AI + Robots, YC Preview & Why the Cool Kids Keep Picking on Tech

[00:00:00] Hi, I'm Eric Newcomer. And I am Madeline Rynbarger. And this is The Newcomer Podcast.

[00:00:10] Each week, Eric and I discussed the VC deals and the drama that went down.

[00:00:15] Let's do it! Here we go! Alright! Let's dive in!

[00:00:20] Jumping into the deals that defined last week.

[00:00:22] AI Legend, Fafely, who's been a huge thought leader in AI doing her own company? Why wouldn't

[00:00:30] she raised $230 million to build large world models to interact with 3D space, came out of stealth

[00:00:36] and raised from a who's who adventure capitalist, and recent horror wits, NEA radical ventures,

[00:00:43] and includes big names like Mark Benioff, Jim Brieyer, Ron Conway, Jeff Dean. Yeah, I'm not going to read them all.

[00:00:49] Lots of them! And fundamentally this is what Robots Plus AI, right, Maddie? I mean this is the world you've written a ton of

[00:00:55] stuff. Right! Yeah, I mean this is a really booming space within as a subsector within the AI funding

[00:01:01] rounds that we've seen, robotics and foundation models that can help translate into movement

[00:01:07] and physical space and understand the world beyond just self-driving cars. It's something that's been

[00:01:13] AI has really provided a big breakthrough for. So it seems like there's lots of different companies playing in this space.

[00:01:19] I will say there's physical intelligence, lucky groups company along side-smeather co-founders raised

[00:01:26] to sizeable round earlier this year. And then there's several other smaller ones that are doing more kind of

[00:01:32] specified in factory robotics versus humanoid robotics companies that we're also seeing and there is kind of a debate around

[00:01:39] whether humanoid is the way to go or we should build different kinds of robots to handle these models

[00:01:46] and so we should build things for their specific tasks rather than try to replicate a human assistant.

[00:01:52] Although the human assistants fulfill a sci-fi fantasy for some people, so that's cool. It's a huge round, lots of investors.

[00:01:59] Alright, that was Faith A. Lee's world labs for our second notable deal defining this week. I think you're going to take us back

[00:02:06] to my old neck of the woods. I use once covered delivery very stittiesly anyway.

[00:02:13] What's the deal?

[00:02:14] Yes, we're jumping back into the heyday of delivery companies. If you remember in 2021 there were quite a few of these rapid delivery

[00:02:22] startups. One of which was Flink, German company in the space and they raised 150 million in new funding from investors bond

[00:02:32] in the world. A Scandinavian supermarket giant all jumping into this round and I think what's most notable to me is how far the valuation

[00:02:40] is fallen. This one places it just under a billion but this was a company that was valued at five billion in 2021

[00:02:48] and there have been a lot of competitors that have been acquired or just fallen out of the space, especially as you know,

[00:03:02] where that space ultra fast delivery, where they stock warehouses with specific items and then drive drivers get it to you as quickly as possible.

[00:03:10] It's one of those things that just seems much tougher in execution when it comes down to it, but obviously Flink has held on and has still running a business,

[00:03:20] but it just seems like the margins don't look as frothy as they once did back in the day.

[00:03:25] It's amazing how much the narrative just forgets about these companies.

[00:03:28] I feel like they never got to the level of Uber. I mean, sort of like lime in the scooter businesses were, they seemed crazy.

[00:03:38] I mean, I would say one similarity between the scooter businesses and these is a required actual spending.

[00:03:44] They're not like Uber was fundamentally a marketplace with sure a lot of subsidies and other costs it burned through billions,

[00:03:52] meaning billions, 20 billion plus but it's still like you could see how it could cut them all away and have a marketplace at the end of the day.

[00:04:00] Whereas you have to build warehouses or you have to buy scooters that's a tough business.

[00:04:05] And you know, we don't even know the terms of this round. It's like under a billion raising 150 million.

[00:04:12] I imagine that's it's a tough round. It'll be interesting to see if there are any soft landings in this space at all.

[00:04:18] It's a tough round for sure. And I do feel like there's been kind of entrance outside of the US that have gained more traction several of these companies are based in India as well.

[00:04:27] Or they've seemed to have more success. So it could be just that it's not a market for US consumers and if you take a more global approach obviously flink is from Germany.

[00:04:35] So they could have, you know, some staying power in a market outside of the US and that could be enough to keep it going.

[00:04:41] But it certainly is a fall from the the Zerberaheiz.

[00:04:45] I'm going to give one sentence on a roll of mansion, which is of course the open AI deal this 6.5 billion dollar funding round is coming together.

