The TechCrunch Podcast

AI art? More like AI fart and other TC news

Episode Summary

This week Darrell talks with Taylor Hatmaker on to talk about Lensa AI and the possible ramifications for artists. Then we’ll hear from some attendees at the TC sessions: Space event  And as always, Darrell breaks down the biggest stories in tech.

Episode Notes

This week Darrell talks with Taylor Hatmaker on to talk about Lensa AI and the possible ramifications for artists. Then we’ll hear from some attendees at the TC sessions: Space event  And as always, Darrell breaks down the biggest stories in tech.

Articles from the episode:

Other news from the week:

Episode Transcription

Darrell Etherington  0:03  

Hi, I'm TechCrunch Managing Editor Darrell Etherington. This is the TechCrunch podcast where we cover everything you need to know about the week's top stories and tech from the people who wrote them. This week we have Taylor hatmaker on to talk about lenses AI and then we'll hear from some attendees at the TC session space event. But first, I'll break down the biggest stories in tech. New rules that Facebook owner Mehta forbid employees from talking about sensitive subjects like abortion, gun control, pending legislation and vaccine efficacy at work. An internal memo circulated of the company says it's intended to help employees focus on delivering their goals and a spokesperson said it's about quote, reducing distractions, cutting off expression of values in the workplace, definitely bucks the trend of the last few years, particularly in Silicon Valley, which is generally swung towards bringing your whole self to work similar attempts at internal policies at Basecamp and Coinbase. face criticism earlier this year. More about this for Amanda silverlink en TechCrunch. In a deeply weird move, Amazon has introduced a means to tip your delivery driver $5 during the holiday season, and it was on his picking up the cost up to a maximum of the first 1 million tips but the weird part is that you have to say Alexa thank my driver to trigger the action. This of course requires that you have an Alexa enabled device in the house somewhere. Despite the apparent feel good vibes this was meant to deliver it's generated a lot of scrutiny regarding Amazon's pay scale and policies around worker organization. Check out more on TC from Amanda silverlink. Salesforce executive leadership drama extends into another week with news that slack CEO Stewart Butterfield will be leaving the company in January, Butterfield's departure comes hot on the heels of the shocking news that Salesforce co CEO Brett Taylor will be leaving which itself was followed immediately by word that Salesforce own Tableau was losing its CEO Mark Nelson. Of course, all these top executives and heads of acquired companies are already rich and have no need to continue working. But it's still bizarre to see so many key people leave all at essentially the same time. Read more about this on TechCrunch from Ron Miller. The US Federal Trade Commission is suing to block Microsoft's proposed acquisition of Activision Blizzard. Microsoft said it intended to buy the gaming giant nearly a year ago in January 2021. Specifically, the FTC takes issue with the impact that deal would have on potential competitors to Microsoft's Xbox and cloud gaming businesses. Predictably, Microsoft is counter arguing that somehow buying up one of gaming's largest developers is good for competition rather than bad. I read all about it from Taylor hatmaker on TechCrunch. First up, I talk with Taylor hatmaker. About lenses AI and the possible ramifications for creators. Hey, Taylor, Hello, how's it going?

 

Unknown Speaker  2:33  

All right, how are you?

 

Darrell Etherington  2:34  

Good. I'm just living non AI generated life thus far. Or maybe I'm not I don't know. Maybe I'm the product of AI generation. That's like a new Do you think? It's the life is a simulation theory, extended life is actually an AI generation. Right? I think I'm the first one to say that. So give me all the money when you make movies out of that or whatever listener,

 

Unknown Speaker  2:56  

this feels like a really generative prompt that you would even respond to. So I buy it.

 

Darrell Etherington  3:01  

Yeah. And Taylor can have half the money too. I've just decided right now. So that's so nice. But we're here to talk about one expression of that we're here to talk about Lanza AI, and specifically some of the some of the complications, some of the blowback because some of perhaps the positive complications to or the positive outcomes that have resulted since this has been out, it's been, what about a week now or something or maybe a bit more than that? Well, just this feature,

 

Unknown Speaker  3:27  

but yeah, they push the feature, I think, at this point, probably like a week and a half ago. And I would say like mid last week is when this like really started taking off and like charting in the app store.

