Listen: OpenAI's 'Great Man,' AI Regulation & Drugs

[00:00:00] Hi, I'm Eric Newcomer. And I am Madeline Rinnberger.

[00:00:04] 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! All right! Let's dive in!

[00:00:20] Alright, let's do our two deals of the week. I'm going to cause trouble with mine on our second episode back.

[00:00:26] I'm beer from the format. Madeline, you have a real deal to do.

[00:00:31] So why don't you start with that one before I offer my Twitter feud deal adjacent news?

[00:00:37] Absolutely! Happy to jump in! We love to break the format right away! That's what it's for, right? To be proven.

[00:00:42] My deal this week that I wanted to point out was UJET, which makes AI powered call center management software.

[00:00:48] They raised 76 million in series D funding led by Sapphire Ventures with GV, Cliner Perkins and City Ventures,

[00:00:56] and all participating as well, so some pretty big valley heavy hitters.

[00:01:00] And it just seems like one of the areas where AI can kind of enhance slash replace jobs a lot more quickly than other industries.

[00:01:09] Call centers are kind of one of the first jobs on the chopping block in my opinion.

[00:01:14] So for better or worse, this is something that AI is good enough to do now,

[00:01:18] and it seems like people are really willing to make bets on different ways to automate it.

[00:01:24] It's not guaranteed that UJET is going to be the call center automation winner,

[00:01:28] especially now with OpenAI's voice tool getting launched.

[00:01:31] I mean there's an argument we made that some of this could just be fully cannibalized by people using AI tools and everything in AI.

[00:01:38] Yeah!

[00:01:38] Any part of it? Who knows?

[00:01:39] Exactly! We don't necessarily need a different company to do this, is kind of where we're going to see where the most lands on these,

[00:01:46] but having some specialization definitely will help, having software that is specific to call center management and automating around those tasks.

[00:01:55] There's a certain degree of training data that you can specify for and fine tune that could help them stand out.

[00:02:01] All right! Well there were bigger rounds this week.

[00:02:03] I think what fix raised 125 million, EGYM, smart fitness software for gyms raised 200 million,

[00:02:12] so let's see if foundation raised 100 million, but I don't care about those big guys.

[00:02:16] I'm much more interested in the Twitter drama with the Par AI.

[00:02:23] This thing was all kicked off with the co-founder of Par AI tweeting,

[00:02:28] I just quit my $270,000 job at Coinbase to join the first-wide commentator fall batch with my co-founder.

[00:02:35] We're building Par AI and Open Source AI code editor.

[00:02:39] Think of better co-pilot or Open Source cursor, but you've heard this spiel already.

[00:02:44] Just total growth hackering tweet and he's sort of got what he wanted, which is a lot of attention,

[00:02:50] but then a lot of people should not have known them for ripping off the code based on open source.

[00:02:58] So being flip about whether they were breaking the law or not following the Apache license,

[00:03:04] and then you've got Gary Tan, the president of Y-Commonator or CEO.

[00:03:10] He's president and CEO.

[00:03:11] Oh yeah, both. Okay, he covers bases.

[00:03:14] We can lord and commander of Y-Commonator, Gary Tan, and Gary Tan's in the reply sort of defending this poor little start up.

[00:03:23] It raises a couple interesting questions.

[00:03:26] One, Mattie, would you say about the time we know that it was?

[00:03:29] Oh my gosh, yeah.

[00:03:30] Well there is a tweet that just said, you know, quitting your 300k job as a senior software engineer in big tech to become founder of the 1000th AI powered co-pilot and raising a modest seed is still the convales version of joining a gym in January.

[00:03:44] There's something to be said for that.

[00:03:46] I mean, I think the people are very excited about curse right now.

[00:03:49] I think you hinted at it or said it in a couple of stories now.

[00:03:53] I keep hearing it and then there's also sort of the augments of the world in a number of code startups already.

[00:04:02] But anyway, besides sort of like the fun people yelling each other on Twitter over some tiny startup and Gary getting in the muck in the replies.

[00:04:09] I do think there's this substantive question somebody tweeted, why fun for companies?

[00:04:15] Paravoyed continue melty doing the same thing diversifying risk and and Gary replies,

[00:04:21] why see us 14 equal investing partners who operate independently.

[00:04:25] It is more like Bell Labs with independent thinkers.

[00:04:28] Any given one of us has wide latitude as General St. Vessers, we tend to do no competing deals ourselves but also people pivot into each other spaces.

[00:04:36] I mean, with the volume, you know, 500 deals a year basically we're saying that the idea that YC picks a team is less and less and so it's.