[00:04:53] I think the information reported the other day that investors had a minimum 250 million investment.

[00:04:59] It's going to be bigger than like all the funding start up world. You know not so long ago anyway. So he's open AI start up in the traditional sense. I'm sure people debate but anyways, it's an epic round and so we'll keep watching that one.

[00:05:14] Jumping into our top story this week, why combinators demo day is taking place Wednesday and Thursday. It's the you know semi annual event that's about to happen more times a year.

[00:05:24] We get four of me year now. Yeah, so many more news to watch.

[00:05:29] You really captured what's happening in YC demo day this week. So I won't steal all your punches but obviously it remains artificial intelligence driven lots of boring enterprise software.

[00:05:42] Were you tried to drill into some of the areas where there's some new stuff. What do you seen with the YC demo day right now?

[00:05:49] Yeah, so this was the first demo day after they released their latest call to start up that mentioned defense tech as an area or space tech as an area that they were interested in missiles.

[00:05:59] I think it was for missiles in the headline. Oh, we had to put missiles in the headline. It is YC's first weapons company period not just missiles company came out of this batch areas industries.

[00:06:08] Well, why see call for startups and their batches post that call usually are pretty reflective of trends are in Silicon Valley at large. They do lead in some ways on the cutting edge. But they also are reflective of where VC interest is a lot of the time.

[00:06:23] And it has shifted more to hard tech than it's been in recent years.

[00:06:27] I will say an investor texted me after our story came out. I think we were somewhat skeptical of areas in the story too, but we were yeah this person said the company is so awful.

[00:06:36] And I was I said, in quality or intentions. You know, he's a missile company. He said quality. So I feel like there's the investors lots of Silicon Valley investors are ready to bet on a missile company, but we'll see if this is the one or not.

[00:06:52] Right. There's a lot will come down ultimately still to the team, especially with a seed stage investing where the team is so important.

[00:06:59] I think also, I mean just in my conversations with some of the people over at founders fund which have been really heavily involved in funding and founding and rural especially trace defenses co-founder over there.

[00:07:10] That company has become sort of the banner success defense company. So obviously there is room for more than one but how many more remains to be seen at this point. So they could a lot of these companies could eventually become prime acquisition targets for like an enderular or a palentier or something like that.

[00:07:27] Or prime. Yeah, or prime exactly. I pun intended another company among the defense tech plays in this batch was the easiest which is building hardware and software for drones.

[00:07:39] So not necessarily weapons based drones that could be among their clientele and that was a really hot round highly contested first round capital ultimately ended up winning that round out from a couple other tier one investors.

[00:07:50] So it'll be interesting to see how that company grows.

[00:07:54] In our story besides looking at the companies we also interrogated this sort of growth story of YC I remember a few years ago I had lined a profile of YC in the Jeff Rolls to Nera as YC equals growth that was during the pandemic where they could do everything remote and they were scaling scaling.

[00:08:11] They became a lot more international too during the pandemic like they accepted a lot of founders that were not based in San Francisco based all around the world and now that has sharply shifted under Gary Tan.

[00:08:22] YC is in person you must relocate to San Francisco. It's been this way since 2022 but they've really honed it on that being the message that YC is also now very much based in San Francisco.

[00:08:32] So we peaked it 749 companies a year in 2021 fell to 639 and 2022 and now we are at 528 in 2024.

[00:08:43] That's part of the new batch strategy is keeping it at that same size around 500 where they want to stick which is still a lot bigger than what it was pretty.

[00:08:52] Almost 5 times 2011 you know 2012 was 151 and 2013 was 99.

[00:09:00] These batches that had like the iconic YC companies like Airbnb in them were much smaller when those took place.

[00:09:06] So this is in some ways a way to maintain sort of the point that they're at now with scaling back a bit from the pandemic but then also offer smaller batch sizes so hypothetically more personalized time with group partners.

[00:09:19] So it's kind of how to scale while remaining small is sort of the puzzle that they're figuring out right now and investors tell me that I mean the batch quality has been great these last few.

[00:09:29] It seems like it's paying off but we'll see once they add in some more batches.

[00:09:33] So there also has been this debate you know you look at YC's winners on their top companies page within their directory and a lot of the biggest companies that have come out of YC really did come from sort of this golden era of the mid 2010s.

[00:09:47] I put it another way when Gary Tan wanted to talk about founders mode he was still talking about Brian Chesky you know Airbnb from the class of 2009 which still there's dining out on certainly coin races and other top win door to an amplitude drop box.

[00:10:05] So a lot of sort of old names.