 

Darrell Etherington  3:36  

And so the feature we're speaking of specifically as a AI image generation that uses as its inputs, well is the dataset. And that's what we'll get into is called from kind of just the web, and includes many, many images, probably 1000s, if not millions of images. And then what you do as a user of this app is you input your own so profile pictures or just photos, selfies, photos of yourself, I think asked for 10 between 10 and 20. And then it generates depending on how much you're willing to pay a number of AI generated profile image outputs in the style of art. I don't know. They look painterly or they look illustrated, I guess is the simple way to say it. And then they have a lot of like, sci fi kind of vibes and other things going on which I don't know how they've selected those, but they definitely appeal to me, and I think many others. But what is the issue Taylor that you addressed specifically in the article you

 

Unknown Speaker  4:33  

wrote about this week? Here a lot of issues this is like an interesting one for a lot of reasons I wrote about it because I thought this is the first time like your average person who isn't like chained to Twitter like you and I are like, you know, typing into, you know, a generative AI chat bot all week has actually run into a tool like this probably that's my guess. And it's definitely the first time that most of the people have paid basically a robot to make art. Yeah, you know, I don't think people people were like, like, you're seeing images, you know, they're they're designed to generate certain styles. And that's one of the whole big controversies too, with these models in general. So you're like, oh, wow, shit, like, my friend has this awesome when the avatar looks like you're in Star Wars, you know, they're really good. Yeah, that's the thing. Like, they're just very good. And like you're saying earlier, like, you know, these models are actually trained on billions of images, like this one alone is like 2.3 billion English language caption images, like the dataset is drawing on. So they're very sophisticated. It's drawing from a lot, a lot of visual imagery. So they're used to like make you a really badass looking avatar, but then everyone did it and was like, hell yeah, I look awesome. You know, there's like a, it's like gender too, which is kind of funny. There's like, a girl one and a boy one. And like, you don't identify as either of those ones. So like, if you do the girl and you get like these beautiful floral images, you do the boy and you look like an action hero. They're very good. And then everyone did it and was like, Oh, shit, like, my friends or I have a bunch of friends who were like, artists or illustrators, and I've never paid them for this. Right? So a lot of people had to kind of reflect on the ethics of all of this in this big moment that happened kind of simultaneously online. Mostly on Instagram, but also on Tik Tok.

 

Darrell Etherington  6:04  

Yeah, and that's like, first level, the first level is like, Oh, wow, like, I know, people who do this for a living, and I had never even considered throwing them. Uh, I don't know who's on the face of your $10. Bill, I just realized I was going to try to be cool. Do I? Maybe, because like, you don't say like a Benjamin or something. I don't know who it would be. But anyways, you're not throwing them 10 bucks for anything, right? And you're more than willing to go like, okay, cool, like in app purchase, done, give me these images, right, I really want to see them. But that what ads and what you addressed in the article is like, it's complicated even further by the fact that the images that it's using to do this are the images that these artists have created previously, right, and have shared on the web, you know, for the purposes of promoting their own businesses, or, you know, an orphan, just artistic expression, but like, not with the intent of making themselves redundant, that's for

 