[00:04:46] I mean, I think obviously much more than spray and pray they're very involved but spray and piss off some founders as you support other founders going right into their turf.

[00:04:57] At the end of the day, I think all these dramas like a good sign for YC.

[00:05:00] They want to show they're still like hackery startups, you know, taking risks like breaking the rules to build stuff that gets people's attention and that's what these more talking about these paravoyed guys so they're winning.

[00:05:12] Oh, no, ultimately this was a classic growth hack tweet and it did work like now everyone knows what paravoy is which is the startup that's two people.

[00:05:19] So there is something to be said that the the YC model in this sense definitely still works will paravoy be the best breakout open source code editor that's a fork or will you know.

[00:05:30] That ultimately be it's unmonetizable downfall we'll see now for the what if you're a listener what what do you think we're about to talk about I think guys it's whatever everybody is chatting about at the moment open AI and the sort of endless drama whatever everybody less left last week right we had.

[00:05:49] Miram or Adi the CTO who's been one of the key leaders of the company a core public face apparently told Sam Alman the day of their announcement that she was leaving devastating and I mean Bob McRue Barrods off key researchers and leaders at the company.

[00:06:07] I mean it was keep researchers on the technical side too, you know, right right after they come up with this reasoning engine they're like oh I guess we're far enough, you know, and of course while the companies.

[00:06:18] In the middle of a 150 billion valuation round.

[00:06:24] The information just reported soft bank is apparently always a good sign soft bank is going to invest 500 million dollars in open AI.

[00:06:32] Malin what what stands out to you from all the little subplots in this open AI news saga well we're in the depths of you know it feels like one of those multi thread prestige TV dramas but maybe more a little succession with some humorous bits thrown in every now and then with the the public Twitter back and forth throughout all of this mostly just strikes me that.

[00:06:57] And ultimately at the end of this it's it's Sam's company and he's basically the last one left of the original confounders and.

[00:07:05] His vision he's returned from being out stood last year it's now his to run and it seems like he's really the driving force behind all of this and.

[00:07:15] And do we think that you know reaching age I should be in the hands of one great man or like it's you know classic Silicon Valley story but that's my my my.

[00:07:26] That's exactly right that this is the Sam Altman story more than it's ever been I mean for loyal newcomer readers you'll remember that last year in November.

[00:07:37] I penned what was almost certainly the most favorable column on the team of the board everybody was losing their minds coming out to support Sam Alman carous wisher was calling the board like idiots you know Ron Conway Brian Chesky at Airbnb everybody was like.

[00:07:54] Sam's a god what what's going on how could they ask him and I was like I don't know we should remember that the anthropic was basically formed from defectors disappointed with Sam Alman.

[00:08:05] Elon Musk who of course I have mixed feelings about but he has a very negative view of Sam Altman Sam Altman was basically forced out people will fight endlessly about whether you can use word fired but it was basically forced out of why commentator so this is somebody with a.

[00:08:22] Baracky track record of getting along with people and someone that early on a polygram was openly saying is perhaps one of the most power hungry people.

[00:08:33] He's ever met of course you was saying it as a positive but I think this is the reality Sam Sam Altman sort of loves power and plays a game even with his own company and it certainly turns off a lot of people.

[00:08:47] How to open AI? So it certainly doesn't turn off investors though. Obviously now soft bank is jumping into this round

[00:08:53] This ever growing 150 billion round. I mean Eric, what do you make of the investor in this round? Like it's you know

[00:09:01] Well yeah thrive as coming back soft bank lots of money still eager to get on the opening

[00:09:07] I train. I mean obviously I mean you know I covered Uber super closely and

[00:09:15] There I mean investors were savvy to put more money behind Uber

[00:09:20] Even amidst all the drama

[00:09:23] I would say the Uber story had had a very different thing playout right so you had this all-powerful leader

[00:09:32] That sort of dropped drove the whole company in Travis Calenek and then he gets outstead with this support

[00:09:38] Of some of his key executives, you know they wrote a letter and helped push him out and then a lot of the the

[00:09:45] Executives who remain I mean if you think about Andrew McDonald basically ran all driver forever

[00:09:52] Pierre their head of Uber eats is still there Joel Hazelbaker

[00:09:55] Communications and marketing is still there so a lot of this team that built Uber into

[00:10:01] Sort of that it's core identity under Travis and sort of embodied some of the Travis stylings are still there doing

[00:10:08] Doing the stuff without sort of the sort of particular pitfalls of Travis Calenek here

