WEBVTT - TechStuff Looks at Three AI Startups 0:00:04.440 --> 0:00:12.520 Welcome to Tech Stuff, a production from iHeartRadio. Hey there, 0:00:12.560 --> 0:00:16.120 and welcome to tech Stuff. I'm your host Jonathan Strickland. 0:00:16.160 --> 0:00:19.400 I'm an executive producer with iHeart Podcasts. And how the 0:00:19.480 --> 0:00:22.319 tech are you? Y'all? Were getting up to the time 0:00:22.360 --> 0:00:25.200 of year when I like to look back on, you know, 0:00:25.239 --> 0:00:27.960 the months that have passed and reflect on all that's 0:00:27.960 --> 0:00:30.880 happened in the tech space, and for twenty twenty four, 0:00:30.960 --> 0:00:34.520 it's been a heck of a lot. But I thought 0:00:34.640 --> 0:00:38.239 today I would talk about three startup companies in the 0:00:38.400 --> 0:00:42.519 AI tech space that are really hoping to make it 0:00:42.640 --> 0:00:45.000 big in the coming years. I'm sure I'll talk a 0:00:45.040 --> 0:00:48.280 lot about AI when we do our year in review 0:00:48.640 --> 0:00:52.960 episode because this year has been largely about AI and 0:00:53.080 --> 0:00:56.480 different aspects, a lot of them kind of scary, right, 0:00:56.760 --> 0:01:00.120 But I wanted to talk about some startups because that's 0:01:00.120 --> 0:01:02.920 something we don't focus on that much in the news episodes. 0:01:02.960 --> 0:01:05.680 Most of the talk has been around companies like open 0:01:05.720 --> 0:01:10.959 AI and Microsoft and Meta and Google and Amazon. I 0:01:11.080 --> 0:01:13.800 thought maybe we could take a look at some startups 0:01:13.840 --> 0:01:16.360 in the space because there are lots of those two. 0:01:16.640 --> 0:01:19.840 There are tons of AI startups, some of which are 0:01:20.319 --> 0:01:23.800 doing all right right now, some of which may be struggling. 0:01:23.920 --> 0:01:27.800 And honestly, like, there's this growing concern in the AI 0:01:28.000 --> 0:01:32.840 field that perhaps some versions of AI are starting to 0:01:32.920 --> 0:01:37.200 hit like a wall when it comes to advancements, like 0:01:37.720 --> 0:01:40.520 not that they're not continuing to evolve, but that that 0:01:40.560 --> 0:01:44.800 evolution is happening on a slower time frame than what 0:01:44.880 --> 0:01:48.080 we saw previously. Which typically that's what happens, right, Like, 0:01:48.600 --> 0:01:52.560 usually you have lots of early gains and then it 0:01:52.640 --> 0:01:55.160 starts getting harder and harder. Those of y'all who work 0:01:55.200 --> 0:01:57.920 out and know what I'm talking about. Anyway, I thought 0:01:57.960 --> 0:02:00.800 I would give a shout out to the true team 0:02:00.920 --> 0:02:06.680 tru I see over at startups Savant. That website has 0:02:06.680 --> 0:02:08.880 put together a list that they called the one hundred 0:02:08.919 --> 0:02:13.399 top startups to watch in twenty twenty four. Now, that's 0:02:13.480 --> 0:02:16.000 a huge number of startups, right and I'm not gonna 0:02:16.000 --> 0:02:18.440 go through all of that. I'm going to look at 0:02:18.480 --> 0:02:20.359 just a few of them, and I want to come 0:02:20.400 --> 0:02:24.760 in a bit on what's been going on. So artificial 0:02:24.760 --> 0:02:27.520 intelligence is still something of a boom period right now. 0:02:27.919 --> 0:02:31.520 You know, Yes, there are these these areas where AI 0:02:31.600 --> 0:02:35.040 could potentially be brushing up against some limitations to certain approaches, 0:02:35.080 --> 0:02:38.800 particularly in the large language model space. But AI, as 0:02:38.840 --> 0:02:41.640 I've said many times, is a very complicated topic. There're 0:02:42.000 --> 0:02:45.560 lots of nuance to AI. It's not just one big, 0:02:45.840 --> 0:02:51.880 monolithic discipline. It's lots of much smaller, subtle disciplines that 0:02:52.320 --> 0:02:56.200 collectively make up artificial intelligence. To that end, I want 0:02:56.240 --> 0:02:58.320 to talk about some of the startups mentioned in this 0:02:58.520 --> 0:03:01.280 piece in Startup Savant. One I would like to chat 0:03:01.320 --> 0:03:06.960 about is Suno. Suno because simultaneously puts on