{ "schemaVersion": "v1", "slug": "ai-hallucination-and-creating-ordered-meaning-spaces", "title": "AI \"Hallucination\" and Creating Ordered Meaning Spaces", "kind": "solo", "listenUrl": "https://podcasters.spotify.com/pod/show/lovemusicmore/episodes/AI-Hallucination-and-Creating-Ordered-Meaning-Spaces-e2hrfal", "publishedAt": "2026-10-14", "source": { "provider": "whisper", "model": "faster-whisper:base", "normalizedSchema": "v1", "transcribedAt": "2026-06-09T08:21:13.029Z" }, "paragraphs": [ { "startSec": 19.15, "endSec": 46.07, "text": "What's good, everybody Scoobert Doobert, and here for this solo version of love music moreI'm really excited to talk to you guys because we're gonna talk AI againBut in a little bit of a different way, so we're gonna zoom out a bit talk about what creativity isWhat AI is what neural networks are and then try and weave it all together so that you can differentiate betweenfaking it and actually making it making the music and" }, { "startSec": 46.91, "endSec": 73.11, "text": "And then you know decide for yourself from thereBut there's a lot of like weird misinformation out there and I'm excited to break down a little bitHow this actually works from the lens of the musician from the producer from the creativerather than just from thetechnologist that wants toYou know inflate their stock prices and whatnotSo I'm excited to get into that and it's gonna be a great little mini" }, { "startSec": 73.59, "endSec": 211.27, "text": "Synopsis of some of the other pods that we've had on on this show, but in a slightly different wayAll right, so how does AI even work?So this is this is one that you knowI've talked to a lot of people that actually do this stuff right that" }, { "startSec": 211.67, "endSec": 224.47, "text": "Like are AI either on the science side or the engineering side because there really are two different sidesThere's a side that says it works greatAnd then there's the other side that says it works how" }, { "startSec": 225.75, "endSec": 231.27, "text": "And I would say that the engineering side's doing a little bit better right now because there's some things that are just working" }, { "startSec": 232.71, "endSec": 261.27, "text": "Because you basically uploaded the entire internet into this thing and for some reason there's some emergentCharacteristics that's a word that they throw out a lot emergent in the same way that consciousness emergesBut one of the things that I alwaysKind of try to take care of or be careful ofIs that there are persuasive tools and the ways that they're namingThese features the ways that they're naming these" }, { "startSec": 262.55, "endSec": 268.31, "text": "algorithmsTo make them seem more sentient to make them seem more than they actually are" }, { "startSec": 269.35, "endSec": 280.05100000000004, "text": "So a good example of this is you know on iPhone and I'm pretty sure android's got this tooIf you type in a word. Let's just I'll do it right now." }, { "startSec": 280.05100000000004, "endSec": 307.91, "text": "I'm going to type in theThen it has a kind of next wordSuggested for you below and what it's really doing is it saying okayI have seen the a lot of time so I'm going to based on either what I have typed or what it has been trained onSuggest the next word So the it's going to go the it suggests first only most I'm going to go mostThe most suggests important common recent are going important" }, { "startSec": 308.79, "endSec": 341.91, "text": "The most important part is that the first thingAnd then it just starts to break downBecause it's really just looking at you know very simpleI guess you could call this sort of AI, but it's predictive textAnd a lot of what AI really is is predictive text on steroidsBut it's gotten so goodThat we are now starting to almost anthropomorphize this thingAnd it's really easy for us to anthropomorphize things and so what I" }, { "startSec": 342.71, "endSec": 356.23, "text": "What I'd like to kind of start off this conversation this one way conversation isThink about how easy it is to anthropomorphize your dogThat is a human characteristic" }, { "startSec": 357.03, "endSec": 365.43, "text": "Yes, the dogs are intelligent, but not intelligent in the way that humans are intelligentBut we are able to gift them through our own perception" }, { "startSec": 366.95, "endSec": 386.5093406593407, "text": "More than they actually have yes, of course they have personalities, but look look at even that word personalityIt's like similar to a person their personhood in the same way that when you talk to dog owners are goOh, it's not a dog. It's a person. It's a dog. It's a dogAnd just because your dog's smartYou know, that's I love dogs grow up at dogs." }, { "startSec": 386.5093406593407, "endSec": 