Brian AI Project.

[Sam_BFC]

To echo Hey Jude a little and repeat what I said on Endless Harmony, a brave move to replace existing vocals with AI generated ones.

Regarding Surf's Up, it is one of my absolute favourites, an yet - is it greed? - I can't help but wonder if a somehow "better" Surf's up could have existed. And the suggestion of harmony vocals during the "aboard a tidal wave" section is tantalizing.

[HeyJude]

Interestingly, this particular work actually has enough decisions that are outside of the “do these AI vocals sound accurate?” debate that it’s actually in some ways easier to debate those.

I think that version of “Surf’s Up” sounds like kind of a mess. I guess, because it’s using a weird AI-assisted vocal, it manages to sync up the “Brian vocal” to the ’66 backing track a bit better than either fan-made sync jobs or the official releases that sync Brian’s ’66 demo vocal to the ’66 backing track. But just about everything else sounds wonky to me. The weird key change at the end sounds totally off. The key change on that discovered “Wild Honey” demo of “Surf’s Up” works because of the nature of that solo piano Brian performance. But pitch-shifting the entire track on this “finished” version sounds really strange.

I obviously have my own take/ethos on this song and the material in general. Much like I think Brian/BB fans ignoring “Guess I’m Dumb” because Glen Campbell is singing it and they can’t have a Brian lead vocal is kind of missing out on a great song with a nice vocal (Glen sounds good on it, and sings it like Brian), I think there’s a point where the novelty wears off on trying to shove Brian on the lead on the first movement of “Surf’s Up” and it makes more sense to listen to what I think is the best version of the song, the 1971 version with Carl singing it. Carl sounds amazing on it, it retains the “Smile” vibe while fleshing the song out with a *very* important bass line near the end, and amazing group backing vocals, not to mention Al’s excellent part at the end. I’m not opposed to the Brian “what if” vocal as heard on the “Smile Sessions” (and that fans had done years before to varying degrees of success). But at the end of the day, in a world where I’m happy to admit when the “official” version of something is *not* the best, I can safely say the 1971 version of “Surf’s Up” released on the album of the same name is the definitive, and best version of the song. Better than the ’66 paste jobs, certainly better than the weird AI versions, and better than Brian solo versions from the 2000s.

I mean, I guess it takes some balls (it takes something anyway) to go beyond the AI generated leads for non-existent stuff and just decide to erase actual Brian/BB leads that are definitive and re-do them with weird AI, and then also start grafting on weird overdubs. I think this stuff, already questionable with the AI vocals, quickly becomes a weird mess when it’s getting futzed with on multiple levels in multiple ways. Yes, yes, I know, “it’s just for fun”, etc. I’m not saying anybody wants to delete the BB albums and put this out. But I gotta rate it and analyze it for what it is, and as much as some fans seem to really, really want this to be a mind-blowing thing (and if it does that for you, I’m of course not going to stop anyone), but to me it’s a weird melting pot of the questionable AI vocal stuff, brazen but ultimately sub-par artistic liberties, head-scratching “fan mix” decisions to move around tracks, and a weird overall goal/ethos (e.g. This seems to be a weird, truncated “Smile” that’s missing key things like “Look” and “Holidays”, etc.).

And, in the realm of the actual AI voices, there are a few here where I honestly couldn’t tell who it was trying to sound like. I think at one point it’s going for 1966/67 Dennis, but I guess the AI has less 60s Dennis vocals to learn from, so that one sounds particularly off.

Also, one more thing and I’ll take a breather: While I guess on some level it’s interesting to hear the alternate phrasing/melody for the “Do You Like Worms” verses employed, I think the 2004 iteration sounds much smoother and less clunky.

[brother john]

I think the whole thing is amazing. A lot of very creative decisions have been made throughout, and I love some of the new backing vocals. On just two listens nothing about it annoys me, except perhaps the strings in Surf's Up which sound a little weedy, and the French Horn tag at the end of same which is just silly, and breaks the mood somewhat. Overall a stunning effort.

[Awesoman]

There's definitely some spine-tingling moments from this version.  I'm not sure that I like the overall sequencing of the songs though and was there any reason to omit "Good Vibrations" which was the song that tied the whole thing together?  Also some of the wacky tempo changes seem jarring to me.

But on the plus side it was nice hearing a fully-realized version of "I Wanna Be Around" and kudos to Dae Lims for being daring enough to update "Surf's Up".  This is probably one of the most remarkable fan edits of SMiLE that I've ever heard.