[00:10:07] I think there's some more promising companies up in recent classes I think cruise was a YC company so I mean there have been some.

[00:10:15] That's not the recent to me.

[00:10:17] Yeah, but more recent than 2009 let's say but that's a good point that there is still much to prove in this new cohort and part of this is also the world of seed investing it takes a long time to see which that's take off and these are pre seed companies so we should be thinking on you know 10 year time frames

[00:10:35] but the size of the batches and potentially the delusion of quality of the companies that came about during this massive time of expansion is something that's clearly on the top of mind of Gary Tan and the YC team going forward

[00:10:48] They want to retain their banner Ivy League brand of start-up accelerators and for the most part investors still view them this way is from my conversations with people who back YC companies they do view them there

[00:10:58] but my commentator is also facing a little more competition from accelerators that it didn't have in the past well tech stars has kind of fallen off the neo accelerator has grown quite a bit and many investors speak to the quality of the founders that come out of that also there's the hf zero hacker house or hacker residency if you will

[00:11:17] I know that they're sensitive about that it's not a hacker house.

[00:11:20] It's a real residency accelerator, but those are deep pockets that that accelerator to fund these you know AI forward companies

[00:11:27] so why commentator is not the only game in town anymore so quality and making sure that the experience is as personalized in depth and what founders want out of it is definitely something that's on the top of mind for them

[00:11:39] Our second story of the week Microsoft taps three mile island nuclear plant to power AI crazy I mean just as

[00:11:46] a nuclear it's the 20th century a sign of how much power inside or is an AI really think it's going to take to get what a G I already get to the

[00:12:00] you know 10 billion dollar compute large language models so if Microsoft is saying oh we need to find nuclear power to make this work I think

[00:12:08] that's as much evidence as you need of how consumptive I think these models are going to be and how much they're going to keep

[00:12:14] investing enormous sums I mean I went to a tech conference a couple months ago now and certainly power was one of their top concerns

[00:12:23] not to you know plug you know sort of valley again but power was a big conversation topic at our sort of a new york event and one of the discussion groups how we will power all of the compute needed for AI to become as ubiquitous as we expected to be and

[00:12:37] the future so I think there's also you know the conversation around clean energy and a lot of the tech world had been moving climate tech had been has is still you know of an

[00:12:48] investing space that some people are putting lots of money into because there's you know an economical side to it we can't have the AI revolution if we're burning earth resources you know where nuclear falls into the

[00:13:04] field and it can be another way to meet the power demands without overly relying on fossil fuels for AI yeah if this is an excuse to invest in nuclear like God speed to the core idea of whether

[00:13:15] AI will require this I think there are two themes sort of at once right it's like okay models keep needing to get bigger and bigger to get smarter and so if you chart that out you're like man we're going to need a lot of power lets

[00:13:27] invest in it on the other hand you're seeing people use sort of like cheap models and open source models and small models to get a lot of stuff done and certainly a player like Microsoft has sort of a

[00:13:40] vested interest in believing like oh our big models are going to dominate so we're going to need a ton of power but there's certainly a world where it's like what's tech good at doing it's like doing things

[00:13:51] cheaply and sort of lightly and like you know the returns from all these foundation models erodes so quickly so it just leads me to believe like are we really on this escalating

[00:14:03] sort of power consumption curve or people going to find ways to do most of the stuff we need to do fairly cheaply as chip competition maybe comes up more to when people are there's you know the

[00:14:13] shortage that is slightly easing but still demand for new videos chips is really high but there's all these ship companies that are pitching themselves to run these models at a much more efficient pace using a lot less energy so I think competition in the market could also just bring that on the hardware side the demand for energy down a bit too

[00:14:30] Matty you want to go run way sure runway reached an agreement with Lionsgate to train its models on the studios library of movies and TV shows this is the first of this kind of these deals

[00:14:41] that a lot of these model provider companies have been making with publications for the most part but this is you know a video model that got a licensing deal with a production studio so I mean there's some debate over whether or not

[00:14:54] these studios are kind of signing their own death certificate with these deals similar with news media companies breaking in these deals with open AI but I don't know Eric if you have thoughts on these

[00:15:08] you can control the movie industry certainly I wouldn't train my potential disruptor to create movies with models but if I'm just one studio among many that has earnings

[00:15:19] pressure and wants to make money today and if not me you know my biggest enemy then sure all strike a deal with run way to let them train and make the money now and hope that I'll figure it out down the road

[00:15:30] so this is sort of yeah it's business incentive same thing we're seeing media and fundamentally I think it's good that run way can take a crack at building models