[00:10:15] We're seeing sort of the opposite we're like the core loot tenants who are doing a lot of the work and research

[00:10:20] Are not there anymore. Yeah, they're all the ones we're leaving

[00:10:23] Right it's Sam's left so come Valley loves the idea the strong man leader

[00:10:27] But I do think we could look back on this and say a lot of what's key about the strong man leaders

[00:10:33] They find people who

[00:10:36] Truly run business units with a lot of freedom and understand those business units and are not the sort of

[00:10:42] Manageurial bullshit or class and that what we have this sort of founder versus manager mode

[00:10:47] But really it's sort of the layer below the the face of the company. You know, Dara gets to win in manager mode

[00:10:54] He keeps the founder mode loot tenants and here we have a founder mode company with Sam Altman

[00:10:59] Bringing in manager mode leaders with Sarah Fryer and Kevin wheel and so I do think it's gonna be an interesting

[00:11:07] Test case of would you rather have the company with the founder mode person at the top or would you rather have that next run

[00:11:15] a founder mode leaders and then fix the sort of PR

[00:11:20] Stylistic problems of the leader on the top and we're really we're really seeing two different cases between

[00:11:26] Uber and open AI

[00:11:28] I mean beyond open AI, there's also you know the tons of regulation flying around California right now of course most famously

[00:11:37] SB 1047 the bill that would have required large AI models to take reasonable care to ensure their technology didn't pose an

[00:11:44] reasonable risk of causing or materially enabling critical harm basically

[00:11:48] SB 47 the bill that is the most restrictive around how companies must manage their own AI systems and

[00:11:56] Put in guardrails to protect against harm was vetoed by Gavin Newsom so the bill that everyone that all of the

[00:12:03] Convalley was up an arms bout no longer on the table. He did however signs several other AI bills

[00:12:08] It's not that he didn't sign any safety bills some of which around protecting likeness and recording of

[00:12:14] Audio representation for commercial purposes, which I'm sure you know the actors and writers unions had a lot to say with that

[00:12:20] But yeah Eric what do you make of these? I mean even though the core one the one the big one Buzzbetode

[00:12:26] Yes Gavin Newsom focus on the idea that this

[00:12:29] Regulation was two focus on large models and not enough on small models

[00:12:34] You know the funny thing about AI regulation is that legislators clearly learned from social media

[00:12:39] That they were too slow. They let

[00:12:42] Everybody get addicted to social media and then it was too late to regulate because people loved what they had I mean same thing with Uber you know if you get

[00:12:51] Your

[00:12:52] Those back to tumors yeah, because everything Uber is my that's how I understand business

[00:12:56] But you know if you get consumers addicted to something

[00:13:00] Hard to rip it out of their clutches and so you want to maybe regulate fast on the other hand

[00:13:06] Then you're regulating without much sense of what what's to come and you're potentially blocking great things and and certainly I'm on the side of like

[00:13:14] Let's not over regulate to me over regulation tends to help

[00:13:18] In comments and big companies who can do sort of all the

[00:13:21] You know they can of course

[00:13:23] Yeah, they can afford the compliance or if they decide to not follow it they can afford the fines to break the rules

[00:13:29] So if you because usually it's a fine-based system right so so I think it's a good sign

[00:13:34] Gavin knew some vetoed this bill for the time being I mean

[00:13:39] I think there's clearly a lot of energy to regulate AI

[00:13:42] I mean a lot of times you say oh

[00:13:44] It's only this bill or bus because they're a legislature will never get sort of the inertia and energy to like do it again

[00:13:51] But I do think on AI people are watching this and take another year see how things develop and then come back in

[00:13:58] I also think with AI that's been proven to be not the case like he did signs several other bills

[00:14:03] This the session that do regulate AI at least around terms of safety and particularly you know child safety

[00:14:10] That kind of image likeness thing rather than necessarily things on the sort of industry

[00:14:14] Reasonable care model side so I think where it comes down to you know some consumer safety protections or

[00:14:21] Workers safety he's been willing to sign regulation the last thing I'll say about it is that this is why

[00:14:28] The character of the leaders of these AI companies matters so much right when people have trust

[00:14:34] In the leaders there's less pressure for regulation

[00:14:38] But then you see when people become disillusioned about Mark Zuckerberg

[00:14:42] There's much more scrutiny on these companies and I think

[00:14:45] Sam Alman has got to whatever sort of internal constituencies media investors like forget all of us

[00:14:53] He's got to keep good faith with the regulatory community or it's gonna be

[00:14:58] Very difficult for AI over all and you saw anthropic basically well, that's their whole