[pixletwin]

There's definitely some spine-tingling moments from this version.  I'm not sure that I like the overall sequencing of the songs though and was there any reason to omit "Good Vibrations" which was the song that tied the whole thing together?  Also some of the wacky tempo changes seem jarring to me.

But on the plus side it was nice hearing a fully-realized version of "I Wanna Be Around" and kudos to Dae Lims for being daring enough to update "Surf's Up".  This is probably one of the most remarkable fan edits of SMiLE that I've ever heard.

Agreed. I'm not into the sequencing either, but an amazing mix. Surf's Up is inspired.

Like it or not, AI is the future of music production.

[rab2591]

I caught a little bit of Smile AD's livestream today where he talked about the making of these tracks. For those who don't really dig this mix, it may be worth scanning through those 'making-of' livestreams. Not because it'll change your mind about how you feel about the mix (though it might), but it really shows the amount of time, energy, and care that went into this project. Just from a technical perspective, this mix is an astonishing achievement. And hearing Smile AD talk about his reasoning for mixing things certain ways, adding in AI vocals in certain places, it really makes sense once you listen to his reasoning.

I was blown away at how many tracks he had for each song - double tracked vocals, bass lines that were beautifully (and seemingly perfectly) extracted so he could mix them better. The clippings/edits he did just to one of his vocal tracks was very cool to see. Other highlights were how he added in various sounds to help achieve a more impactful sound. He would scroll up/down through the tracks for whatever song he was talking about, and I feel like sometimes it was an endless scroll because he had so many tracks (in a good way, it just shows how much work he put into each of these songs).

The artist/musician in me is completely floored by the creativity of this mix. There are so many beautiful moments. Not to mention how AI Brian, Mike, Al, etc sound just like the real thing.

Some songs I'm not too crazy about, some have really grown on me (especially Do You Like Worms - I love how he tackled the melody of the 2nd half of the verses - but I admit it took some getting used to!). But overall I love how genuine it sounds - genuine in both quality and in the happiness the tracks evoke...and I can't stress that enough. It really is worthy of a smile!

Above anything else, I have to commend Smile AD for not being afraid to try new things.

[All Summer Long]

I caught a little bit of Smile AD's livestream today where he talked about the making of these tracks. For those who don't really dig this mix, it may be worth scanning through those 'making-of' livestreams. Not because it'll change your mind about how you feel about the mix (though it might), but it really shows the amount of time, energy, and care that went into this project. Just from a technical perspective, this mix is an astonishing achievement. And hearing Smile AD talk about his reasoning for mixing things certain ways, adding in AI vocals in certain places, it really makes sense once you listen to his reasoning.

I was blown away at how many tracks he had for each song - double tracked vocals, bass lines that were beautifully (and seemingly perfectly) extracted so he could mix them better. The clippings/edits he did just to one of his vocal tracks was very cool to see. Other highlights were how he added in various sounds to help achieve a more impactful sound. He would scroll up/down through the tracks for whatever song he was talking about, and I feel like sometimes it was an endless scroll because he had so many tracks (in a good way, it just shows how much work he put into each of these songs).

I know this wasn’t directed at me, rab, but I realized I should (briefly) further explain my thoughts on the AI topic, since I can’t remember if I said one specific part before. I am not denying the effort that some users such as Smile AD are putting into these projects (though I will still insist that many other users are just doing something without effort for the sake of it). I definitely understand the intrigue and what if’s regarding the band’s music. There is likely some alternate or something that never existed or used to exist that each and every one of us wants to hear. However, I don’t understand why we’d listen to AI recreate The Beach Boys when we have the real thing, and so much of it? The amount of material the band has released (solo material included) is a ton of material, let alone the amounts of alternates and unreleased and/or previously unknown material surfacing constantly on the archive sets (since 1981/1993, but especially since 2013). Ok, I think I’ve bored you all enough now.  I’ll keep quiet.

[rab2591]

I caught a little bit of Smile AD's livestream today where he talked about the making of these tracks. For those who don't really dig this mix, it may be worth scanning through those 'making-of' livestreams. Not because it'll change your mind about how you feel about the mix (though it might), but it really shows the amount of time, energy, and care that went into this project. Just from a technical perspective, this mix is an astonishing achievement. And hearing Smile AD talk about his reasoning for mixing things certain ways, adding in AI vocals in certain places, it really makes sense once you listen to his reasoning.