Unknown Speaker  6:55  

sure. Yeah, the dataset is so huge, I think people really had a hard time wrapping their heads around this. And even those of us who've been following this stuff, have a hard time wrapping our heads around it, like this is it's not just a lot of images, it was like the entire visual Internet. In this instance, you know, it's like a better way to think about it just like, it's not like somebody fed, it's not as much somebody like was like, Oh, here's a bunch of cool images, let's like feed it into this AI generator, you can do that with the models, like, you know, you're like, I want a Star Wars looking one, you can like create a model that does that. And arguably, these rip off artists too. That's like another conversation kind of, but this is like a cross section of the internet period. It's like somebody just like took like one of those like, you know, little cake server things and like got a chunk of the internet and then like service into, it's like, it's just so much it's hard to even comprehend. It's like, it's all kinds of stuff to Andy bayeaux, who's actually a friend of mine, and a brilliant, really thoughtful person has been blogging about the origins of some of this stuff in general. And he's had a lot of really good blog posts on a waxy.org, I believe, is his website. But it's like, you know, it's just from everywhere. It's like separate social media sites from SmugMug from Flickr DeviantArt art station. And like, I think 8% of the images he sampled, which were just like, some millions, which is still a small portion of the whole image set were from Pinterest. So it's just all over. And those images have been scraped from somewhere else or double uploaded and had the watermark stripped, you know, is it? It's just everything?

 

Darrell Etherington  8:10  

Yeah, yeah, there's no consideration to copyright or rights management or anything like that. That's not even that doesn't enter into it. Right. It's just, whatever. It's like the bottom of the ocean. It's like when you do the job, you just trawl the bottom of the ocean? Yeah. And you're just like, whatever. I got some crabs in here, but also everything else. But I got the crabs I want.

 

Unknown Speaker  8:30  

It's crabs are great. Yeah. And it's like an organization that did this. So it's like, you know, you have the the person who made this model, which is called Stable diffusion. There's so many layers, right? You have like Lindsey AI, this is a it's not the company and Lindsey is app that is Prisma. Labs, they've been around for a minute. And then above that, you have the model they're using, which we don't really know like, what the model is that they made that kind of like tweaks the data set, you know what I mean, to get what they want to get the the cool girl images or like the like, badass looping boy. But then below that you have, yeah, you have Stable diffusion, which is the AI model in general. And that's actually free open source tool, anyone with like, a decent PC can run. I mean, it might be a little beefier requirements than that, but anybody can run it. And then above that, you have Stable diffusion. And the guy who runs that he pulled this data from like a nonprofit organization, I believe their nonprofit organization who just have these giant datasets. So there's like so many layers of I want to call it passing the buck. But like, if you're looking for, like, Whose fault is this? Is this ethical? There's like a lot of different people you need to be asking.

 

Darrell Etherington  9:28  

Yeah, for sure. And you talked about this a bit in the article where you're like, I think the barn door is already open or whatever. Like the metaphor is like the water can't go back in the jug in some ways, right? It's like where do we go from here? People are already paying for it. It's already kind of like jumped past so many stages that other things would have had to go through in order to get to this point in Canada ever get back to a place where it's acting in a way that you know, artists believe is fair and ethical and a general observers I mean, maybe general observers feel this is fair and ethical, right? Because it doesn't I don't see a lot People to your point, I still have people using this in my social media. My other circles are not techie. But I don't see a lot of them weighing the ethical concerns, I would say. So are we past that point? Like is Do you see a way to kind of resolve some of those issues ever in future?

 

Unknown Speaker  10:14  

I mean, I think we have to grapple with the issues. And you know, depending on what your like prior, I guess, your biases going into this, or you'll you might be like, This is no big deal. This is cool stuff, or like, this is an existential threat to artists, we know it and actually feel both ways, depending on what minute you asked me, you know what I mean? So it's like, everybody's kind of gonna have their own stuff they bring to this conversation. But I will say, like, there has been not with Lindsey AI, but with the model that's trained on this Stable diffusion, AI model and open source tool, there has been pushback that resulted in an updated version of that tool. So like, a big thing that happened, that kind of intersects with this whole conversation is like, basically, you could be like, Oh, I'm, you know, I'm a badass illustrator, Darrell, I'm an awesome illustrator, I do things I draw, like the original Mickey Mouse style, or like, whatever, I got these great crabs, these, my crab illustrations are the best in town. And so your name is actually like a style that might be recognizable for people. So the model used to be you could go to this and you could be like, hey, I want like 30, badass crab images that all look like they were made by Darrell, you could like type in the artists name. And there would be tags associated with that. And people were building models like this. There were like, literally ripping artists, whole styles individually, sometimes, like kind of indie artists, or people who didn't even work for like, you know, giant corporations or whatever. And then these artists would be like, they've come across this and be like, Oh, my God, I was like, What is this is a terrible. So they actually did update the model with a version two that has that removes the ability to like, at least linguistically tag an artist to a model. So you can't be like, I want to rip this guy Solomon's search. Daryl, like that doesn't work anymore. But again, the horses left the barn, like all the models are already trained on the prior data set are all there, you know, all of this is already kind of out. And they're iterating on it. So it's complicated.