a really 0:03:07.160 --> 0:03:11.120 darn impressive display while also being indicative of some of 0:03:11.120 --> 0:03:15.360 the more challenging aspects of generative AI in particular. So 0:03:16.040 --> 0:03:19.840 Suno is based out of Cambridge, Massachusetts. Technically it's a 0:03:19.880 --> 0:03:24.079 twenty twenty three startup, and it released its first build 0:03:24.240 --> 0:03:26.880 of the generative AI tool that they created back in 0:03:26.960 --> 0:03:30.640 December twenty twenty three. But the first stable release of 0:03:30.680 --> 0:03:34.640 that application came out just this past November. So I 0:03:34.680 --> 0:03:37.720 think Suno really qualifies as a twenty twenty four AI 0:03:37.840 --> 0:03:41.240 company if you ask me. I actually downloaded the Suno 0:03:41.320 --> 0:03:43.560 app to give it a try, and I have to 0:03:43.640 --> 0:03:49.120 admit it is pretty impressive. So Suno uses generative AI 0:03:49.200 --> 0:03:53.800 to create music based off user prompts, so those prompts 0:03:53.800 --> 0:03:57.640 can be really specific or really vague. So, for example, 0:03:57.880 --> 0:04:01.720 you might write a prompt like create a high energy 0:04:01.920 --> 0:04:05.800 dance track with lots of synthesizers and drums about going 0:04:05.840 --> 0:04:08.280 out on the town with a group of friends, sung 0:04:08.400 --> 0:04:12.360 by a female vocalist right, Or you might give it 0:04:12.480 --> 0:04:17.640 much more broad direction, like create an Appalachian folk tune 0:04:17.839 --> 0:04:24.440 about witches. Suno can compose music including lyrics and synthesized vocals, 0:04:24.640 --> 0:04:27.359 which on a phone speaker sound pretty darn convincing. Like 0:04:27.400 --> 0:04:29.880 when I made it make folk tunes. You could even 0:04:29.920 --> 0:04:33.560 hear like the breathing sounds. I guess someone will go, 0:04:35.320 --> 0:04:39.000 you know, it sounded organic, at least on a phone speaker. 0:04:39.040 --> 0:04:42.040 I'm sure if I were listening on really high end speakers, 0:04:42.080 --> 0:04:45.279 I could probably detect a bit of the artificiality. But 0:04:45.440 --> 0:04:49.600 to my dumb ears, they sounded pretty good. Now, what 0:04:49.680 --> 0:04:52.080 you do with that music track from that point forward 0:04:52.160 --> 0:04:55.039 depends on whether or not you're a paid subscriber to 0:04:55.120 --> 0:04:58.480 the service. If you are using their free basic plan, 0:04:59.080 --> 0:05:05.040 then you can use those tracks for any non commercial purposes. 0:05:05.560 --> 0:05:10.279 The actual ownership of the music tracks itself, those belong 0:05:10.360 --> 0:05:12.760 to Suno, so you don't have ownership of the tracks 0:05:12.760 --> 0:05:15.479 you make if you are a free user of the 0:05:15.520 --> 0:05:18.600 basic service. But let's say that you were running a 0:05:19.120 --> 0:05:22.480 role playing game session, like you're the game master and 0:05:22.520 --> 0:05:25.760 you really need some ruddy mysterious music playing in the 0:05:25.800 --> 0:05:28.920 background as your adventurers walk through a dungeon. Well, you 0:05:28.960 --> 0:05:32.320 could use Suno to generate that music for you if 0:05:32.320 --> 0:05:35.680 you liked like it could be instrumental pieces orchestral pieces, 0:05:35.839 --> 0:05:38.479 you know, synth wave, whatever it may need to be 0:05:39.120 --> 0:05:41.920 to suit the needs of your game. You could do that. 0:05:42.240 --> 0:05:44.560 And because it's non commercial, you know, you're just playing 0:05:44.560 --> 0:05:47.360 with friends, there's no problem there. Now, if you wanted 0:05:47.360 --> 0:05:50.920 to make commercial use of the music generated based off 0:05:50.920 --> 0:05:53.599 your prompts, then you would need to be a subscriber 0:05:53.680 --> 0:05:58.400 of the Pro or Premier level, and then you own 0:05:59.040 --> 0:06:03.640 whatever tracks are generated. You own the intellectual property of 0:06:03.720 --> 0:06:08.160 those songs, and Suno grants such users a commercial license. 