421.8029574541681, "text": "That's awesome But it's easy for us to dilute ourselves into thinking that they're more than they actually areApply that same kind of logic to like Pixar movies toy story or whateverIt's easy for us toKind of get lured in byAssigning our perception to the object of our perceptionIt's part of the reason why anime fans can get pretty weirdYou know, there's a there's a fine line, right? I love anime." }, { "startSec": 421.8029574541681, "endSec": 427.99, "text": "I love Japan There's like a fine line where it's like hey, they're not real. These are characters" }, { "startSec": 428.95, "endSec": 445.4194214876033, "text": "It's like it's like the people that try to marry their favorite fictional character and have likeForceding or something or the people that have the body pillows and they like worship the body pill, you knowI'm all for havingHarmless fun and fandom. That's all cool." }, { "startSec": 445.4194214876033, "endSec": 462.7850187362045, "text": "But there are interesting fine linesMaybe the big red lines the people cross that is like hey, yo That's you're putting a lot of yourself into this inanimate thing or into this non-human thingYou are personifying it you are anthropomorphizing it." }, { "startSec": 462.7850187362045, "endSec": 494.47, "text": "It's very easy to do thatTo these machines that are very good at predicting The next thing that you want them to say or the next thing that most people would saySo if you think about it in terms of probabilitiesA lot of what they're really doing is just using advanced stats to sayusually when I end a sentenceLike this or a clause like thisI would follow it with a statement like that" }, { "startSec": 496.68, "endSec": 511.64, "text": "It's it's using these grammar rulesThat some of us you know are completely unconscious of doing and that's one of the things that separates a native speaker from a non native speakerIt's also one of the things that separates bad AI from good AI. It's learning some of these subtle" }, { "startSec": 512.76, "endSec": 534.4625857519788, "text": "Kind of linguistic patterns the things that you expect the expectations and the one the moments that you break themThe moments that you break them are the truly like transcendent like wow. I never would have expected thatButThink about the permutations of language and then come back to that conversationSo rather than like immediately jumping to oh my god." }, { "startSec": 534.4625857519788, "endSec": 557.21, "text": "It's a sentient robot It's it there's a person trapped inside of the server which some you know engineers have gotten intoAnd it's kind of led them down some dark paths in terms of psychology where they think like every time they turn off their computerThey're killing their friend we might get thereBut I think I think that a lot of that is coming from the person" }, { "startSec": 557.93, "endSec": 585.978345323741, "text": "In the same way that those anime fans are believing that their anime characters are realAgain, I love animeBut like there's a limit when something gets really good at predictingHuman pattern human speechIt can feel emergentBut take a look under the hoodAnd really what's going on is that we are creating these little tokensThese tokens." }, { "startSec": 585.978345323741, "endSec": 599.6445062692703, "text": "So it's not actually learning speechIt is assigning each of these words a number And then associating those numbers with other numbersYou feed it. Let me look at this." }, { "startSec": 599.6445062692703, "endSec": 623.3396658082514, "text": "This is a crazy thing that I readThere are 176 billion parameters in GPT-3We're on four and five is coming soonYou basically uploaded a good chunk of the internet into this thing Which is a whole other legal question mark right without most of people's consent." }, { "startSec": 623.3396658082514, "endSec": 647.66, "text": "It was publicSo it's mineI can train this robot on your reddit postData is power and If you look at that you go okay yet nobody could program 176 billion parametersAnd nobody's really able to have oversight with that unless they use another AI to look into itOr have this absolutely massive team that nobody's gonna do thatAnd that's the difference between those" }, { "startSec": 647.66, "endSec": 674.42, "text": "Scientists versus the engineers the engineers go it works parameters whateverIt worksBut the way that it works you tokenizeSo the doesn't become the it becomes like let's say the number oneBall becomes the number twoThe ball those sometimes go togetherNot super often, but like say you're in sportsSo that's gonna be moreSo then it starts to create these little families called clusters" }, { "startSec": 675.38, "endSec": 700.52, "text": "These clusters