 

Darrell Etherington  11:52  

Yeah, absolutely. Yeah. Because I admit to having tried some of that with the first version, because I was like, I think I saw someone share one of the arcane, like the Netflix series, which is like an excellent, beautifully animated series. And I was like, Oh, that's cool. I want to do like me in the arcane style. And I think I went and built the model and like, followed their instructions to install that particular branch, or I forget how, what it's called, but you can install like, someone will have the pre formatted things, if you have a GUI version of Stable, the fusion or whatever, you can, like put this in a folder, and it will adjust the model to that specification, right, which is, which I think you'd still be able to do under this because it takes away the key wording, but not necessarily if you want to adjust the output or tailor or customize the output of the model at a specific direction, which is what

 

Unknown Speaker  12:38  

you can absolutely still do it. And yeah, I mean, one question, I would love it. And I don't think this is what I'm going to get an answer to this is earliest, like, their models are so good. And the images look so cool. Like, what did they make their models off of? I mean, they're like an actually, like personal labs, I guess it's actually a company. It's been around for a while we've written about a TechCrunch before, but like, you know, it'd be interesting. We came out like, oh, like they ripped some other artists. I mean, I'm not saying that happened. Like, we don't know that, but like, these models are really good. images look really cool. You know, I did the one for the purposes of research and also vanity. You know, I feel really conflicted about it, like I said, but I did the one that, you know, spits you back, like 50 images. And I would say like, like, some of my friends had like horror show images. They're very bad. And like a solid 40 of mine were like, awesome looking. Yeah. And I'm just like, so shocked that the results are that refined out of the gate. And I mean, this is like catnip for social media, obviously. Yeah,

 

Darrell Etherington  13:25  

absolutely. Yeah, I did. So I did the 100 block, which I think is 10 bucks Canadian or whatever. But it was, yeah, it was a very good hit rate, especially versus because I have played around with this previously, like you said, and some of the more like, are the less successful, I would say ways of doing it, but like the hit rate was not nearly as good, not nearly as good. So they are doing something to refine the model that is quite effective. And yeah, I've I too, would be very curious to find out what it is. And maybe we can ask them and they'll let us know. We've asked them a lot of questions already about this. And they seem to be eager to answer some of the others because I know Ohio wrote about this too, this week, right? And how easy it was to trick the model into outputting. Essentially pornography, right, like with non consensual pornography using real people's faces, right?

 

Unknown Speaker  14:07  

Yeah. And it says it says don't submit nudes, you know, so they're kind of like, well, we hope that nobody sends us this stuff. But if they do, you know, our models will output it. Like I definitely have a few friends who I don't know if they've submitted topless photos or whatever. They're like a few like partial nudes. And I was like, Well, those are Tasteful Nudes, but I'm like a little alarmed about like, what's going on?