0:06:08.440 --> 0:06:12.040 Copyright gets really complicated because, at least here in the 0:06:12.120 --> 0:06:18.400 United States, you cannot copyright a work generated by a 0:06:18.560 --> 0:06:23.160 I And what you can do is you can monetize 0:06:23.600 --> 0:06:27.320 the track, right like if you wanted to make a 0:06:27.360 --> 0:06:30.080 commercial podcast, Like you want to monetize the podcast, and 0:06:30.120 --> 0:06:33.440 you wanted to use Suno to generate the theme song 0:06:33.720 --> 0:06:36.000 for your podcast, you could do that at the pro 0:06:36.120 --> 0:06:38.760 or premier level, right, because that's a commercial use. You'd 0:06:38.800 --> 0:06:41.360 be using it for a commercial podcast. You could then 0:06:41.440 --> 0:06:44.279 do that. This does start to raise some questions, however, 0:06:44.400 --> 0:06:47.359 I mean, you could just start churning out songs, you 0:06:47.400 --> 0:06:50.120 could switch out prompts, you could tweak approaches in an 0:06:50.120 --> 0:06:52.640 effort to make something, you know, anything that resembles a 0:06:52.640 --> 0:06:56.000 catchy hit, and you don't have to hire musicians or 0:06:56.040 --> 0:06:59.280 singers or anything. You're just working with a computer and 0:06:59.320 --> 0:07:01.200 you're you know, if you have lots of time in 0:07:01.200 --> 0:07:04.080 your hands and you have one of those subscription levels 0:07:04.160 --> 0:07:06.840 where you don't have a lot of limitations on how 0:07:06.880 --> 0:07:09.960 many times you can use the app, well you can 0:07:10.040 --> 0:07:12.240 just keep rolling the dice and maybe you get lucky 0:07:12.400 --> 0:07:14.800 and you have a track that ends up getting crazy 0:07:14.840 --> 0:07:19.120 amounts of attention, and you're just flooding streaming services with 0:07:19.240 --> 0:07:23.320 track after track after track of AI generated music. This 0:07:23.440 --> 0:07:28.520 is something that is happening, and it's somewhat concerning because meanwhile, 0:07:28.560 --> 0:07:31.320 you have actual human being musicians who are trying to 0:07:31.320 --> 0:07:34.800 get noticed, and if there's just a glut of AI 0:07:35.000 --> 0:07:39.760 generated material hitting the various streaming platforms, it gets harder 0:07:39.760 --> 0:07:42.520 and harder to get discovered as an artist. And that 0:07:42.720 --> 0:07:46.360 doesn't seem terribly fair, right. But to the person who's 0:07:46.400 --> 0:07:49.240 just using this tool to make track after track, it's 0:07:49.240 --> 0:07:51.880 a licensed print money baby. Of course, it's not as 0:07:51.880 --> 0:07:54.240 simple as that, and there's definitely some discomfort in the 0:07:54.280 --> 0:07:57.720 music industry over the rise of these AI tools. The 0:07:57.840 --> 0:08:02.040 stuff made by Suno sounds retive of the various genres, 0:08:02.080 --> 0:08:04.080 at least to my ears. In other words, if you 0:08:04.200 --> 0:08:06.360 give it the direction to make a song in a 0:08:06.400 --> 0:08:11.040 specific genre or subgenre, what you get sounds like it. 0:08:11.040 --> 0:08:14.680 It fits pretty much. I didn't try things like rockabilly 0:08:14.800 --> 0:08:17.559 I should. I should try and do a rockabilly song 0:08:18.000 --> 0:08:21.120 about something and just see what it sounds like. But 0:08:21.600 --> 0:08:26.320 you know, for very broad categories like classical or folk 0:08:26.560 --> 0:08:30.080 or R and B or funk, the stuff I was 0:08:30.120 --> 0:08:35.079 getting sounded fairly representative of those genres. Not necessarily brilliant, 0:08:35.360 --> 0:08:37.280 but you know, it's not like something that you would 0:08:37.280 --> 0:08:39.720 hear if you were tuned into a radio station that 0:08:39.760 --> 0:08:43.120 catered to that specific genre of music. One thing to 0:08:43.200 --> 0:08:47.280 keep in mind is again, anything generated by AI here 0:08:47.320 --> 0:08:51.360 in the United States is ineligible for copyright protection, because 0:08:51.360 --> 0:08:54.439 in order to qualify for a copyright, a work has 0:08:54.520 --> 0:08:56.640 to be created by a human being, and writing a 0:08:56.640 --> 0:08:59.439 text prompt into a field and then having an AI 0:08:59.559 --> 0:09:04.240 model generate music does not qualify as a work created 0:09:04.280 --> 0:09:07.120 by a human being. That means that if you did 0:09:07.559 --> 0:09:10.679 create this musical track, and if you did have a 0:09:10.720 --> 0:09:14.319 commercial license to make money from it, there's no copyright 0:09:14.360 --> 0:09:18.040 protection that would allow you to go after anyone who 0:09:18.160 --> 0:09:22.040 infringed upon your intellectual property and made copies of your music, 0:09:22.400 --> 0:09:25.960 whether in whole or in part, or sampled it or whatever. 