start to form these variousWhat's the best way to put this disambiguations it's a big wordthese varioussubjectsthatDifferent words become more or less associated withSoSay for example, you're a training AI to be a really good sports writer" }, { "startSec": 701.08, "endSec": 712.84, "text": "You would take all of the sports writing that you could possibly find an inherent in that writingWill be a bunch of different patterns. It's going to be for basketballThere's gonna be a bunch of people that are basically using the same sentence" }, { "startSec": 713.8, "endSec": 737.7, "text": "So you can take that sentence and say hey, it really feels like a basketball article when you write something likeblank player was able toPut the team on his back and carryAnd will themselves to the free throw line, you know will themselvesThat's like a very basketball thing" }, { "startSec": 739.22, "endSec": 764.74, "text": "And so now the AI will see that in a bunch of different sources and say assign itOkay, when talking about sports I will use will themselves to theSports term more oftenAnd what that is called is waitingSo you assign these different weightsSo just like you're measuring thingsBut again, we're not talking about actually the languageWe're talking about a disambiguation of the language" }, { "startSec": 765.22, "endSec": 796.18, "text": "A in intermediary we put in our prompts whichWords the computer translates those words into different tokens and then analyzes how often are those tokens associated with other tokensAnd then it uses probabilityAnd a lot of what ends up happening is that you have these scientists or theseEngineers that train it until it goes out to publicSo it self-trains for a good amount you just upload a whole bunch of crap" }, { "startSec": 796.58, "endSec": 809.54, "text": "And then it like I'm saying it says okay will themselves to it identifies thatBut then after that you'll have humans which is happening in a very massive scale on ChatGPTWill like thumbs up or thumbs down the response" }, { "startSec": 810.74, "endSec": 825.14, "text": "And so it that gives it additional weights. Did you do your job? Did you not do your jobAnd over time it ends up creating the appearance of being intelligentAppearance of being emergent and again that might get there" }, { "startSec": 826.02, "endSec": 859.13, "text": "But I think that there's a really big difference between this is another big wordLike a meme but a mimesis like a false falsification aShadow of the thing rather than the thing itselfandone could sayLike this is the way we learn is that we ingest a bunch of stuffWe apply different weights which you would call taste or preferencesAnd then you output new stuff" }, { "startSec": 861.5, "endSec": 884.14, "text": "But there's a difference between outputting new stuff that sounds like a permutation as in a shufflingOr truly new like actual creativityAnd sometimes it's hard to tell the difference especially in someMediumsYou know one of the things I like to talk to guests about are the differences between making art like" }, { "startSec": 884.46, "endSec": 899.99, "text": "Your art your music the stuff that you're going to put as my artist name or work for hireLike somebody comes to you and says hey, I wantI'll spoil somethingThis is something that I think I'm gonna do" }, { "startSec": 900.79, "endSec": 921.99, "text": "I want to do a train beat which is teke-teke-teke-teke-teke-teke-teke-teke-teke-tekeUnlike a snare drumWant to do that but really fast because I want to do a song called ShinkansenReally fat teke-teke-teke-teke-teke-teke-teke-teke-teke-teke-teke-teke-teBullet train in JapaneseAnd I want somebody to take a Coltrane-esque solo on top of that really fast train beat so it's like double trains" }, { "startSec": 923.43, "endSec": 935.6925, "text": "You know that's the prompt right? I'm taking disparate thingsSpeeding one up that's usually used in like you know country or folk musicAdding a more of a jazz thing because the guy's name is trainPutting them together." }, { "startSec": 935.6925, "endSec": 949.03, "text": "That's the prompt now the execution is where it really comes into playBecause a robot would get that and be like okay train beat faster a human would get that and go Okay, how could I apply what I know and the way that my limbs work" }, { "startSec": 950.31, "endSec": 967.27, "text": "On fast beats and then apply what I know about train beats to do that in a way that's totally newI'm not just gonna speed up like I would in a row like a computerI'm going to actually have to biometanically play this beatSo I'm gonna do that and it's gonna be something similar to a train beat" }, { "startSec": 967.27, "endSec": 978.1563215258856, "text": "But not quite it'll be something