[00:15:39] and do so you know licensing with the licensing agreement to where it's not scraping these kind of movies to train on you know right as long as the argument isn't well we'll seal it if you don't give it to us certainly that's always the other attention with this like they're going to be rogue actors are going to train and then you'll see them in a decade you know absolutely

[00:15:59] and the other funny thing is you know just run way cTO and a status spook at cerebral valley New York and June and he was the one who was sort of saying you know we're going to have full AI movies on run way like imminent what like two years or something

[00:16:13] and our very AI friendly audience was like laughing at him fundamentally so I love run way I really like the team there but they certainly are projecting this sort of revolution and how movies are created that even some of their fellow technologists seem skeptical of so

[00:16:32] we'll see what they can make it this well obviously they've been used in production with a 24 movies before with everything everywhere all once

[00:16:40] yeah but they're not like you know there there's sort of promising like the whole movie which is narrative it's like I don't think anyone doubts

[00:16:47] like you're going to do special effects stuff with AI but I do think if it's just special effects my wife is a documentary filmmaker

[00:16:55] I don't know I'm pretty bullish on Adobe a company that really understands creators and can plug into creative tools that are already using so you can see why there's pressure on a company like run way

[00:17:05] to offer sort of the full answer because that would be truly disruptive and not something maybe that Adobe has prepared to do

[00:17:11] yeah that's a good point so it seems like with run way there is also the pressures on to be able to deliver a quality full AI made film whether short or feature film that people will enjoy watching

[00:17:27] and so training on you know narratives that people have already enjoyed certainly will help with that but it remains to be seen if that really is within you know the next two years

[00:17:34] or it's going to take a little bit longer to get there.

[00:17:37] Our last story the week more lighthearted or sort of sad I don't know the coming out you look at it.

[00:17:44] comedian John Malaney who I love was I think the close out speaker at Dream Force or sort of helping wrap them up with entertainment

[00:17:53] and he just sort of roasted the audience, pulled a couple of the quotes.

[00:17:58] Imagine me as John Malaney I'm not a not a person impersonator here.

[00:18:03] You look like a group who looked at the self checkout counters at CVS and thought this is the future Malaney said.

[00:18:10] If AI is really smarter than us and tells us that humans should die and I think we should die he said looking out to the crowd from center stage so many of you feel

[00:18:20] Incidentally replaceable can AI sit there in a fleece vest can AI not go to events and spend all day at the bar.

[00:18:30] I mean well it has flashbacks of when TJ Miller hosted the crunchies and insulted everyone for sure.

[00:18:37] I think it's a little more lighthearted than that to be fair and I think John Malaney as a comedian tends to land a little bit more softer but it was not exactly a receptive audience perhaps at the same way.

[00:18:49] Although yeah what do you think Eric I mean I also a huge fan of John Malaney I love John Malaney.

[00:18:53] I love John Malaney I love John Malaney.

[00:18:54] I love John Malaney I love John Malaney, he's kind of funny.

[00:18:55] He's dog much darker.

[00:18:56] I saw that TJ Miller sketch your reference which I'll touch on.

[00:19:01] Oh right you were there.

[00:19:03] I mean it's a way sales force is an event of saying what you do matters to all these people who do fairly boring stuff you know it's like I'm Mark Benny off

[00:19:12] I'm gonna donate a lot, we're creating a culture here.

[00:19:14] We have all this sort of entertainment and it's party

[00:19:17] and we connect with each other and this is more,

[00:19:21] in some ways it's an whole event oriented round,

[00:19:24] justifying people who are doing important parts

[00:19:27] of capitalism but not necessarily the most glorious

[00:19:30] and so great.

[00:19:32] So I get paid for it and the money is your reward

[00:19:36] at the whole Mad Men clan.

[00:19:39] But I feel a little bad for them that come on,

[00:19:41] this is their event to sort of celebrate themselves.

[00:19:44] I get your hiring John Malayne,

[00:19:46] but I think a lot of people think of,

[00:19:47] I don't know the pre-hab John Malayne

[00:19:49] who is a little more upbeat.

[00:19:51] Yeah, so I don't know, I feel unboven

[00:19:54] so the teaching Miller won.

[00:19:56] This was at TechCrunchies, which were these weird awards

[00:20:00] where they give out trophies to companies

[00:20:02] that the real award was going public

[00:20:05] and making a bunch of money,

[00:20:06] but it'd be like best startup

[00:20:07] and Stuart Butterfield would go up and collect

[00:20:10] for Slack or Travis Calenek would go for Uber.