0:09:26.360 --> 0:09:28.600 And I say your music in the sense of you 0:09:28.679 --> 0:09:32.119 own it, not you created it. And that's a real issue. 0:09:32.280 --> 0:09:35.000 Beyond that, there's a bigger concern in the music industry 0:09:35.040 --> 0:09:38.720 over how SUNO trained its models in the first place. 0:09:39.080 --> 0:09:42.400 Like AI doesn't just magically know that a folk song 0:09:42.480 --> 0:09:45.320 sounds like a folk song, or that you know, rhythm 0:09:45.320 --> 0:09:49.200 and blues sounds a specific way, or Chicago blues having 0:09:49.280 --> 0:09:52.880 a different style to New Orleans blues. The models had 0:09:52.920 --> 0:09:57.479 to learn the rules of music theory and the styles 0:09:57.559 --> 0:09:59.959 and the things that go along with the various style. 0:10:00.559 --> 0:10:02.920 It had to have some form of understanding, which is 0:10:02.920 --> 0:10:04.680 a tough word because I don't mean to imply that 0:10:04.760 --> 0:10:09.119 the model has general intelligence. It doesn't understand like humans understand, 0:10:09.280 --> 0:10:13.880 but that it recognizes these qualities that are associated with 0:10:13.960 --> 0:10:18.920 different kinds of music, right like, it has to have 0:10:19.040 --> 0:10:22.360 that reference bace, and it has to be able to 0:10:22.400 --> 0:10:25.440 generate music that we find nice to listen to. Otherwise 0:10:25.440 --> 0:10:28.920 it's just spitting out random noise. So to get to 0:10:28.960 --> 0:10:32.679 that point, Suno presumably trained its models on a data 0:10:32.720 --> 0:10:35.480 set that included lots and lots of songs made by 0:10:36.120 --> 0:10:39.440 real life, living, and breathing human beings, and that also 0:10:39.520 --> 0:10:43.720 raises concerns about the potential for plagiarism. Now, Suno asserts 0:10:43.800 --> 0:10:46.720 that its model has guardrails in place to prevent it 0:10:46.760 --> 0:10:51.000 from just say, lifting chord progressions and melody lines and 0:10:51.040 --> 0:10:54.280 such from existing songs. But the music industry isn't so 0:10:54.360 --> 0:10:57.320 quick to accept that explanation. I mean, we just kind 0:10:57.320 --> 0:11:01.240 of had a pop culture moment in the form of Heredic. 0:11:01.440 --> 0:11:04.040 If you saw the movie Heredic had Hugh Grant in it. 0:11:04.400 --> 0:11:06.959 That brought up the issue of plagiarism a couple of 0:11:06.960 --> 0:11:09.840 different ways, and one of the ways was he plays 0:11:09.960 --> 0:11:12.680 a song by the Hollies called The Air that I Breathe, 0:11:12.960 --> 0:11:16.120 and then he plays Radioheads Creep and shows like, here's 0:11:16.160 --> 0:11:18.520 two songs that have the same chord progression. It's so 0:11:18.640 --> 0:11:23.960 similar that the Hollies sued Radiohead and were ultimately allowed 0:11:24.000 --> 0:11:28.800 to have some writing credit and royalties from Creep because 0:11:28.840 --> 0:11:32.400 of the similarity between the two songs. Hugh Grant's character 0:11:32.440 --> 0:11:37.080 also points out that Lana del Reys get free shares 0:11:37.440 --> 0:11:40.480 that same chord progression and most of the melodic line 0:11:40.559 --> 0:11:44.280 from Creep, and then in fact, Radiohead or Radiohead's agents 0:11:44.320 --> 0:11:47.680 or whatever sued Lana del Rey for plagiarism as well, 0:11:47.800 --> 0:11:51.520 which is wild because Radiohead had already been sued and 0:11:51.640 --> 0:11:56.320 essentially admitted to or at least acknowledged the similarities with 0:11:56.520 --> 0:11:59.400 the Hollies, and then the Radiohead then