episodes VII, VII and IX")

Gahr.  Someone decided to "make better" what is probably the closest to a perfect record that has ever been made?

It's of course, much louder.  Which in the 21st century means "better".

In doing this, they've changed the nature of the way the volume peaks change, and the way the low end sounds dynamically, and the tonal balance shifts around as the compression messes with it.

The way the volume peaks from the screaming intro is faster now. *The rate of which that occurs - which I presume came from Alan Parson's hand - is now different*.

The bass doesn't decay correctly.  The quarter notes previously poked out and then faded PERFECTLY.  IMO, this is one of the most important Pink Floyd signatures.  I remember that as a kid hearing it, *in my memory* it's a significant effect: boom BOOM baboom BOOooomm BOOM BOOM booboom...

Now it's just BOOM BOOM BOO BOOOOOMMM BOOMMM...

Ahrgh.

Likewise, the guitar strums are squashed down.  Despite the overall level being louder, the initial attack character of the guitar is gone.  The opening Em add9 is the OPENING SIGNATURE TO THE ALBUM.  It washes into the track.

 Now you have to "squint" as you listen to hear it at the start.  

And because of the overall level now being a completely different LUFS value, the entire vibe has been affected.  The pulse (ironic) of the low end is different.  The bass has to throb, the guitar has to have a "tinkly" effect - and level balanced to the bass (which it no longer is thanks to limiting/compression).  

 THE ADSR CHARACTERISTICS OF THIS RECORD IS SACRED.   

 I'm sure a lot of Monet paintings could be "made better" by color correcting sky tones, making the reds pop out more, maybe make some of the orientation more 16:9 friendly.  

Ahrgh.   #Ihatethe21stcentury.

Post AGI, Guitar and Art: Are You Tired of the Over Impressive Gimmick Yet?

  The end of human civilization, thousands of different ways?

 Eventually, "soon", there will be a robot that plays guitar solos.  Anyone want to take a bet?  Maybe even "rotoscoped" from video.  

 There is no reason it won't happen.  

Fake A.I. Prompted Image I Made Using DALL E

 
 What does that mean for the human pursuit of technical, kinesthetic facility?

 Nothing.  It will turn off a lot of people.  It will diminish the social perspective on another human accomplishment.  Will that change what you do, musically?

 It makes no difference to me.  I will continue to play guitar, and try to create what's in my head, a potpourri of different influences.  Influences that can be turned into a prompt for a LLM (large language model) like ChatGPT4, coupled to some sort of a GAN/ML audio generating system.  

 Except, making a prompt for a future LLM that will output exactly what I want will NOT be exactly what I want.  The detail required is redundant; I may as well do it myself.

 What I look forward to will be better SOUNDING tools, that are easier to deal with.  There will be "Engineer Mr. Scott from Star Trek" creators, directing a.i. to output art.  Which will be as fanciful and overblown as what we see today.

 Unfortunately for talented painters, illustrators, a.i. text prompts are cranking out Epic Landscapes like Doritos in a bag. It's gotten kind of boring, hasn't it?  For myself it has; I see people are still enthralled by endless comic book "kinesthetic super power display" movies, maybe it's no different?  

 I'm still as interested in French impressionism.  I still want to go see Monet, Renoir paintings in real life.  Am I imagining a subtle quality that I think I see, the human not in the machine?

 I don't think so.

 There is a difference between the Immediate Presentation - the first sip of Mountain Dew, versus the gestalt of it.  Don't we all know we're looking at something artificial, when viewing whatever Amazing Visual a.i. cranks out?

 Music seems to be lagging the visual.  It's way further along than most people realize, but it's just barely dipping it's toe in the Uncanny Valley.  But it will come out the other side eventually, soon.  

 But again, I doubt something will come out that affects everyone, universally, the way McCartney's "Yesterday" does, or Bach's "Air on the G String" does.  

 Not at first.

 The thing I don't look forward to is when some sociopath decides to train outputs based on human reinforced alignment from an emotion standpoint.  Because it's going to happen, and not for the right reasons.  Visuals, video, music, that is designed to affect humans emotionally. 

 Perhaps to a degree we're not prepared for.  This could be just one a.i. apocalypse scenario.  That sounds overblown if you're not aware of where the technology is already, but just consider the ramifications of such a thing.  The perfect car commercial, video and music that is subtly affecting you.  Or things in social media affecting in ways you're not aware of.  

 If we survive that, what's left is being intrinsically unique as a human.  What can you do that makes you a human?  Skillsets you have decided upon.  Accomplishments that you have bothered to do, from your will power.  Because at that moment, what I think we see now as the YouTube Effect - a futility based on being overwhelmed - will seemingly cover everything.

 We'll have to either try to be immune to it - reject seeing or listening to any media, lest we be swayed by the a.i. - or push past it *despite* it.   Or do we turn into non-emotional, grey aliens in a 100 years?  