similar to punk music, but not quite it'll be something similar to drum and bass, but not quiteThat's what I'm looking for and that's really hard." }, { "startSec": 978.1563215258856, "endSec": 990.9135624672323, "text": "It's likeWhen I'm creating music, especially the weird stuff like thatI'm reaching out into mist and trying to grab a little bit of itIt isn't as simple as saying I want x plus y equals z It's like not like not at all." }, { "startSec": 990.9135624672323, "endSec": 1008.87, "text": "I'm trying I'm saying x because I know what the musician is going to interpret that asI'm calling a certain musician because I know they're tendencies And then I'm gonna record it a certain way like it's like all of these decisions start cascading upon themselvesAnd none of them are as concrete as the words that I said" }, { "startSec": 1010.23, "endSec": 1029.2321705426357, "text": "And then I have a feedback a conversation with those musicians in a way that I cannot quite yet with an AIThis would be a fun prompt to try both, you knowAnd maybe I willBut then same thing with the Coltrane thing like I don't want something that is just trained on ColtraneAnd it's like I am Coltrane bot." }, { "startSec": 1029.2321705426357, "endSec": 1038.36, "text": "I'm going to play Coltrane musicYou know it takes a crazy tenor sax solo It's like no, I want somebody that's gonna bring themselves, but like think Coltrane" }, { "startSec": 1039.96, "endSec": 1043.56, "text": "And play something completely new just you know, give me a little flash of Coltrane" }, { "startSec": 1044.6, "endSec": 1064.44, "text": "Versus something that is purely trained on Coltrane is going to sound like a weird shuffling of that deckOf course, you can always stack parametersYou can always get things farther, but again, it's the difference between a shadow and the thing itselfAnd the fun thing that I like about all of this is it's so philosophicalIt's crazy" }, { "startSec": 1064.44, "endSec": 1078.52, "text": "But I like this philosophical stuff especially in creativity because like that's what this isCreativity is philosophical creativity is the highest height that a humanity can doAnd we're trying to create a robot that can kind of do that I guess some people are" }, { "startSec": 1079.4, "endSec": 1086.4933333333333, "text": "But it's a great measuring stick and it's a really interesting thing that's like hey, it tricked meHe got really close." }, { "startSec": 1086.4933333333333, "endSec": 1109.32, "text": "That's awesomeAnd then but the leap between saying and so if we just add more parameters, it'll get thereIt will be truly emergentBut a lot of the scientists are sayingWe might need to rethink the whole thing or at least figure out the way that the human brain works because we don't know To be able to get thereAnd on that point the human brain thing I think is really interesting because" }, { "startSec": 1110.12, "endSec": 1141.35, "text": "These things are called artificial neural networks and again look at the languageThey're using those words on purpose to make you think it's a big brainWhen it's really is inspired by a limitedUnderstanding of the way that the brain worksWe are limited in terms of ourUnderstanding of this physical thingAnd we're going to create a digital shadow of itBased on that limited understanding, but we'll call it neural" }, { "startSec": 1142.79, "endSec": 1149.83, "text": "So that you know you fund our stocksBut it's it's still it's it's a cool idea. So basically the idea is like a neuron" }, { "startSec": 1151.03, "endSec": 1172.23, "text": "These little two electrical signals that go on in your brain this happening right now as I'm spewing language into your brainAnd you are digesting it or ignoring it as you're drivingThat littleNeuron that's firing if you zoom way in just a one of themIt's sending electrical signals through the use of chemical thresholding" }, { "startSec": 1173.19, "endSec": 1178.79, "text": "And the way that we did again like kind of break this down for digital is it goes zero or one" }, { "startSec": 1179.83, "endSec": 1196.65, "text": "Does it meet this threshold by which it fires from one neuron to the next or does it fall below that thresholdAnd that's the basic idea and the same way that you zero one for machine codeYou're doing that, but you're doing it through a network of neurons" }, { "startSec": 1197.45, "endSec": 1220.25, "text": "And you're building connections between different artificial virtual neurons so that you can scale upIt's it's a genius really cool ideaUsing the physical world to inspire a digital world" }, { "startSec": 1286.17, "endSec": 1302.41, "text": "I like the idea of like what's lostIn the simplificationAnd then you build it back up right through