goes and sues 0:11:59.440 --> 0:12:02.160 Lana del Y. This is a really sticky subject in 0:12:02.240 --> 0:12:06.720 music already, like this issue of how similar can one 0:12:06.760 --> 0:12:09.520 song be to another? Before you start to say, hey, 0:12:09.520 --> 0:12:12.400 wait a minute, I think you actually copied this other 0:12:12.600 --> 0:12:15.880 piece of music, as opposed to you both independently arrived 0:12:15.880 --> 0:12:20.080 at the same structure from different pathways. But imagine how 0:12:20.120 --> 0:12:22.240 much more complicated it gets if a company were to 0:12:22.320 --> 0:12:26.000 argue that an AI model was plagiarizing protected works by 0:12:26.040 --> 0:12:28.360 generating songs that conform to the rules of music in 0:12:28.360 --> 0:12:31.720 a way that's already similar to existing pieces. In June 0:12:31.720 --> 0:12:34.760 of this year, the Recording Industry Association of America the 0:12:34.880 --> 0:12:38.480 RI double A, filed a lawsuit against Suno. The RI 0:12:38.640 --> 0:12:42.640 double A claims that Suno has engaged in copyright infringement 0:12:42.960 --> 0:12:46.679 by training its models on protected works. The lawsuit seeks 0:12:46.760 --> 0:12:48.920 damages in the form of up to one hundred and 0:12:48.920 --> 0:12:54.559 fifty thousand dollars per copyrighted work that was used in training. Now, 0:12:54.600 --> 0:12:56.760 I'm going to guess that would be a huge amount 0:12:56.800 --> 0:13:00.200 of money, but I don't know for sure because we 0:13:00.240 --> 0:13:03.640 can't actually see the data set that Suno used. Notably, 0:13:03.960 --> 0:13:07.440 it is behind closed doors, so we aren't allowed to 0:13:07.480 --> 0:13:10.120 see how many songs or what kinds of songs are 0:13:10.120 --> 0:13:13.880 from what catalogs Suno dipped into in order to train 0:13:13.960 --> 0:13:16.120 up its AI. I mean it had to be a lot. 0:13:16.360 --> 0:13:19.560 When you're training AI, you need lots and lots and 0:13:19.600 --> 0:13:22.760 lots and lots of training material to get your model 0:13:23.080 --> 0:13:26.760 to start to hone in on what it is you 0:13:26.800 --> 0:13:29.160 want it to do. In the case of image recognition, 0:13:29.440 --> 0:13:33.280 that could be millions of images in order to train 0:13:33.400 --> 0:13:38.760 a model to recognize one thing versus another. So presumably 0:13:39.360 --> 0:13:43.760 the training material for Suno was pretty darn extensive. I 0:13:43.800 --> 0:13:46.960 think it's pretty safe to say that every single record 0:13:47.040 --> 0:13:51.360 label out there likely has some music from their catalog 0:13:51.440 --> 0:13:53.480 that was used in the training. I don't know that 0:13:53.600 --> 0:13:57.160 for sure, that's just a guess, because again, you know, 0:13:57.280 --> 0:14:00.680 it's not like you can have the AI scan the 0:14:00.800 --> 0:14:03.200 radio for like a second and then say oh yeah, 0:14:03.280 --> 0:14:05.320 I got this, and then they could just go and 0:14:05.400 --> 0:14:07.560 generate all that. Okay, well, we've got more to say 0:14:07.559 --> 0:14:11.440 about Suno plus some other AI startups Before I get 0:14:11.480 --> 0:14:22.960 into any of that, though, let's take a quick break. Okay, 0:14:23.000 --> 0:14:26.160 we're back. We're still talking about Suno here, the AI 0:14:26.280 --> 0:14:29.400 company that makes a tool that allows you to create 0:14:29.440 --> 0:14:32.280 a song just based off a text prompt. So I 0:14:32.320 --> 0:14:35.880 think Suno really illustrates the pros and cons of generative 0:14:35.920 --> 0:14:39.320 AI in a neat little package. So the pros are 0:14:39.800 --> 0:14:43.000 if you are not musically inclined, but you need to 0:14:43.080 --> 0:14:47.320 create some original music for whatever reason. Suno is incredibly 0:14:47.360 --> 0:14:50.600 easy to use, and you can even continue to work 0:14:50.640 --> 0:14:54.000 on a piece. Like let's say you prompt Suno to 0:14:54.040 --> 0:14:56.360 make something, you listen to it, you're like, well, that's close, 0:14:56.480 --> 0:14:58.240 but it's not really what I want. You can keep 0:14:58.280 --> 