 

Darrell Etherington  14:26  

Yeah, and that's right. So their response was that it was a violation of the terms of service, which is we already talked about on another podcast, one of those answers, that's like, okay, I guess that's all right. But like is the Facebook Cambridge Analytica answer, right, like, well, they weren't supposed to do that. Well, it does not do the thing that they did that you allowed them to be able to do technically,

 

Unknown Speaker  14:45  

right? Yeah. You have to plan for how people are gonna abuse whatever it is that you make on the internet period. Like that's just Yeah, due diligence. Yeah,

 

Darrell Etherington  14:52  

exactly. But I mean, they are they're trying now to resolve some of that with some post generation kind of filtration or detection of content like that, and then auto filtration or censoring, right, and we'll see how successful they are. That is probably the internal cat and mouse game like everything else on the internet, but at least they're trying. But I did want to talk about because you mentioned it in this article too. Like, I mean, aside from people just really enjoying these clearly, there's also some other like positive outcomes and possibly things that wouldn't be possible or wouldn't have happened. Yeah, otherwise, I think specifically for non binary community or trans community, like, do you want to get into a little bit of that?

 

Unknown Speaker  15:28  

Yeah, this is based on my personal experience. So like, my Instagram is like, very, very, very largely queer people that I know, I don't follow like brands very much. You know, it's just mostly a kind of small cross section of like, the queer community that I know, personally, and a few that I don't, but it's in a big Congress, the conversation immediately was like, wow, there's this cool new avatar tool. And then, you know, the pushback kind of like, bubbled up a little but then the conversation that happened, like simultaneously with that, which I found really interesting, because I do have so many close friends who are trans and non binary, was it a lot of people were like, putting the stuff in the image generator, getting these images back. And they just felt really like affirmed. I mean, we felt like, you know, the community, we felt like gender euphoria people are like, it's kind of the opposite of, you know, what the media likes to talk about a lot about gender dysphoria, you know, given the negative side of being like, I'm trans, and I don't feel like my body represents how I feel. And I want to like, kind of adjust that mismatch, whatever I need to do. But the opposite of that is feeling really good. And being like, that looks like me like, this feels like me. And I saw a lot of people expressing that kind of sentiment, which was fascinating, because like that, that stuff just rarely intersects with the tech world in general. Usually, these are just kind of like diametrically opposed. I mean, I would kind of say, like, the queer committee is very sensitive to new technology, like, you know, we've been jerked around enough by corporate America and technology and everybody else that like, you know, religious, very sensitive to developments and technology, understandably. So it's weird seeing people be like, This is awesome. And I feel good. And like, here's my images where I feel awesome.

 

Darrell Etherington  16:50  

Yeah, yeah, I mean, it is I think it's definitely the exception from what we typically see where it's like a new feature comes out, and it's like, Oh, good. No one over there at a product development company acts like took into consideration the massive potential for abuse for this thing for this particular community, or these individuals. One

 

Unknown Speaker  17:09  

gay person in the room, you just need like one white gay person, one trans person, just like meet one, find one on the street interviewer like, whatever you need to do you that person in that room?

 

Unknown Speaker  17:18  

Yeah. It's cool. They don't do that. And I highly doubt that sounds

 

Unknown Speaker  17:24  

really hard. Yeah. I mean, how would you find them? I wouldn't find these people. They're these legendary people who live their lives.

 

Unknown Speaker  17:33  

Are they searching for Bigfoot? I don't know. I feel like,

 

Unknown Speaker  17:36  

I know a lot of them. But you know, I do. I guess they're very difficult.

 

Darrell Etherington  17:41  

But I do. This is another thing I would be interested to hear more about from Prisma. Right. Like they at least gave enough thought to introduce the three radios or whatever options they give you, when you go through the flow, right? Where you have, like, I think it's the gender selection screen, right? Or like gender identity selection screen, and it's like, male, female, or what was the other option?