[00:15:04] Supported that's their whole value at right like anthropic is for for better for worse

[00:15:09] But I think very sadly positioned itself as you know we spun out of open AI

[00:15:13] Because we wanted to prioritize safety. We are you know much more focused on making sure

[00:15:19] AI is benevolent going forward

[00:15:21] That's sort of their branding and they do release a lot of research behind their models in sort of a gesture to show that

[00:15:28] Safety and transparency they do to some degree you know

[00:15:31] Like walk the walk so I definitely so I think that there's you know different ways this could play out

[00:15:37] It's not just you know AI versus regulation and perhaps Sam Alman maybe you know should he

[00:15:44] Played by the rules a little bit we've saved the most fun conversation for last

[00:15:49] You know you have to eat your carrots and think about before you get to deserve bills and

[00:15:55] Legislation before you get to talk about drugs

[00:15:58] We had we I wasca particularly it was in the spotlight

[00:16:04] But but drugs generally we're we're getting discussed Madeline. Do you want to give us a rundown of

[00:16:11] The different pockets of the conversation before we weigh into

[00:16:14] Drugstuff

[00:16:15] Happy to Eric so

[00:16:18] It all kind of stemmed from Bloomberg's Ashley Vance tweeting about an unnamed VC complaining that they had lost many

[00:16:25] Founders to I wasca. It's been in Vogue and tech circles, you know for a while classic you know

[00:16:31] Loosen a genic plant that you know makes you have go on vision quests and trip in every

[00:16:37] Founder or you know days potential founder college student dreams of doing these trips but

[00:16:43] It seems to be you know it's sparking a color roll back and forth around drug use in silicone Valley with

[00:16:49] I think Mark and recent coming out with the most refer madness take I've seen in years about the issue

[00:16:55] Saying that there was absolutely no value. I can I can pull up the tweet our lead drug scene is tracking the 60s and 70s quite closely

[00:17:02] What starts with pot and acid moves to speed and heroin and cocaine and ends in madness and death

[00:17:09] I mean literally there's the reader we have already product ties pot speed and heroin

[00:17:14] We're currently productizing acid cocaine will be next

[00:17:19] Mark in your first

[00:17:23] Yeah, I mean I

[00:17:26] Don't know about i wasca. I feel like that one I'm not gonna like real it we sort of dodged that in the newsletter

[00:17:32] I do think some I'm I think we both agree that like this should be a more drug specific

[00:17:38] Conversation like whatever you just like zoom out and just talk about drugs. It's like some of these drugs are like opposite

[00:17:44] Inspired and effect and so it's like being pro one

[00:17:48] I don't know to me has very little to do with the others. Well silicone Valley you know is in some ways the result of the mixing of the counter

[00:17:57] Culture and academic and financial institutions meets drug use and

[00:18:02] Sex and free love and rock and roll and so I mean that's always gonna be a part of silicone Valley culture

[00:18:08] I think you know also like it's it's just ingrained in sort of this, you know

[00:18:12] American dream be a California ethos like I think that the conversation should really come down more to what

[00:18:20] drugs are you doing like if you're if I awasca has you know plenty of documented

[00:18:25] You know side effects. I'm just saying let's not drag down acid with all this other stuff

[00:18:31] I mean that's the quintessential Steve jobs, you know

[00:18:35] Who does acid is like man? I'm gonna do that like every day like you'd have to be like a lunatic

[00:18:40] I just think like cocaine and acid should not be in the same

[00:18:44] sentence and yeah, that's that's honestly my main take and the as long as silicone Valley is aligned with this

[00:18:52] Or like once in a blue moon opening your eyes to the world a little bit. I think that's a good thing

[00:18:57] Certainly I don't think good to give a lot of your have your employees doing heroin and cocaine

[00:19:02] So it all depends to me what what drugs are talking also like to the further point

[00:19:08] Show me the the founders and with successful startups that are like you know raging around

[00:19:15] Constant drug use all of the time like usually all of those companies do crash and burn and if they as happening we hear about it eventually so

[00:19:22] I just I find that to be I find some of the takes

[00:19:26] You know circling around this debate of drug use to be absolutely ludicrous hyperbole

[00:19:32] I mean, I am sympathetic to silicone Valley creates these personas and so then you know not everybody is Travis

[00:19:39] Calenic or Adam Newman or Steve jobs but then

[00:19:42] 5500 other startup founders dress up as them and so if you give those sort of wannabes the idea that oh they need to go do drugs like

[00:19:51] Steve jobs they don't necessarily have the sort of genuine like spirit or disposition or know how to know what to get out of it

