Sam's Checkmate: How Open Source AI and Silicon Valley Kingmakers dethroned the OpenAI emperor! Plus, Tesla's API Apocalypse has arrived.
This is Coder Radio, episode 599 for December 10th, 2024. Hey friend, welcome in to Jupiter Broadcasting's weekly talk show. It's taking a pragmatic look at the art and the business of software and the whole gosh darn world of technology and everything else in there. My name is Chris, probably, and our host, I know that, his name is Mike. Hello, Mike. Hello, hello, hello. How are you?
I'm all right. I'm all right. I got a fired up email this morning from a very disgruntled listener. He wasn't mad at us for a change, though. That was good. But he turned me on to the situation going on over at Tesla. I guess he writes an app, I guess, that you can install on your Tesla car. And there's a big change happening where I guess he just had free access to the API in the past.
Then last year they released documentation kind of like making the API seem official. And then this week they've now announced that the APIs in the Tesla system are going to be pay per use and they're massive fees. And you have to pay per vehicle. And some developers with their current API usage would be looking at millions of dollars annually in fees.
So there's this developer that makes this app called Tessie. He's looking at $60 million under the new system. And so the guy that I heard from, he's just going to have to stop using it altogether. He's just done. Other people are going to try to break it and use unofficial APIs. Other people will try to bypass it by connecting over Bluetooth and doing things that way.
There's a developer of an app called StatsApp for Tesla. He says it's totally unsustainable for any independent developer. The BetterRoutePlanner developer faces similar challenges, saying he's probably going to have to drop Tesla support entirely. It's a lot like... Well, it kind of reminds me of the moves around Twitter's APIs, and it reminds me of the changes around Reddit.
Very, very similar to what he did with Twitter, right? Yeah. It feels like that to me. I don't know. It... Is it a rug pull if they were never like super official? Rug pull is the word I want to use, but I don't know if that's fair.
You know, is it a rug pull? I don't know. I mean, they weren't official, right? They were kind of just there. I don't think, at least I couldn't find anything where Tesla came out and said that they were going to support these APIs. Yeah. For third-party use.
Well, they launched documentation, but it was kind of vague. At first, people started using it before there was documentation. Right. That was the gist I got. I mean, clearly, they're trying to monetize their API, right? Yeah. But... Yeah, but that's crazy, though, because the software platform of the Teslas is supposed to be one of the value adds.
And then, you know, having a rich app ecosystem could be potentially badass, especially if some of them are giving you cool stats or helping you find better places to charge. Yeah. That's value-add that Tesla doesn't have to bring. All they have to do is make the API available.
Yeah, that's tough. It's a shitty situation all around. This is the risk you always run when you're building on top of somebody else's API.
You know, I will admit, too, that I have a bias because it's a car. And there's just something about it being a car and pulling these kind of, like, App Store games on, like, a $60,000, $70,000 vehicle. Right, right. That I have to maybe, that I would like to own for a decade at least. I don't know. Like, it just sort of, like, touches an extra nerve. But anyways, there is a possible solution.
Get ready for this. Tesla's heard some of the complaints, and their workaround is to offer users a $10 credit towards personal API use. So the individual users, in theory, would then start paying for the API usage.
It's kind of a weird, jank solution.
Yeah.
God. Well, speaking of open AI, I wanted to chat with you about something that is a fundamental land shift that I don't think the rest of the tech world has fully processed yet. Some corners have, but definitely not the folks over at The Verge and the general like tech bro press. And that was since last episode, Donald Trump named David Sachs as the White House AI and crypto czar.
So this is an interesting position because it's one of these czar positions. Here's what Trump said. Quote, David will guide policy for the administration in artificial intelligence and cryptocurrency, two areas critical to the future of America's competitiveness. David will focus on making the clear – making America the clear global leader in both areas. Trump said on a true social post.
So David Sachs, I've actually – we may have mentioned him on the show before because he's one of the original PayPal mafia founders. He's kind of a big Silicon Valley VC guy. He was kind of famous for dumping on Solana users back in the bear market. And now he's going to be running AI and crypto from the White House. But there's an interesting thing about David Sachs.
First of all, David Sachs is besties with Elon Musk. Elon Musk is in heavy conflict with Sam Altman and OpenAI. And David Sachs, from the very beginning, has clearly seen Sam's playbook, the same playbook that we noticed. I've been doing some research on David. I don't know a lot about him. I can't speak to him very comprehensively.
But I did find this clip of David Sachs talking about Sam Altman's strategy the day that Sam Altman went to the Senate and tried to scare all the senators about the dangers of AI. These were David's thoughts. And this is going to be now the guy making – AI policy for the White House and making recommendations to the councils.