I was blown away at how many tracks he had for each song - double tracked vocals, bass lines that were beautifully (and seemingly perfectly) extracted so he could mix them better. The clippings/edits he did just to one of his vocal tracks was very cool to see. Other highlights were how he added in various sounds to help achieve a more impactful sound. He would scroll up/down through the tracks for whatever song he was talking about, and I feel like sometimes it was an endless scroll because he had so many tracks (in a good way, it just shows how much work he put into each of these songs).

I know this wasn’t directed at me, rab, but I realized I should (briefly) further explain my thoughts on the AI topic, since I can’t remember if I said one specific part before. I am not denying the effort that some users such as Smile AD are putting into these projects (though I will still insist that many other users are just doing something without effort for the sake of it). I definitely understand the intrigue and what if’s regarding the band’s music. There is likely some alternate or something that never existed or used to exist that each and every one of us wants to hear. However, I don’t understand why we’d listen to AI recreate The Beach Boys when we have the real thing, and so much of it? The amount of material the band has released (solo material included) is a ton of material, let alone the amounts of alternates and unreleased and/or previously unknown material surfacing constantly on the archive sets (since 1981/1993, but especially since 2013). Ok, I think I’ve bored you all enough now.  I’ll keep quiet.

I do understand where you're coming from. From my point of view, while the real thing is SO much better, but it's also kinda neat to step into an alternate world for a few minutes and pretend The Beach Boys completed something that never was.

The only three projects associated with this AI trend that have truly captured my long-term interest are the recreation of 'Thank Him', and FOTM's & Smile AD's Smile projects. For me, it's the act of completion of something that should have been completed long ago. So I think that's where my interest stems from and likely where it will stay. The Beach Boys recording SpongeBob music, or Brian covering the Ronettes - while technically interesting, doesn't hold the same value to me, because it's not completing something that was, historically, almost in the grasp of the band.

But that's just me. Honestly, I'm still on the fence about the morality of it all, and I worry about what could come from all of this (the entirety of the AI boom). For now, I'm glad that people are using it creatively, in an attempt to bring more interest to the band.

[Old Rake]

https://www.youtube.com/watch?v=MDcJKC5RWpI

Wanted to post a link to my AI-enhanced version of Worms. I tired to keep it a little more minimalistic -- my verse vocal is a little bit like a stripped down version of Dae Lims' version, that was pretty much where I was heading too. Also added in the "East or West Indes" vocal at the end. I think maybe it's a little more historically accurate than his but less fanciful? If that makes sense?

I am working on the others!

[El Molé]

The Daelims AI Smile is my favourite ever Smile listening experience, without a doubt. Smiley Smile, GV box Smile tracks, bootlegs, the H&V demo from Endless Harmony, BWPS, Smile Sessions etc - I love it all, but I've never enjoyed a fully formed mix as much as I've enjoyed this Daelims one. Given where things are with the development of AI and how quickly this was created, I'm stunned at how good the vocals are, and the Q&A youtube stuff was really illuminating in terms of choices and why AI vocals were used in certain places. The expansion of the backing tracks to enhance etc is just as impressive to me.

I'm not a purist, but I can enjoy the integrity of the Pet Sounds mono mix (Wouldn't It Be Nice in mono is as good as pop music ever gets) and the Stereo, and the sessions etc. BWPS is stunning as a completed album, but the vocals will never compete with the Beach Boys own voices, no matter how well executed. I have always found any blend of BWPS vocals with original tracks jarring and distracting in a way that I simply don't find with this AI mix. There is no historically accurate final Smile mix by the Beach Boys and there never will be, but my own view is that this is the closest anyone has come to a full-album mix that is cohesive and not hampered by loads of missing vocals or varying fidelity/sonic issues etc. This mix is an extraordinary piece of work and I love every second of it.

There is no perfection possible with this sort of thing - we'll all differ in how we experience it and what we prefer, and I'm still working out what I think about the ethics of AI in general etc, but for anyone who might want to hear the Beach Boys sing In Blue Hawaii, then what's not to like about this? Or a version of You Were My Sunshine without Dennis being squashed in the mix - listen here for that too.

I don't disagree particularly with views of those that aren't so keen (different perspectives etc) or who simply prefer to hear the original voices rather than AI recreations etc - but this whole topic (smile, this board etc) was always very much based around a celebration of a lost album, missing pieces and the many "what if" scenarios and this mix responds to so much of that with an incredible sounding response. I have literally woken from dreams in which I'd heard rediscovered original Beach Boys vocals on some of these tracks that were either never recorded or lost, but in this mix I can hear them and they're just as amazing as the things I heard in my dreams. To someone like me, this mix is an extraordinary gift and I am beyond grateful that someone took so much time and effort to make something so wonderful.