 

Unknown Speaker  17:59  

Did you so like, you go through and you're like, Okay, so like 10 to 20 photos. And you do that? And then it's like, do you want I think it says male, female. And then it might say neither, I mean it like, I wasn't affronted by how, whatever language they used, it was like, fine. But the first time I did it, because I was like, I saw ED, my wife did it. And she got these like, really beautiful, like very feminine images. And I was like, Oh, this you're pretty, but I wonder what other ones look like and like, I don't know, my gender presentations all over the place. So I like I was just playing around with it. And then I did the I did the one that was like, you know, kind of like non binary other. And actually, I would say like, the first set of images I did were like, less impressive. And I would say this is like a very small sample. So I'm not trying to make any assumptions based off of this. But like those images I got were like less, I would say they're less good. They were less cool looking. They weren't as like epic, but also I might have sent different photos, you know, not enough angles or not enough variations. Right in the photo set. I didn't send the same photo sent twice because I was trying to get better results a second time. And the second one I actually did the male version, which is funny because I was like there's like actually nothing like male about it. I think males language they use right? Yeah, I think so. Yeah, I got the images back. And they just like looked cool. And like, you know, they didn't even do anything weird or like my long hair or anything, but there was like one image and the 50 It's so funny off to share with you after this. It's like, there's just this dude with like, what looks like a like a patchwork coat. And it's very cool looking. And I have just a giant beard. Like stuck a beard, like a huge, like lumberjack beard. I like Okay, that's interesting. That's the only it's just one image and that's the only like, you know, very overtly gendered experience with that, that others set which I thought

 

Darrell Etherington  19:31  

it would be super interesting to see this is again, this gets the model it's all it all goes back to the model because I would love to see what they specified for that because you're 100% Spot on when I think about those images. I'm like, what is it that's masculine about them? And it's like, a space

 

Unknown Speaker  19:47  

super sci fi like and I was like, I'm a space person I love sci fi and space like this is perfect for me because it basically looks I look like Jodie Foster recontact and like 60% of them. Basically, like one of them. I actually look like a But, but it's just like it's just straight up space stuff. Yeah. So you're like, Okay, I'm like a sci fi nerd and other ones kind of like nature goddess. So it's like the two genders are like sci fi nerd and nature goddess. Yeah. And then a non binary one is like, I'm not totally sure. Yeah,

 

Darrell Etherington  20:15  

yeah, exactly. Yeah. And then maybe they aren't either is the issue right there.

 

Unknown Speaker  20:20  

The datasets aren't I don't know what they're drawing on from that

 

Darrell Etherington  20:22  

maybe they didn't tweak that model as much or something they did the less

 

Unknown Speaker  20:25  

totally or they're trying to draw for I have no idea whether it's trained on but like, I would generally assume if they're doing something that's like not gendered, like that they would be more difficult to, you know, sort people's profile images and stuff. So it might be less than data that it's trained on. Yeah,

 

Unknown Speaker  20:36  

that's work. Yeah, I don't know. But

 

Darrell Etherington  20:39  

I mean, prisma, open invitation comes actually, because I feel like there's like a entire series of theses in the like, sociological approach that you took to making this stuff. It's very complex, and very weird, but

 

Unknown Speaker  20:53  

a lot, a lot of issues. And I think this is interesting, too, because it's like, this is the first one I think that most people have kind of bumped up against into like, Oh, weird AI thing. Because we complex? Yes. Yeah. So like, what's next? Like, what's the next or like, there's other services that do this, that are paid. And I think the paid part is the kind of the part kind of like, raises red flags for artists, which is what I wrote in my story that I did, but like, there's other services that do this, and people are going to, you know, run into these models in the future. So like, what's the next big viral AI art thing going to be? Is it going to be like this again? Is it gonna be something different? You know, they designed this, whether intentional or not, in a way that made it really conducive to going viral online?