You know, Sam just went straight for the endgame here, which is regulatory capture. Normally, when a tech executive goes and testifies at these hearings, they're in the hot seat and they get grilled. And that didn't happen here because, you know, Sam Allman basically bought into the narrative of these senators. And he basically conceded all these risks associated with AI. He talked about
Chat GPT-style models, if unregulated, could increase online misinformation, bolster cybercriminals, even threaten confidence in election systems. So he basically bought into the senator's narrative and, like you said, agreed to create a new agency that would license models and can take licenses away.
He said that he would create safety standards, specific tests that a model has to pass before it can be deployed. He says he would require independent audits who can say the model is or isn't in compliance.
And by basically buying into their narrative and agreeing to everything they want, which is to create all these new regulations and a new agency, I think that Sam is pretty much guaranteeing that he'll be one of the people who gets to help shape the new agency and the rules they're going to operate under and what these independent audits are going to – how they're going to determine what's in compliance.
So he is basically putting a big moat around his own incumbency here.
Something we had noticed as well, this – Regulatory moat that Sam was trying to build. And to do it, he scared the out of everyone about the dangers of AI, about how it could impact the election, misinformation, all these things leaned heavily into the regulation and the entire play. And with this one appointment, all of that has been essentially washed away.
I mean, not the fear that he created, but all the power he gained and influence over the executive office and And all of that has essentially been washed away because now Elon's best buddy is in the White House making AI policy. And Sachs is on the record saying that Sam was trying to make a regulatory moat.
And I believe Sachs is also on the record for being a proponent of open source AI and specifically what Meta is doing with Lama. It is, Mike, the train, the AI train, with all this regulation and burden licensing that was just speeding down the track, just did an impossible 180 in my opinion.
I mean we're not there yet, but this could be a ginormous land shift in the power structure and regulatory structure around this future industry.
Yeah, it's interesting because – One, it's going to be imperfect no matter what, but it would be interesting to see someone in a position of power. If we're going to have some kind of regulation, be an advocate for the open source models rather than, you know, not. Now, you know, when the rubber meets the road, where does he really land, right? Who knows, right? This is all new.
This is all, it's kind of exciting. It's a lot of it's to be determined. It's also the federal government.
Yeah, how much can an AI czar impact can they really have?
Right, how much can they actually do, right? I mean, open AI doesn't seem to be slowing down. They just got my $200.
We're going to talk about that. Even without Sachs there, though, Elon is besties with Trump. And this is a major bone of contention that Elon has. You know, like, I don't know if you saw, but Sam was being interviewed by CNBC. And he says, oh, I really hope that Elon and I can get past this. It's really kind of a sad statement. He seems really kind of down about the whole thing.
Like, this is a big shift for Sam. He's got Sachs possibly influencing AI policy, and Elon and Trump's here saying that AI is one of the next frontiers for America, and I don't like this Altman guy.
Well, in a lot of ways, Elon ended up doing a better version of what Sam was trying to do, right? He got cozied up to the folks in power by taking what some might have called a long shot on Trump. It's going to make life hard for Altman, right?
Yeah. I don't know if it was intentional, but it was a master flip of the situation.
It will be interesting to see if OpenAI starts to signal some more friendliness to the incoming administration. I think one easy way to do that that's not that impactful, but is more of a signaling move, would be to loosen ChatGPT's... Let's try to say this in a non-controversial way, but some of it's more pearl clutching norms. Sensitivity.
Sensitivity. That's a good way to put it. Yeah. I couldn't even get it to give me some answers about the White House or generate an image of a rabbit in the White House because it's so sensitive about that stuff right now.
Yeah, I mean, I would strongly consider flipping that switch to be a little more free-for-all and then making a big splash, a PR splash. You're going to get pilloried by your California friends. But yeah, I would certainly be scared if Elon Musk is Grima Wormtongue into Trump's ear and I'm open AI, considering the just massive amounts of money they're burning.
Agreed. Agreed. I think it's a fascinating development that we don't fully understand yet, but feels like not a lot of people are fully appreciating. It's a huge power shift. For a minute there, Sam was playing the government like a fiddle. He had them eating out of his hands, and he leaned into that. And it's part of what gave OpenAI, I believe, its enormous valuation is that
Sam was openly embracing regulation. So he was signaling to Wall Street that he was going to play ball and that he could become one of the big players and have a moat. He was signaling to investors, too.
Only to have perhaps the key aspects of the strategy just cut off at its knees literally a year later that one of the hearings were the most recent hearings where Sam was scaring senators like a year ago, almost to the week. And then here we are. It's fascinating. I didn't see this one coming. No, it's definitely going to have an effect. So we'll see. Did you know you can join us live?