mathematical tools" }, { "startSec": 1304.41, "endSec": 1314.2032673267327, "text": "Stochastic gradient descent was a was a term that my my friend that knows this stuff really well taught meBasically like you're you're creatingYou're reducing dimensionalities." }, { "startSec": 1314.2032673267327, "endSec": 1336.0747421565977, "text": "I don't anywayI'm not smart enough to really explain it at a good levelBut I'm like smart enough to vaguely understand what he was sayingAnd he maybe I'll get him on the pot so we can actually describe it in a in a coherent way But basically the idea is thatYou need to lower the amount of dimensions that all the stuff is going downSo you some really complicated math." }, { "startSec": 1336.0747421565977, "endSec": 1358.25, "text": "It's actually sort of simple math To make it more digestibleAnd you take the path of least resistanceDown the dimensionalityAnd then you kind of like go back and see how many errors you were and do it againThis basically is a way to like run through training dataAnd see if something's sticking see if something's training this AIIn a way that you can use" }, { "startSec": 1358.49, "endSec": 1381.69, "text": "So the classic example cats cats versus dogs you upload a whole bunch of cats to this machineAnd then you tell itIdentify whether or not there's a catThis is the early early internet stuffSo like we've had this for a long long time and you can make this if you wanted toI don'tBut you canYou can build these like very simple AI models if you if you have a" }, { "startSec": 1382.57, "endSec": 1385.85, "text": "A bunch of time to burn I guess it doesn't take that long though apparently my friend at it" }, { "startSec": 1387.05, "endSec": 1404.3774074074074, "text": "But basically you upload a whole bunch of training data to this thing tries to create the visionThe platonic ideal if you will of a catAnd over time it kind of gets thereAnd then you can take it the next step further and you can say add a catAnd that's what Photoshop can do now you can do this thing." }, { "startSec": 1404.3774074074074, "endSec": 1416.17, "text": "It's AI generative fillSo you can circle a little thing like say for example I took a picture myself sitting in my studioI want to cat on my shoulder. I can circle the dead space around my shoulder and say add a cat" }, { "startSec": 1417.05, "endSec": 1425.4851592356688, "text": "And a lot of cat because it knows what a cat is because it's been shown a bunch of pictures of a catIt's had a researcher there saying no you got it wrong. Yes, you got it right." }, { "startSec": 1425.4851592356688, "endSec": 1441.1270445859873, "text": "No, you got it wrongOver and over and over againcontrolling those inputs which the inputs are the cats and then controlling the outputs Which is whether or not it's able to draw a cat and you tell yes or no and then it creates the weights itselfIt's kind of a black box." }, { "startSec": 1441.1270445859873, "endSec": 1461.16, "text": "You can open up that black box and look inside butAgain, how many was it?176 billion parameters in GPT-3 But yeah, I think it's interesting to think like you're creating ordered meaning spaces was one of the things that they called itAnd like I like to think of that too as writing to referencesSo like the culture and example, but I think a perfect example" }, { "startSec": 1461.16, "endSec": 1479.56, "text": "And this is one or I think AI might end up eating our lunch is in like made to order commercial music. SoI was asked at one pointJust a cold email to write the temu adFor that ended up being in the Super Bowl" }, { "startSec": 1480.76, "endSec": 1485.88, "text": "Um, it was not a lot of money. It was super tight turnaround. I was busy. I said no" }, { "startSec": 1487.16, "endSec": 1507.88, "text": "And it's like one of the worst sounding bits of music. I don't know if you guys heard it if you didn't check it outBut it is a brutally grading advertisement that would have been absolutely soul crushing to work onEspecially because they like I pretty sure it was like $2,000 for a Super Bowl ad to make the musicI don't know if that's what it ended up going with but it was something very very small" }, { "startSec": 1508.2, "endSec": 1532.8059113300494, "text": "For a very very stressful turnaround time to make some really really bad musicBut that music was likeIt could have been done by a robotMight have done a better jobBecause it was intentionally lifelessIt was intentionally like everything you've ever heard before just kind of like temuYou know like fake shop like a billionaire thing." }, { "startSec": 1532.8059113300494, "endSec": 1554.5277868480725, "text": "It's just it's so hollowIt is like the the final boss of consumerism app So you want the final