[00:20:13] So it's super weird, it was to me one of the last,

[00:20:17] like Silicon Valley is really small

[00:20:19] where it's like, oh we can have this,

[00:20:21] and it was so unholywood and really uncool

[00:20:24] and VCs were the celebrities

[00:20:26] and so they bring in TJ Miller

[00:20:28] and he's like really like,

[00:20:31] I mean, it felt like this guy drunk

[00:20:33] and he was just shitty not Silicon Valley.

[00:20:36] I think he also called Travis his girlfriend

[00:20:38] one of the time a bitch,

[00:20:40] which obviously soured a more progressive audience.

[00:20:42] So there were lots of layers to it.

[00:20:45] I mean, in some ways I thought it was funny

[00:20:46] because it was like so hostile to the crowd

[00:20:49] that people were like ready to leave

[00:20:50] that as a bit, that he was just sort of at this award show

[00:20:54] turning on the audience.

[00:20:56] Like I was laughing a lot, but the crowd

[00:20:58] who is not laughing very much,

[00:21:00] but I can see why, you know,

[00:21:01] it's like part of hosting the Oscars

[00:21:03] is finding the right balance between how much you

[00:21:05] should on Hollywood and how much you don't.

[00:21:09] And some of these comedians feel like they don't appreciate

[00:21:11] tech at all, you know?

[00:21:13] At least if you're in a Hollywood

[00:21:14] shooting on a Hollywood, there's a level of like,

[00:21:17] I mean the industry, obviously I have some fundamental

[00:21:20] respect of what you guys do.

[00:21:22] Whereas I feel like for tech,

[00:21:23] it can be pretty disconcerting when people come in

[00:21:26] and it feels like they have no love for this industry

[00:21:29] and they're just taking the check

[00:21:30] and going to sort of hate them to their faces

[00:21:32] and yeah, it's somewhat distressing.

[00:21:35] It's not the same as, you know,

[00:21:36] as the inside we're all laughing at ourselves

[00:21:39] kind of way that it takes place, you know,

[00:21:42] at the Hollywood award shows.

[00:21:43] Like it does feel a little bit more like people coming

[00:21:45] in from the outside to say,

[00:21:47] like really just leaning and roasting

[00:21:49] and insulting people to some extent,

[00:21:52] despite some those insults being funny to be fair.

[00:21:55] So yeah, I think for potentially not the right crowd,

[00:21:57] but John Delaney is still a good get.

[00:21:59] I'll give them that.

[00:22:00] I mean, like can AI sit there in a fleece vest.

[00:22:03] To some degree, it is sort of like your loser.

[00:22:06] You're not cool.

[00:22:07] Like I don't know, it does read as little like bully.

[00:22:10] To meet you know, it's the classic like

[00:22:11] are you punching down or punching up?

[00:22:13] And I think Hollywood often has this sense

[00:22:15] that you're always punching up a tech

[00:22:18] because they're the real rich people,

[00:22:19] but then Hollywood is sort of all things cool

[00:22:23] and tech is all things not.

[00:22:24] And so that if you're like your loser,

[00:22:26] where are you fleece vest?

[00:22:27] Like are you punching up or punching down?

[00:22:29] And I think part of this is that the constant

[00:22:31] who's more powerful Hollywood or tech,

[00:22:34] the one hand tech has all the money.

[00:22:36] And they're the ones sort of hey,

[00:22:38] hey, monkey come dance on our stage

[00:22:40] and then the monkey sort of like mean.

[00:22:42] So it's their weird power dynamics

[00:22:44] and I think that's part of what we're seeing here.

[00:22:47] Well looking forward to the next week.

[00:22:48] I mean, Demo Day is happening Wednesday and Thursday.

[00:22:51] I will be tuning in among hundreds of seed investors.

[00:22:55] So excited to be in the zoom audience for that.

[00:22:57] But what other stories are you looking forward towards there?

[00:22:59] I'm interested to see this open AI around come together

[00:23:02] who's willing to fork up the money who gets in.

[00:23:06] And certainly the, you know,

[00:23:07] the continued evaluations of strawberry remain interesting

[00:23:11] to me and sort of where we are on improvements

[00:23:14] in open AI's tech.

[00:23:17] I've some interesting off-the-record events this week

[00:23:20] so we'll be inhaling information from investors.

[00:23:23] We will not be discussing those.

[00:23:24] But I have some names for my schedules,

[00:23:26] not fun to preview.

[00:23:27] So I'm like, oh that's secret.

[00:23:29] So, but um, that's our show.

[00:23:32] See you next week.

[00:23:33] Bye!