0:15:02.080 on refining it by continue to type in and get 0:15:02.120 --> 0:15:05.880 that shaped into a place where you really like it. Now, 0:15:05.880 --> 0:15:09.080 if you don't know any musicians or you can't afford 0:15:09.160 --> 0:15:12.480 to hire musicians, arguably this is a tool that could 0:15:12.520 --> 0:15:15.480 work for you. But personally, I feel really icky about 0:15:15.520 --> 0:15:19.080 the idea of leaning too hard on Ai to create 0:15:19.360 --> 0:15:23.000 art of any kind, or even calling AI generated material 0:15:23.120 --> 0:15:25.960 art in the first place. That doesn't seem right to me. 0:15:26.200 --> 0:15:29.280 I like to think that art has to have some 0:15:29.320 --> 0:15:33.320 sort of human intent behind the piece. There needs to 0:15:33.320 --> 0:15:36.120 be a human motivation beyond just a text prompt for 0:15:36.160 --> 0:15:39.280 it to be art, at least in my opinion. But 0:15:39.440 --> 0:15:41.960 art is a very subjective thing. And also I'm very 0:15:42.000 --> 0:15:45.360 old fashioned, so it's entirely possible that I'm out of 0:15:45.440 --> 0:15:48.320 step here, but it doesn't feel that way to me. Now. 0:15:48.360 --> 0:15:52.240 The cons are that Tsuno could really impact human musicians 0:15:52.240 --> 0:15:55.080 who otherwise could be hired to devote their skill and 0:15:55.120 --> 0:15:58.360 craft toward creating new stuff, thus making it harder for 0:15:58.440 --> 0:16:00.880 people who have honed their art to make a living 0:16:01.000 --> 0:16:04.640 off of that art. Like think of the countless hours 0:16:05.040 --> 0:16:10.000 that songwriters and musicians and vocalists spend to get good 0:16:10.000 --> 0:16:13.239 at what they do. You know, people aren't just naturally 0:16:14.040 --> 0:16:18.560 flawless in their execution, right. They take years of practice 0:16:18.800 --> 0:16:21.800 to get to where they are. And they do this 0:16:22.320 --> 0:16:25.760 not just from a love of the art, although I'm 0:16:25.800 --> 0:16:27.680 sure that's a big part of it, but also with 0:16:27.720 --> 0:16:30.440 the desire to make a living off of all that 0:16:30.560 --> 0:16:35.160 hard work. Well, tools like SUO arguably create a shortcut 0:16:35.240 --> 0:16:38.880 that might be very tempting for some folks out there 0:16:39.160 --> 0:16:43.560 to just bypass all that messy human stuff and the 0:16:43.640 --> 0:16:46.600 costs that come with it and go for you know, 0:16:46.840 --> 0:16:50.200 let's just go with AI if it's popular. Who cares 0:16:50.680 --> 0:16:54.080 if humans didn't make it, because, you know, to the 0:16:54.120 --> 0:16:57.400 producer side, anyway, the goal is to make money off 0:16:57.440 --> 0:17:00.240 of music. You don't really care where the music came 0:17:00.280 --> 0:17:02.880 from or how it was generated. You just care that 0:17:02.920 --> 0:17:05.879 it makes money. You don't even care if it resonates 0:17:05.880 --> 0:17:08.760 with your audience. You just need it to sell well. 0:17:09.119 --> 0:17:12.680 So that makes me feel icky. The balance between commerce 0:17:12.720 --> 0:17:15.040 and art is always a tricky thing. I had a 0:17:15.080 --> 0:17:18.240 really good conversation with an artistic director of a theater 0:17:18.480 --> 0:17:22.879 about this and how that's always a difficult balance, Like 0:17:22.920 --> 0:17:26.520 how do you balance between commerce and art? And it's 0:17:26.560 --> 0:17:29.040 tricky because you know, we don't live in a world 0:17:29.160 --> 0:17:31.919 where we can just create art and not have to 0:17:31.920 --> 0:17:35.000 worry about paying the bills. We have to do both. 0:17:35.560 --> 0:17:38.640 Then there's also the plagiarism issue. So at what point 0:17:38.720 --> 0:17:41.520 do you say an AI tool is creating music the 0:17:41.600 --> 0:17:44.919 way a person might create music. So for example, people 0:17:45.040 --> 0:17:47.960 lean on inspiration from previous works all the time, right, 0:17:48.280 --> 0:17:51.800 Like you might hear something in a musical piece and think, oh, 0:17:51.880 --> 0:17:53.879 that's interesting and you want to build off of it. 