boss of consumerist music to match it. It's perfectBut that that's a great that's a place where it's like hey youI want the least amount of creativity out of you." }, { "startSec": 1554.5277868480725, "endSec": 1568.31, "text": "I want the shadow not the objectBut don't confuse the two And that's not to say I'm not excited about it or I don't think that there are great use cases for itAnd I think sometimes it is fun to show the shadow before you fill it in" }, { "startSec": 1570.15, "endSec": 1600.9235885167463, "text": "One thing that I thought was really interesting is people have been calling again listen to the languageChoice of words people have been callingAI kind of going off the rails hallucinationBut I listened to one research and they were talking about it how hallucinationActually is a good thing in certain contextsYou know on the like classicMemem from like a crime show zoom in enhance enhance again." }, { "startSec": 1600.9235885167463, "endSec": 1623.75, "text": "Oh, it's the killer That whole kind of thingThey actually have some tools that can kind of do that nowYou can enhanceWhere basically what's happening is just like the words the tokenized thing you can tokenize pixelsYou can tokenize expectations of large chunks of pixels and when you see like oh this usually is matched with that" }, { "startSec": 1623.75, "endSec": 1636.4701985111662, "text": "Which is usually matched with that you can make these disambiguations like putting the cat on my shoulderYou can do the same thing for video framesIt's like oh, I shot this at four frames per second." }, { "startSec": 1636.4701985111662, "endSec": 1660.1578320249841, "text": "Can you upscale it toYou know 30It's gonna do its best might look a little weird, but you can you can do thatBecause it's guessingIt sees the you know, let's say that let's say it's two frames per second It goes one frame and then the next frame is the second later. It can say okayWhat is most likely happening between those two frames?" }, { "startSec": 1660.1578320249841, "endSec": 1665.83, "text": "I'm gonna fill them inIt might look weird And the less information you give it the more guessing it's having to do" }, { "startSec": 1667.35, "endSec": 1682.31, "text": "But to an extent you know, that's kind of what it's being trained to do is hallucinateSee something that's not there based on previousprobabilitiesAnd it's funny because uniqueness" }, { "startSec": 1683.35, "endSec": 1690.23, "text": "Uniqueness is not that unique. It's it's pretty easy to come up with a unique sentence that may be has never been said before" }, { "startSec": 1691.35, "endSec": 1713.27, "text": "As long as you have the words and you can make it sound convincing as long as you understand the patternsthe inflectionsAnd AI obviously I meanIt's it's getting scary in terms of copyingJust like right now think about all of the different nuances of speech patterns that I've been pouring into this now almost 28 minute episode" }, { "startSec": 1715.03, "endSec": 1719.99, "text": "It can copy all of that and it can use that data to call my parents and try and extort them" }, { "startSec": 1720.79, "endSec": 1730.9376056338028, "text": "Do some really bad things or I could die and then you could keep listening to this podcast because somebody else could be writing the scriptIt's pretty weird." }, { "startSec": 1730.9376056338028, "endSec": 1752.415151587491, "text": "That's the realityIs I'm training it right now spotifies listeningIt's it's transcribingAnd I've seen some pretty cool stuff in terms of that that is on not not that scary of side where people are Posting videos and then the webcam is able to make it look like they're speaking fluent JapaneseAnd guess what? It sounds and looks pretty good." }, { "startSec": 1752.415151587491, "endSec": 1763.51, "text": "It evenAdjusts the way that their heads tilt and nod in the appropriate Uh, kind of rhythm and pattern of speechSo you can say one thing completely different and it can" }, { "startSec": 1764.23, "endSec": 1779.99, "text": "Translate not only your speech to another language in a convincing manner, but also your body languageBut again, that's a shadowThere's a different skill set to actually speak Japanese" }, { "startSec": 1781.11, "endSec": 1785.99, "text": "And to me the shadow might look pretty dang real, but I'm not a native Japanese speaker" }, { "startSec": 1789.19, "endSec": 1800.5713623978202, "text": "That's one of the things I think that's going to drive musicians mad if we're going to hear all this stuffIf you're like, no, it's so wrong. Why did they let that happen?But a bunch of dog musicians feel like no, it's it's cool." }, { "startSec": 1800.5713623978202, "endSec": 1825.59, "text": "It sounds greatBut then again like