[El Molé]

https://www.youtube.com/watch?v=MDcJKC5RWpI

Wanted to post a link to my AI-enhanced version of Worms. I tired to keep it a little more minimalistic -- my verse vocal is a little bit like a stripped down version of Dae Lims' version, that was pretty much where I was heading too. Also added in the "East or West Indes" vocal at the end. I think maybe it's a little more historically accurate than his but less fanciful? If that makes sense?

I am working on the others!

Love the East or West Indies bit! Good work! Also just picked up on your Smiley Smiley remix - great to listen to.

[Acechaser]

I like the idea of using AI to fix or complete BB/BW songs.  With respect to fixing, the intro to The Spirit of Rock and Roll (by BW) on the Hallmark album is a top candidate for repair.  Other candidates include several songs from BW's Getting In Over My Head and the Danny Hutton duet (Hold Back Time) on the VDP/BW Orange Crate Art album. 

[guitarfool2002]

Jon Hunt's "Smiley Smile" remix is DEFINITELY worth a listen...very, very cool. Smile with the Smiley Smile production and sound. For those who haven't heard it:

https://www.youtube.com/watch?v=Fh5v80aaweM

[Don Malcolm]

AI and creativity raises a fascinating complex of issues (in part because it is also so harrowing in its implications). That said, I think we need to give folks leeway to try things out as much as possible--within some established and fairly exacting set of parameters. In saying that I am focusing mostly on music/musical reconstruction as we've begun to experience it in general (and, of course, specifically with SMiLE--a project that in its still-infinite aftermath is made for AI tools and approaches.

For SMiLE as tackled by Dae Lims, we're probably going to need a full transcription of the vocal reconstructions in order to fully assess the ultimate level of efficacy in his SMiLE. That may be possible for some unusually enterprising individual who can collate his lengthy explanations of his process. Of course, such a compendium won't eliminate lingering skepticisms about the decisions made (particularly the choices between AI vocals and actual vocals). But we probably need to see all of that in some kind of consolidated flow chart or other type of summary before we can draw any truly substantive conclusions.

One thing that I hope we'll get from one of the SS stalwarts is a song-by-song interpretation/analysis, which doesn't seem to have emerged yet. I wonder if one or two of our folks who are established as research experts might take a crack at that at some point. Those folk may wish to tackle the subject of "liberties" taken and how likely some of those choices might really have been, just for the sake of grounding that in a way where it can be dealt with and not continually threaten to sidetrack the discussion. Dae Lims made some major choices (particularly in jettisoning GV, Look and Holidays) in making what we might call the January '67 version of SMiLE (although with some of the '71 "Surf's Up" managing to travel back in time).

It's clearly a very personal mix, and a labor of love, and we should grant him license to build such a version as a testament to what we can never know, but still have an intense desire to imagine--and now we can hear so many more things that can be imagined than was previously the case. I hope this is just the beginning of an effort to produce "editions of the in situ SMiLE" that will have similar astonishing results. BWPS is clearly not the SMiLE that would have emerged if it had not gotten sidetracked, pole-axed, abandoned and reconfigured, so all's fair in love and ProTools. Let this be just the beginning!

And on such a note, let me second GF's endorsement of the "Smiley" SMiLE by Jon Hunt. More food for thought, more intriguing sonic variations for music that still has mysteries left to be revealed to us...

[HeyJude]

I’m not trying to be reductive, but I think that if one take the maker of these AI projects at their word, and I see no reason for them to lie about their motivations/reasoning for their decisions, then I don’t think analyzing it is too complicated. Meaning, the people on this board know their stuff, and can pretty immediately spot what’s AI and what isn’t, and also know the stuff well enough to develop an opinion about what they like and what they don’t like.

I guess, for folks that are *really* into this stuff, flow charts and detailed notes might make it all more enjoyable. For me, as someone who, setting aside the more broad implications of the use of AI on music (a most decidedly complicated topic), finds projects like this to be more a curio than anything else, I don’t need voluminous notes and charts. I can detect what’s going on more than well enough to *understand* it; I don’t need the precise reasoning behind each of the decisions. Again, this is all said with an understanding that I trust the maker of the project is being honest and that they certainly feel they had reasons for the decisions they made.