[00:19:58] And so then they're just playing dress up using drugs and that that probably is dangerous on on the other hand

[00:20:04] I do think there's you know if we take

[00:20:06] Since yearly this idea that people go on Iowa scotrips and then they like give up their startup

[00:20:12] There is of course the trouble a question of

[00:20:15] Maybe they did find some profound realization

[00:20:18] They're like start up was like pretty dumb and they don't want to do it anymore, you know

[00:20:22] You could go either way like maybe the the business world was was false, you know

[00:20:28] Perhaps they gave up on their you know fifth coding assistant of the week plan exactly

[00:20:35] Instead of yeah growth agging there. They're they're gonna go

[00:20:40] Farmer

[00:20:40] I don't know perhaps if you know the bad ideas disappear

[00:20:44] Maybe you know not that those ideas are all necessarily bad if you have an amazing a

[00:20:48] Power-coding assistant stay in line. I mean on the other hand you know on the opposite extreme

[00:20:53] There are you know these cases like Tony Shay right yeah mate there is this chance

[00:20:57] Marky and Dreson is living in a different world of wealth than we are and so you you know you get Tony Shay

[00:21:04] Zapos co-founder who basically had an entourage like feeding him drugs

[00:21:09] I think what nitrists was sort of key to his repertoire but you know just like an atotal drug bender and like too much money

[00:21:16] for anyone to tell them to stop given they seem to be along for the ride and so

[00:21:21] I mean there's also you know discussions around Elon Musk's drug use too that have floating around as

[00:21:26] Yeah, whether being helping or a hindering his genius yeah primarily kind of mean

[00:21:31] So so yeah I guess my message is you know keep your drug use to once a month

[00:21:40] Don't do the don't don't don't have an entourage fall you're like what are we talking about you know

[00:21:44] I mean just like get yourself under control. Yeah

[00:21:48] Everything in hot air even right exactly this is your semi-adair talk for the day

[00:21:53] Over here in newcomer

[00:21:56] Looking ahead for the week Eric you have some pretty big travel plans, right?

[00:22:01] Yes, I am flying to Seattle to moderate a panel at Madrona's conference. I'll be speaking to an

[00:22:07] P of engineering co here the CEO of the Allen Institute for AI and the SDB for product and AI platform Ashash

[00:22:16] Sharma so it is gonna be my favorite topic where are we in foundation models in the bra all how differentiated are they?

[00:22:23] How much were they gonna improve how much are we gonna spend to improve them so yeah excited

[00:22:28] Excited to talk about that. It always stays relevant, you know, especially with this open AI funding round yet to close them being

[00:22:35] You know the biggest foundation model provider private company that's out there right with have

[00:22:40] Hepatized Microsoft so how can you justify that valuation when companies like co here are worth much less and offering their own business

[00:22:48] Focus

[00:22:49] Foundation models. Yeah, that are directly targeting enterprise super excited about that

[00:22:54] All right, Madron what what do you think you're about this week? We do a lot of reporting here at newcomer about

[00:22:59] University Endowment fund performance data

[00:23:01] You know some of the largest LPs a lot of these venture funds so getting data behind the scenes to great way to see how

[00:23:08] We love you Timco shout out University of Texas systems get a little bit people behind the curtain of how these

[00:23:14] You know funds are performing how these returns on investments are looking and the biggest story I think in that world

[00:23:20] Came out last week that University of Virginia missed their their fun target for their private investments

[00:23:26] If you have any thoughts or data or between you know these university systems or pension funds

[00:23:32] My inbox is open we're very collaborative with our readers. We would like guidance on FOIA requests or

[00:23:40] Websites where public universities and public endowments are sharing their venture returns

[00:23:46] We've got you Timco. We just talked about what UVA

[00:23:50] We know obviously like Berkeley as some though there are so many different entities anyway. Yeah, you have good ideas on those

[00:23:56] We'd like help

[00:23:58] Yeah while we're just making asks we love

[00:24:01] Internal presentations on how funds are doing

[00:24:04] Obviously we would never say where we get them but those are some of our most popular stories with readers and even if they're

[00:24:11] Six month six month old whatever, they're still very interesting to us

[00:24:16] So yeah email

[00:24:17] Madeline at newcomer.co

[00:24:21] I'm always Eric and newcomer.co

[00:24:24] Yeah reach out to us and we'd we'd love your ideas on how to get info on

[00:24:30] Venture fund returns you can you can find us online

[00:24:34] That's our show tune in next week for the newcomer podcast