I want to do this stuff that a robot could never doThat's what I want to focus most of my time on doing Um, but like Jon Sandler said last week, you know, you take the jobs you getIt's great to be doing musicIt's not about that, but it is about optimizing yourself to do the thing itself rather than the shadowBecause the shadows what's easily replacedYou are not" }, { "startSec": 1827.86, "endSec": 1854.62, "text": "You can have a hologram perform in your placeBut it will be a lot less compellingAt making your next albumSoWhen you take that to your own creativityExamine the way that you are ingesting referencesCreating your own weights your own tastes" }, { "startSec": 1856.38, "endSec": 1894.63, "text": "But try to challenge challenge and channel your creativity to create something very newThat might make it less commercially successful short-termBut it's something that only you can doSomething that is new not only in the shufflingNot only in the permutations for the probabilitiesBut something that is truly emergentTruly uniqueYou have it in you because you are a being of infinite complexity as a humanInfinite" }, { "startSec": 1894.63, "endSec": 1921.03, "text": "You have every capacity for good every capacity for evil every capacity for excellenceAnd it is your choice as it is mine every single dayTo chart that course or not, you knowThere's nothing that bad about being lazy for a dayThat's cool tooAnd sometimes I think that's where the magic of creativity does kind of come fromIs from the the human pacing of the whole thing" }, { "startSec": 1921.03, "endSec": 1945.59, "text": "You know I get some really cool ideas when I'm out on a run when I'm being physicalWhen I'm like in the hills touching the dirtYou know like touch grassThat that can unlock something that machines are never going to have access toSo go do that be part of thatUse the rhythm of the seasons and of your excitement and your energyThe channel unique creativity" }, { "startSec": 1946.63, "endSec": 1957.137493403694, "text": "That only you have in whatever it is you do if you make music or whateverIt's not it's not really about that. It's about being a creative complete humanYou've got it." }, { "startSec": 1957.137493403694, "endSec": 1981.98, "text": "You've got it in youDon't be scared by these machines because they they're not going to have thatThey might have something that looks like itBut who caresUse it use it where you need it see if it can speed you up I've been using it for my Dungeons and Dragons sessionsSometimes like I'll get into a situation like for example oneThey wanted to go inside of a video gameSo they were going to play Pac-Man" }, { "startSec": 1981.98, "endSec": 2003.74, "text": "And I had to quickly figure out ways that I can you know use dice rollTo determine whether or not they're winning or losing the Pac-Man on Dungeon mapAnd ChatGPT really helped me out to figure that out really quicklyAnd of course I read it and I made some changes in terms of what I knew my players were aboutAnd she made some changes during the game just slightly" }, { "startSec": 2003.74, "endSec": 2033.3, "text": "To tilt the rules and make it a little bit more challenging a little bit more funBut it helped me kind of get the ball rollingAnd there's no shame in thatIt's funUse the toolsBut ultimately cultivate yourselfThat's where it really comes fromHope you guys enjoyed this podcastIf you would please rate me a five starsThat would be much appreciatedJust click on the dot dot dot if you're on the Spotify app" }, { "startSec": 2033.3, "endSec": 2055.46, "text": "And then it will open up rate showClick on that five starsAnd there will be some incredible guests coming upEliott Glinn is going to be next weekReally a great person kind of on this topic of mixing and creatingMusical sound in a just" }, { "startSec": 2056.58, "endSec": 2083.14, "text": "Just different level right like that's what we're all trying to do. That's that's what we want even as music fansAnd that cannot be so easily replaced by the machinesI hope you enjoy these little AI kind of tidbits and turnsI'm trying to follow it every single day just because I find it exciting and challengingIt's like bring it onSo let's do thatIf you'd like to hear my music go to www.scoobertdoobert.pizza" }, { "startSec": 2083.86, "endSec": 2097.78, "text": "And you can hear all of my stuff find my linksAll right, guys, see you next week" } ], "correctionsApplied": [ { "id": "fuzzy-host-name", "count": 1, "note": "Fuzzy host name + glued suffix (rapidfuzz) (100.0%): 'Scoobert Dooberted' -> 'Scoobert Doobert, and'" }, { "id": "scooberdubert-url", "count": 1, "note": "Outro URL ASR typo" }, { "id": "scuba-duber-name", "count": 1, "note": "Host name ASR typo (not scuba diving)" } ] }