0:17:53.920 --> 0:17:57.000 I mean, sampling is built off this. Their entire genres 0:17:57.000 --> 0:18:00.919 of music that are rooted in this idea where you 0:18:01.000 --> 0:18:04.359 take something that was used in one piece and you 0:18:04.480 --> 0:18:09.280 repurpose it to create something brand new and transformative. Doing 0:18:09.320 --> 0:18:13.040 that is okay, right, even copyright losses, that's okay if 0:18:13.040 --> 0:18:15.960 it's transformative to a certain extent. Like you can get 0:18:16.000 --> 0:18:19.200 into trouble if you don't do it correctly, but you 0:18:20.080 --> 0:18:21.800 kind I would get to a point where you say, 0:18:22.240 --> 0:18:27.440 is the AI, you know, taking inspiration from this previous 0:18:27.440 --> 0:18:30.760 corpus of music or is it actually just copying something 0:18:31.000 --> 0:18:34.000 that's already been done because the AI is determined this 0:18:34.119 --> 0:18:36.480 is the best way to do it. That's tricky. Now 0:18:36.520 --> 0:18:39.240 we'll have to see where companies like Suno go in 0:18:39.280 --> 0:18:43.200 the future. I know that media companies are simultaneously curious 0:18:43.320 --> 0:18:47.120 and worried about this technology. If the media companies can 0:18:47.119 --> 0:18:49.720 make a buck off the tech, then it will come 0:18:49.760 --> 0:18:52.040 as a surprise to no one when we're flooded by 0:18:52.080 --> 0:18:55.840 AI generated tunes. But if those media companies determine that 0:18:55.880 --> 0:19:00.399 working with AI puts their relationships with human artists at risk. 0:19:00.560 --> 0:19:03.080 You know, these are artists who could potentially be earning 0:19:03.160 --> 0:19:06.320 media companies billions of dollars every year, then it's a 0:19:06.320 --> 0:19:08.840 little different, right, Like it would be a dumb move 0:19:08.920 --> 0:19:12.680 to experiment with AI if in the process you're alienating 0:19:12.960 --> 0:19:16.480 the superstars you regularly work with. I say this because 0:19:16.960 --> 0:19:20.320 I've listened in on meetings where there have been discussions 0:19:20.400 --> 0:19:24.000 about using AI to create music, and these were with 0:19:24.119 --> 0:19:28.399 people who worked with companies that had tight relationships with 0:19:28.560 --> 0:19:31.879 musicians and artists, and I had to ask, like, what 0:19:32.000 --> 0:19:34.160 do you think that's going to do with your relationships, 0:19:34.240 --> 0:19:38.199 your professional relationships with these other people who clearly have 0:19:38.280 --> 0:19:43.240 a vested interest in not having AI flood the market. 0:19:43.400 --> 0:19:46.000 And that really gave pause to the meeting. I'm known 0:19:46.000 --> 0:19:49.960 as Debbie Downer at those types of meetings because instead of, 0:19:50.240 --> 0:19:53.000 you know, kind of blue skying the whole AI thing, 0:19:53.160 --> 0:19:57.680 I say, let's take this into context. Let's really think 0:19:57.720 --> 0:20:01.800 critically about this, because otherwise we're putting lots of people's 0:20:01.800 --> 0:20:05.320 careers at risk. And not only that, but potentially we 0:20:05.359 --> 0:20:09.760 are suppressing artistic expression because again, I don't think you 0:20:09.800 --> 0:20:12.760 can call it artistic expression if it's AI generated. It 0:20:12.840 --> 0:20:16.959 might be conforming to certain conventions and rules in an 0:20:17.000 --> 0:20:20.680 effort to create music that sounds like, you know, whatever 0:20:20.760 --> 0:20:23.080 the prompt was, But that's not the same thing as 0:20:23.200 --> 0:20:27.200 artistic expression. All right, let's switch to a different startup. 