I'm not too concerned with giving people leeway or license to make this stuff. They're going to do it and are doing it. Nobody is stopping them. I truly actually do keep an open mind about this stuff. I think the AI music stuff online went off the rails long ago (if it was ever on the rails in the first place). This particular maker of this AI Smile, and their previous work, is really an anomaly as far as time spent, knowledge of the source material, and proper labeling of the end result. I appreciate all of that. But even this maker's stuff has kind of shot past what *I* feel is a possibly useful ethos in making the stuff, and has landed in fan mix/fantasy tracklisting territory, all while using those AI tools. As I said in a previous post, it's not so much a question of "What's the point?", and more a question of "Is my time well spent listening to this stuff rather than X?".

[HeyJude]

I think one of the issues here is the push and pull between AI creations as a technical application of software, versus being actual “art.”

We all have varying degrees of familiarity not only with AI-sourced material, but varying degrees of experience coming up upon (or against) people who create it and/or are very bullish about it (not surprisingly, those two things are usually one and the same).

However many months back (not that it has let up more recently) when people started posting visual/digital art and saying “There, see? I’m an artist now too!”, it immediately got pretty icky. There was and is a large chasm between people who feel those people are *not* artists versus those who feel they are, and also a wide gap between the people who think the art they’re executing is just awful versus those who think it’s amazing.

When it comes to AI-sourced music (meaning mainly in this case generating fake voices meant to sound like a real person, using AI trained on recordings of that person), there is a very similar gap on the *reception* side of things. Some people listen to this stuff and think it sounds amazing, others think it sounds awful. Also, within the group of people who think it sounds *accurate*, there is some level of disagreement on the choice to make the stuff, or the implementation of it.

My subjective observation has been that, especially because a lot of the AI music stuff has been admittedly done to sound like a person, trained on recordings of that person, the people *making* this stuff have generally seemed to be much less egotistical and display much less hubris about what they’re doing as opposed to, say, visual AI artists. I’ve gotten the sense in the past that, when seeking out a specific prompt (e.g. make this recording sound like *that* person from *that* year), the people making this stuff have given off a vibe less of “I’m an artiste!!!” and more like “This is a technical process that requires some level of skill.” While that doesn’t address the misgivings some have about the choice to make it in the first place, it is a lack of bullishness that I appreciate.

However, regardless of the intent or the expressed level of humility or hubris from any given person making this stuff, when it goes from a specific prompt of just filling in a vocal to sound like a specific person at a specific time, and moves to making subjective, artistic decisions about melodies, chords, arrangements, and eventually ends with wiping actual existing vocals to then replicate those *very same vocals* with AI, that’s where it gets a little more dicey. I’m not even talking about the morals/ethics and all of that. Just in terms of how I feel about the end product, that level of “fan mix” added to the “AI” thing makes the whole thing just a lot more wonky.

And then that leads to the question of how to talk about this stuff. I have no desire to sh*t on someone after they spend hours and hours working on something, whatever it is. But also, the presentation suggests wanting people to consume it and provide feedback. So there’s an uncomfortable point where it ends up being a reaction of “Uh, yeah, I’m not sure if it’s worth that much effort to produce *that*.” But clearly there is an audience for this stuff, so I don’t feel too bad saying maybe *I* think my time is better spent listening to something real and made with whatever original artistic intent it was, rather than listening to too much of these large scale AI music projects, because I don’t think my mixed feelings about it are going to hurt anybody’s feelings.

[HeyJude]

I like the idea of using AI to fix or complete BB/BW songs.  With respect to fixing, the intro to The Spirit of Rock and Roll (by BW) on the Hallmark album is a top candidate for repair.  Other candidates include several songs from BW's Getting In Over My Head and the Danny Hutton duet (Hold Back Time) on the VDP/BW Orange Crate Art album. 

I'd say that 2006 Hallmark take on "The Spirit of Rock and Roll" is too limp and lethargic to salvage. It's got more problems going on than some wonky vocals.

I also think the "Orange Crate Art" material is fine as is. It was already futzed with back in 1995, and frankly the "1995 Brian voice" is a specific flavor that I think is worth preserving. Note how his '99 re-done lead on "You're Still a Mystery", for all its smoothness, lacks of the energy and quirk of the '95 vocal.

Yes, I believe "Hold Back Time" is the one that sounds wonky enough that one rumor persisted for a while that the album "accidentally" used a guide/temp set of vocals. I don't know if that's the case, but if it is then it implies a later, more polished set of vocals exist somewhere, which is the main thing I'd want to hear. If it actually exists.