 

Darrell Etherington  21:29  

Yeah, absolutely. And I think probably intentionally, but it is, I think you have the one artists quoted in your piece talking about do I even have a job in however many years, right, like, if you're doing digital Commission's a lot is if that's a significant portion of your business, a lot of that is like, hey, I want something from my d&d character, they're like a rogue half elf, or whatever. And then you can probably get that quite easily now. Right? So what does that business go away,

 

Unknown Speaker  21:57  

and it's gonna cost a lot more because paying for human laborers are expensive, because it's thoughtful a lot of time goes into it. I mean, there's Well, like I follow up with artists who like that seems to be one of the main things they do like I play d&d, and like, you know, they're always just doing like Commission's like, every single day they have like a bunch of different Commission's so it's like, this would be trivial for an AI to output this stuff and kind of put these people, especially independent artists, you know, out of a job, because I think the bigger, you know, the corporations who are, you know, I'm Disney and I pulled a bunch of animators and stuff like they're going to have to grapple with this on an ethical level, there might be protests, or maybe boycotts. Who knows, like, where this is all going ethically, you're going to independent artists, and you're just relying on like, people, you know, in your circles and people who advertise to online to support you and buy your work. Like you could get like this can make a big dent in your, your income, and you know, what you do for a living very quickly.

 

Darrell Etherington  22:43  

Yeah. And to your point, like, versus those large companies, there's not really, it's going to happen silently. It's not going to happen with a huge outcry, or with a lot of like public support in your corner or anything, necessarily.

 

Unknown Speaker  22:54  

Maybe I'm being optimistic that that companies will be dealing with that. But I think there's going to be I mean, especially with like the kind of the labor stuff is going on right now. I think there's gonna be a lot of companies are going to have to make statements, they're gonna have to clarify what percentage of their stuff is AI generated? Yeah, it's gonna be Oh, boy, civic,

 

Unknown Speaker  23:08  

and ethical, like, yeah, and ethical. And who knows? Yeah.

 

Unknown Speaker  23:12  

Or there's gonna be organizations, I'm sure that like, pop up runs kind of thing. I mean, it's, we're, this is an interesting year, I think we're seeing the beginning of a lot of this stuff. And it's complicated.

 

Darrell Etherington  23:21  

Yeah. Yeah, for sure. I mean, it, frankly, is a very exciting time to be talking about and writing about and thinking about this stuff, and how, you know, we've seen some positive outcomes. And hopefully, there'll be more as well, we don't want to paint a huge doom and gloom thing, but absolutely worth making sure that the people who stand to be most impacted most negatively most quickly, are represented in this as well and not lost in the mix. But thank you so much, Taylor, for joining us and for talking about this. Yeah. Next, we have something a little different for you. We could have just told you about this week Space event, but we thought it'd be more fun if you heard it from the attendees themselves.

 

All right, great to be here. This is our third annual TechCrunch space demand. And many of you may have been here before for some of the other ones. So the original plan that we had for this year was we're going to host it on one of the commercial space stations in orbit that remember those as Yeah, so we went with our number two choice, which was at the Sheraton ballroom course. We will rely on today. But let's get started. So first up, we have a terrific panel of investors familiar with the space as it were, so please welcome to rebel for playground Arboga from Serafin and Emily Hendrickson from Route ventures and your moderator Kiersten.

 

Unknown Speaker  24:57  

So, my name is Mackenzie how Justin and I work for Silla minute As I'm the Lead policy and client relations expert, so I'm help building our client facing side of our startup, where a hard and deep tech company that focuses on improving behavioral mental health in space for astronauts, and really anyone who, who's entering the space.

 

Unknown Speaker  25:17  

Awesome. So I guess, when we look at today's event, what's been the most exciting thing so far?

 

Unknown Speaker  25:23  

Yeah, I think the most exciting thing for me so far has been the integration of startups, and government executives, company executives, it's really a fostering ground for relationship building. And I've really appreciated how startups are put on an even playing field here in terms of making business connections and learning new things. Awesome.

 

Unknown Speaker  25:44  

What's been your favorite session so far, and what's been your biggest takeaway

 

Unknown Speaker  25:47  

from them? You know, they've all been great. One of my favorite sessions was with the Assistant Director of acquisitions for the US Air Force, I really learned a lot about what the government is looking for in startups, and how to partner with them successfully, as well as how the industry is moving forward.

 

Unknown Speaker  26:07  

So I guess when we look at the industry as a whole, what's the most exciting thing that's happening?