0:20:27.560 --> 0:20:29.920 This is one that is also a couple of years old, 0:20:30.200 --> 0:20:34.080 but it completed its Series A funding just this year, 0:20:34.320 --> 0:20:37.600 and that's web Ai. Now, I guess I should kind 0:20:37.600 --> 0:20:41.439 of cover what Series A funding actually means. So in 0:20:41.520 --> 0:20:44.080 the world of startups, and this is not just in tech, 0:20:44.400 --> 0:20:47.120 this is startups in general. But in the world of startups, 0:20:47.119 --> 0:20:50.680 there are typically multiple rounds of investment funding that are 0:20:50.720 --> 0:20:53.359 necessary for a business to get to a point where 0:20:53.359 --> 0:20:56.400 it can operate like a business. So first up, you've 0:20:56.400 --> 0:20:59.359 got your seed funding, and this is used to get 0:20:59.359 --> 0:21:02.480 a company established in the very early stages. You might 0:21:02.560 --> 0:21:06.560 use seed funding to do things like, you know, incorporate, 0:21:06.640 --> 0:21:10.400 to design a logo, to secure some office early office space, 0:21:10.400 --> 0:21:14.480 which might be like a sharing situation. Early on, there 0:21:14.480 --> 0:21:17.439 are a lot of businesses that started off as taking 0:21:17.520 --> 0:21:21.320 up a corner of an existing business's office, for example. 0:21:21.720 --> 0:21:24.600 And I think of seed money kind of like how 0:21:24.720 --> 0:21:28.280 buskers will put a few dollars in their hat before 0:21:28.320 --> 0:21:31.199 performing on the street. Right, you see that street musician, 0:21:31.200 --> 0:21:32.919 they get their hat out, it's got some money in 0:21:32.960 --> 0:21:35.480 the hat. Often these musicians will put a couple of 0:21:35.480 --> 0:21:38.359 bucks in there to start with, and that money is 0:21:38.600 --> 0:21:40.800 seed money, and you're hoping it's going to grow and 0:21:40.880 --> 0:21:44.760 blossom as more people throw dollars into the hat. Seed 0:21:44.760 --> 0:21:48.160 funding is kind of similar. Seed money typically comes from 0:21:48.280 --> 0:21:52.240 folks like angel investors, so these could be really influential 0:21:52.320 --> 0:21:55.920 people in the sector, maybe people that the startup owners 0:21:56.280 --> 0:21:59.240 already know personally, could be someone they went to college with, 0:21:59.680 --> 0:22:03.400 or advisor or a relative. It could be something like that. 0:22:03.720 --> 0:22:07.840 It could also be from an incubator group. Incubators exist 0:22:07.880 --> 0:22:12.160 in order to foster ideas that could potentially grow into 0:22:12.680 --> 0:22:16.600 viable businesses. The incubator gets a stake in whatever the 0:22:16.600 --> 0:22:19.879 company is and thus profits if the company does well, 0:22:20.160 --> 0:22:23.840 and in return, the startup gets a little bit of 0:22:23.880 --> 0:22:28.000 stability and access to some assets. But seed money only 0:22:28.040 --> 0:22:30.359 goes so far, and typically it's enough to keep a 0:22:30.400 --> 0:22:33.240 company afloat for just a relatively short amount of time. 0:22:33.640 --> 0:22:37.720 What typically comes after that is Series A funding, and 0:22:37.760 --> 0:22:41.080 in this series the startup opens itself up to investments 0:22:41.119 --> 0:22:45.560 beyond that initial group of seed money investors. This could 0:22:45.600 --> 0:22:48.439 then be followed by additional rounds of investment. You can 0:22:48.520 --> 0:22:52.480 have Series B, Series C, et cetera. And it does 0:22:52.560 --> 0:22:56.639 mean that the investment crowd gets more dense as you 0:22:56.720 --> 0:22:59.359 go on because you get more investors, which means you 0:22:59.359 --> 0:23:01.840 have to pay out more shares. In the long run, 0:23:02.400 --> 0:23:05.439 the goal is to either become a scaled operation that 0:23:05.480 --> 0:23:09.000 has sustainable growth, potentially going public at some point and 0:23:09.040 --> 0:23:12.440 everyone makes their money back plus interest, or you get 0:23:12.480 --> 0:23:15.360 swallowed up by some bigger fish for a handsome payout 0:23:15.359 --> 0:23:20.040 and everyone goes home rich. So webai concluded its Series 0:23:20.160 --> 0:23:23.760 A funding this year and saw about sixty million dollars 0:23:24.160 --> 0:23:29.000 flood into the company. Coffers analysts value Webai in the 0:23:29.160 --> 0:23:33.040 seven hundred million dollar range, So yeah, they got sixty 0:23:33.080 --> 0:23:36.080 million in investment, but they're valued at around seven hundred 0:23:36.119 --> 0:23:39.000 million dollars, So they're closing in on that Unicorn status 0:23:39.000 --> 0:23:42.160