I think the only way to "fix" this later-era solo stuff would be to still make it sound like latter-day Brian, just more polished. And I don't know what the efficacy would be (or possibly the point would be) of making "2004 Brian" sound more like, what?, 1998 Brian?

[HeyJude]

I think at this stage that the AI-assisted isolation/demixing technology is a million times more interesting and useful (and ethical if that's one's concern). I love Brian's productions, but it would be interesting to hear some of the backing tracks pulled apart more. Joshilyn's videos have certainly opened many eyes to all the intricate stuff buried in those recordings/mixes.

Another interesting application? Vocals. AI is getting better at separating multiple singing voices locked into one microphone/track. I'd *love* to hear the harmonies picked apart further. I was recently listening to some Beatles vocals pulled apart (e.g. the separate Paul, John, and George parts on "Eleanor Rigby" and "Because"), and it's super interesting. It really drives home how those middle parts, those George Harrison/Al Jardine parts, are some of the most difficult.

[FatherOfTheMan Sr101]

Some thoughts: https://youtu.be/nrEoGv13Cm8?si=yovpp-n2WHO2GqL1

[FatherOfTheMan Sr101]

Can't wait till you guys hear the rest of this, but I wanted to show a bit of how the sausage is made.
So many folks have the wrong idea about these tools!

https://www.youtube.com/watch?v=Tnm_DW0azIc

Any guesses what this goes to? Hmmm...

Oh! And these models, much better ones than my Brian model, will be available soon as well. 

[FatherOfTheMan Sr101]

BOOKOFF SMILE PLUS - Completely new ai models, tracks, and mix! Coming... when it's done!

https://www.youtube.com/watch?v=ep_6aF32OjI

Song for Children - featuring Al, Mike, Brian and Carl!

https://www.youtube.com/watch?v=C_LBW8qDZFM

[thelonleyc]

><

[FatherOfTheMan Sr101]

Fair enough :/ This is literally a bedroom project ahaha, I think it sounds pretty good but... valid criticism!

[Don Malcolm]

I think the constructive criticism here should be based on a better understanding of the "training" process and how it's applied to each song. In comparing Mike's work to Dae Lims (you folks out there realize, of course, that "Dae Lims" is "Smile AD" backwards, yes?!) we should ask whose models/process has produced the most credible "AI surrogate" vocals, and why this is the case--better vocalists being used as "source," better/more exhaustive-thus-closer-to-the-sound-timbre-vocal style of each actual BB, etc.

AI has got to get closer and closer to a more authentic (if that's the right word, and if it's actually possible...) replication of the BBs voices in order for re-envisioning of vocal arrangements that never made it to tape in '66/'67 to pass muster. Many of them kinda sorta do that now, but this is the battleground for whether this promising area of work can go beyond the stage of being a fascinating curiosity. Dae Lims has (IMO) gotten there on more than a few occasions; Mike is lagging behind in that respect, but his work still holds promise.

It's clear how much work is involved in all of this, so none of the above is meant in any way to be perjorative or judgmental, and I hope it's not taken that way. Hoping that folks will keep at this, as the music is special and unique enough (and still tantalizingly incomplete despite the existence of BWPS '04) to warrant ongoing investigation and engagement...

[guitarfool2002]

It's the advance in the technology that is pretty stunning, in that none of this was available nor was it happening even a short few years ago. What did exist was pretty bad, and now at least the median standard is that it could be called "convincing" in the average AI modeling products. I compare it to the various technologies that were used back when they cleaned up the Beatles BBC tapes and released them: At the time, it was new and exciting, however now it can sound kind of wonky and primitive. And look where the extraction and cleaning technology has progressed solely in the Beatles' universe, where Jackson's Get Back and the newest remixes used that same extraction software to basically reinvent the original mixes.

So that's 30 years difference, from BBC and Anthology to Get Back and Revolver. Compare that to where AI was in 2021 to where it is now, and it's staggering in how fast the tech progressed. I predict within a year, there will be AI that will allow fans to create full videos for a random song, using AI technology to create the imagery of the artist, the music, and whatever visual cues the maker wants to feed into the AI generator to create the video. And then consider what that will mean for film and the whole process of making a movie from writing the script to shooting the scenes to editing it all together. That's the area which I think many have issues with. It's already affecting the world of academia where ostensibly a student could write and hand in a paper generated with AI and the teachers/professors would have no way to know if it were the student or the AI creating it.