 

Unknown Speaker  26:12  

I think the most exciting aspect of the industry right now is just how quickly it's evolving, and how that speed is really opening up areas of opportunity for diversification. And for smaller companies who I work for a woman run under 30 company. So we're really young, but we're really talented. And we come from very different areas of the world and backgrounds in education. And I think the industry's opening the door for that diversity, diversification more than it ever has.

 

Unknown Speaker  26:44  

My name is Lauren, Seamus, and my company is epsilon three, and we're here to meet everyone else in the space industry. And our founder is also a keynote speaker this afternoon, Laura Crabtree. Well, the most exciting thing for me has been meeting all the other people who have been at this event and hearing their stories about their businesses in the industries and just understanding where they are, and then their own journey in the space industry.

 

Unknown Speaker  27:07  

What's been super exciting in the aerospace industry,

 

Unknown Speaker  27:11  

I would say the privatization and how fast we have as individuals been building up the industry moving away from something centralized with the governments like NASA and things before, and the incredible innovation that's come out of that. And I'm constantly blown away by the things that people are thinking up and hearing stories just from how we're impacting not just launch and the things that come top to mind. But our case like software and helping the back ends and, and then things I wouldn't even think about like advanced safety systems for sending our actual humans into orbit. The biggest takeaway for me is just been hearing about how people are now getting involved in the green industry, and especially with things new new ventures that are gonna affect just the planet and how we're thinking about that as a whole. Hi, I'm

 

Unknown Speaker  27:58  

Robin Stonesifer. And I'm chief of staff of people connect. And we're a boutique recruiting firm focused on startups. So helping startups grow their company, for any level, any position for any type of company. And we also do pitch force, which is a weekly pitch event for startups to pitch looking for either their seed or a round funding. I think what's really great about where all industries are, at this time, especially the space industry is it's getting to be more reachable for startups to enter the space. And it's not just those big box names anymore. You hear a lot of people at all different levels of companies interacting at the seabed, sharing ideas and connecting and being open to that.

 

Unknown Speaker  28:44  

Yes, Eric Frank's work for a company called sis loon that has a couple of NASA contracts for lunar construction technology.

 

Unknown Speaker  28:51  

So I guess my first question is, what do you think about that? What's been so exciting?

 

Unknown Speaker  28:56  

Yeah, it's been amazing for networking and talking to people regarding opportunities for collaboration and partnership on some of the some of the work that we need to do that. We don't necessarily need to do by ourselves, but we can work together to actually achieve these goals for NASA and the country. I think the fundraising session was really good. I got to hear from Mark Bogut it sir, from space capital. And some of the ideas that they had regarding the change in the fundraising environment are promising and especially for early stage companies, that they're still able to raise money, things aren't that different and kind of a flight toward quality? I think they mentioned as well. I think the biggest thing, the macro trend that's critical for all of us here at the conference in the space industry is the reduction in launch costs that's coming with competition, particularly with SpaceX and their use of the Falcon nine for launch of the Starlink satellites. I mean, over 3500 have been launched already. And they've recently got approved for many 1000 more so it's a real priming of the pump for all of us that are trying to get to Space, being able to just hop on a ride and get up there any week of the year that we want to

 

Unknown Speaker  30:15  

everyone you know what, thank you so much for sticking around was a great program. I want to thank everyone at TechCrunch for helping to put this on. Thank you all very much

 

Darrell Etherington  30:38  

that's it for this episode. Thanks for joining us and thanks to Andrew Mendez for his help this week. Be sure to check out all the other TC podcasts found equity chain reaction in the TechCrunch live podcast. We'll be back next week. The TechCrunch podcast is hosted by myself Managing Editor Darrell Etherington are produced by Maggie Stamets with editing by Cal Keller Bryce Durbin is our Illustrator Melissa stringer leads audience development and Henry pic of IT managers TechCrunch his audio products. Thanks for listening and we'll be back next week.

 

Transcribed by https://otter.ai