 You are the sum of your experiences, trained in reality. Learning to appreciate music, guitar playing, is something an a.i. cannot do, because it will never have the context of being a human.  

 

Learn Guitar from Alan Alda?

   I saw an interesting little bit of advice from Alan "Hawkeye Pierce" Alda yesterday.  

 The gist of it was, 

repeat things you're trying to tell someone 3 times, 3 different ways.

 

With the additional aspect of

 only try to remember / learn 3 things at once, no more.

 I think Alda is a brilliant actor, but also has some pithy, wise things to say.  In the above, I think it's something I wish I could do better as a guitar teacher; convey a new concept or technique consistently 3 different ways, 3 times.  

 I've tried to do something like this as par for the course, but what Alda adds to the process is the (obvious) part whereby the person you are communicating with needs to be at the point of "oh yeah, yeah, right, I've got it" when you're at that third repeat.  

 I will tend to try to not annoy when I think a person has it; but perhaps that is wrong?  I probably should take it to that 3rd, "ok, you're annoying me" point?

 I'm not sure.

 But what I am sure about, is the part about remembering only 3 things.

 A lot of students want to cram as much as possible into a lesson.  My personal perspective is that the optimal lesson doesn't require me to send notes to the student, the student is familiar enough with the concept after the lesson that they know what to do without a reminder, for the following 7 days until the next guitar lesson.

 Alda's Limitation (as I'm going to coin it) is pretty logical, and concise.  If you think about most things you do, or want to do, do you really grasp or fully accomplish things when taken in more than 3 steps?

 I don't think so.  At least not when it comes to learning how to play guitar,

 The takeaway to this being, you should lean into to more fully digesting no more than 3 concepts/ideas/techniques in one go on guitar, at a time.  You're diluting your effort, and probably practically wasting your time otherwise.  It may seem *possible*, but that's not the same thing as *optimal*.

  

#1 Guitar Lesson Issue in the Year 2023?

  Skipping the fun.

 What do I mean by that?

 The fun of playing guitar is.....

PLAYING GUITAR.

 In my opinion, you should be playing along to music 10x or more than doing any kind of exercise.  Not only that, but for the most part, an "exercise" should really be in the context of a song.

 

 You don't wait until you reach Magical Guitar Wizard level before playing to a recording.  You do it constantly.  It is the best practice in itself, even when you're playing something you *think* is "easy".  Doing repetition never hurts your technique or skill.  If you only know one little part of a song - do it over and over and over, until it's like breathing.   

 You'll get much more benefit for the time put into it than doing 4 different things separately, by themselves, as an exercise.   The more you do something "professionally", the more professional you become.

ENNUI: Guitar Lesson Observations Today

The number one problem I'm seeing "these days" is 

ENNUI.

 A feeling of listlessness or lack of excitement.

 

  I think I grok the base cause of this, although it's very much hidden behind layers of Modern Sociological Sludge.

 Every major music genre - EVERY one, classical, baroque, big band, modern jazz, rock and roll, country, metal, nu-metal, mathcore, whatever - was energized in the following way:

1) An initial cadre of people struck out on their own, doing something that was unique and possibly heretical/unrelated to the "mainstream".

2) By serendipity what they were doing happened to get popular.  It's not necessarily by being great on it's own initially; it was the right thing at the right time.  

3) Secondary to that, people get excited - and build on it.   The excitement propels people forward, gets them to create and DO THINGS.

4)... until something new comes along.

  There isn't a clear cut New and Shiny Thing as there has been in the past.

 Step 1 is started by pioneers.  When you have a SUCCESSFUL pioneer, you don't see all of the unsuccessful ones.  People who may be equally worthy, but were at the wrong place at the wrong time.

 Jumping into music at Step 3 is easiest.  The problem is, we're at step 4.


  But the thing is, there is excitement to be had.  Buddy Holly didn't know he was going to be a star; not even the Beatles at first.  Eddie Van Halen, Jimi Hendrix are stars, but is Tosin Abasi a "star"?  How about Yngwie Malmsteen?   For all of these people, whether they're considered "stars" or not, they sort of pioneered a new style of music, with no distinct expectations.

 If we didn't know their names - and perhaps the reader *doesn't know who all of them are* - does it matter to them?  They were doing what they wanted to do, in a UNIQUE manner.  McCartney and Lennon didn't bemoan trying to write music that sounded EXACTLY like Elvis, they had fun writing music.  Music that was derivative at first, but after the initial 50 songs as they say, they got to the good stuff.  

 You've got to have a pioneer attitude today.  It's the time, and the way things work now is that by the time you realize Step 2 or 3 is happening, you're really back at Step 4 again.  

 So you may as well do your own thing, and get started with it now.