We do the show live typically on Tuesdays at noon Pacific, 3 p.m. Eastern. We also have a chat room going all the time, but they're helping us a lot during the show. You can hang out in there at coder.show.matrix. The stream usually gets started about a half hour or so early, and I'm now making that available to our Coder QA members as well.
You can listen in a podcasting 2.0 app, or you can just plug jblive.fm into your internet radio app of choice. And you members out there, you don't have to worry. You just now have your own private feed, and you get it. In fact, you get the show a day early as well. So if you want to become a member, that's at coder.show slash membership. All right, you hinted at this.
I couldn't believe it when I saw this. OpenAI has launched a $200 a month subscription, and it's pretty tough to even tell what you get because Pro is $20, right? Or no, no, I'm sorry. I guess it was – I don't even know the names, dude. But like the old version – which I thought was called Pro, is $20. And now the new version, which is called Pro, is $200. And you're trying it out.
I am. So Bombshell, I can't tell the difference.
Are you serious, dude? $200 a month? So you get unlimited access to the new reasoning model. You get unlimited access to the voice model. Oh, so it's okay. So it's plus is the $20 tier is what it's called.
Yeah, it's plus than pro. Yeah.
Yeah.
So, I mean, it takes longer to respond to me, which I guess is their way of showing surfacing the additional processing. Yeah. But the answers are roughly the same and it still has what I would consider the same gaps that the pro version has.
That's what I felt like when I went from the free to the pro tier or plus tier plus tier. God. Yeah.
I mean, I am going to stick with it for a month and see where I land. But I, you know, because like I have it doing a lot of like the marketing stuff, right? Just like, you know, here's an ad I want to do for Alice and I just make me like five variants and I manually check them and tweak them. But there you go. Huh, okay. I do ask it coding questions.
It still has this weird black hole with Doku and some Rails stuff that I don't understand where I guess it's just a bias of the information that it has is so heavily skewed towards other solutions.
You know, I did see them specifically mention it's supposed to be better at assistive coding.
Yeah, I think it's better. So I found ChatGPT to always be much better at Python than Ruby. But I think they made an effort to make it good at Python, right?
They really seem to be, well, their positioning of this seems to be, well, if you need advanced reasoning, then you know why you're spending $200, right? They're kind of like, if you have to ask why $200, then you don't need this tier. It's for people that really use ChatGPT. Yeah, I mean, I'm going to try it out more.
I'm doing more complicated things with it. I'm definitely leaning pretty heavy on the business side because it's the kind of work that's expensive for me to outsource or to hire in for. You know how it is, man, right? You run a campaign, and if the campaign fails... It's nice to not have paid some consultant five grand to set that up, right?
In the last two weeks, I have found it immensely helpful sorting out just like small business questions and things that I would probably almost be willing to pay somebody to ask, but I wouldn't even know who to ask. So there's things that can be very useful. And I've been using it more and more. And so I signed up.
I unsubscribed from Claude, and I subscribed to ChatGPT because of the custom GPTs, basically. Which I noticed that they said upcoming features will be the ability to browse the web, which is going to be really huge, and file uploads, which I didn't really realize since I use custom GPTs, which you can do file uploads. But, wow, I can't believe they don't have file uploads. They must. Yeah.
Maybe they mean for the reasoning stuff. They also are doing a grant program. They're offering 10 grants of ChatGPT Pro to medical researchers. So you can see how they're kind of positioning this is you need something that's for research. It's going to do the thinking for you. So $200 a month is totally worth it. To me, it kind of sounds like a desperate play for revenue.
It feels like you're, yeah, it's not done yet. Yeah, it's not worth $200 because a lot of what they're, if you go through their promotional material, is they're selling future stuff. They're selling stuff like they have this 12 days of shipments where OpenAI is going to ramp up releases and they've announced the new Sora video model. They're trying to get buzz going again.
And Sora does look very impressive. I don't know if you had a chance to see it. No, I haven't played with it yet. I'd say a lot of what Soar is producing, which is their video model, not always, but on occasion it can produce results that are past the uncanny valley that you can't tell are.
MKBHD has a YouTube video on it where he plays like AB of, you know, guess if this one's AI or real, and you just legit can't tell on some of them. Jesus Christ. Yeah, and you just go in there and you do it by prompt. Now, initially, they've only given certain people access to this, but it's the same thing.
You just go in there and you give it a text prompt, and it produces not just like an animated Jeff, but like a full video of, you know, it can be multiple minutes in length. And then, you know, we've also seen experimentation with live 3D game rendering, where the 3D game engine is just live generating this stuff. So there's obviously a huge, huge future there for this technology.
Visual generation and video generation. But the reasoning stuff, I was playing a clip in the pre-show, like a lot of people are saying like we're throwing more and more horsepower at this and we're not really getting better results. Like that was the question we had just a couple of months ago. We played that Zuckerberg clip.
You know, can we throw more horsepower at this and it gets more and more advanced? Yeah. Now, some of them are saying you cannot. And it comes down to, like, you know, you just can't move the data around. I don't know if you saw. This is really interesting.
But XAI engineers are releasing a new Ethernet standard where essentially you can wire GPUs with this hyper-efficient packet transfer over Ethernet so the GPUs can communicate directly. And they've had a breakthrough in the amount of data that they can pump into these things. So there is some changes to be had there. But... I don't know, man.
It feels like we've kind of hit this, I don't know, like this peak. It's sort of where we're sort of like if you look at like a price, like almost like NVIDIA's price chart, like it shot up, shot up, and now we're just kind of chopping sideways on what this stuff can do. And so it feels like this $200 a month pro, there's 12 days of ship miss. All of this is really to try to get
Open AI and all of this out there as the leader still. But we'll see. You know, I mean, we'll see. I think it's going to be interesting to take people's take polls at Christmas and see how many people have the chat GPT app installed on their phone. And if they're using it, I have been using it. I actually found it kind of useful. You know, the voice mode and all of that.
So it's not a bad product, but I'm going to take a little informal poll of the family and see how many people have a chat GPT app on their phone. I'm betting it's not. I bet almost none of them do. We'll see. Do you have a sense, if you were to ask your family, would anybody have a chat GPT app on their phone?
No. I mean, I do.
Yeah, me too. Yeah, me as well. But I don't know. Maybe they've gone to the web. I'm actually pretty curious. I'm going to ask them to find out. I'll have to report back. Four score and seven boosts to go. All right. Well, we're moving right along this episode. And we have a few boosts to get into. And one of our boosts is our ball. Oh, we have a baller booster this week. And he's back.
It is Mr. Turd Ferguson.
Turd Ferguson. For 88,222 sats.
Hey, Richland. early congrats on episode 600 I hope I'm the first in honor of episode 600 which past hot take haunts you the most which past hot take on the show haunts us the most hmm so many bad ones yeah yeah there's there's some there's some juicy ones this is a question I'd like to turn around and ask the audience Which hot take sticks out the most?
All right, I got one that's pretty bad. It's been bad on multiple levels. All right. My just general hatred of Swift when it came out.
Sure, okay.
One, irrational. Two, it actually didn't matter in a lot of ways, right?
Oh.
Yeah, it was a nail in the coffin of Objective-C, but a more...
calm and let's just say uh i don't know how you say thoughtful person would have been well it doesn't matter because the trend is is and was always going to be a lot of the enterprise type development i do was going to end up you know relying on web standards rather than uh native development or any of the shims like xamarin okay that's i guess that's not like a cringe take it's just wrong right
It was hot at the time, though.
The delivery was certainly spicy.
I got one of mine. I feel like the audience could probably come up with a better one. But during the absolute peak of the butterfly keyboard and just the touch bar and the soldered-on parts and the $800 fixes for the keyboard, I was really convinced that... The Mac line was essentially abandoned and that Apple was just focusing on iOS and that, you know, the Mac was going to wane.
This is years ago. But then, of course, they pivoted a little bit with that MacBook Pro 16 with the fixed keyboard. But really what they were doing was buying time. for the M series launch. And so they had kind of languished, you know, the Mac mini was a piece of crap for years.
Things had languished indeed, but that was because they were working on the M series stuff and it proved to be very successful for them. And, uh, you know, I don't know if I saw the M I, I, I don't know. Maybe I can't remember. We'd have to go back and see what our take was on if there would be a transition to the M series. Thank you, Turd, for the baller boost this week.
You're definitely doing the heavy lift for this episode. And if anybody else has any hot takes that were just awful, let us know. It'd be fun to cover them on episode 600. Tell us about our failures, please. CB comes in with 12,001 sats. Hey, guys, I'm working on a project that I'm hoping will be the home assistant of home video security. Here's some context.
My last job was writing enterprise video surveillance software and none of the current available projects met my standards. I've tried them all. Some get close, but just no home run. I've been doing the initial dev privately in my spare time, which I have very little of, but eventually I want to open this project up.
So I'm not doing it all solo, but I'm wary of doing all the work and then someone else grabbing my code and doing their own thing before the project has a chance. The goal is to be fully open source and, But I would like to have some way to financially support my work that goes into it. With financial support, I can devote more of my time towards the project.
What is your guys' thoughts on the approach to do something like this and the kind of open source licenses that would be appropriate? That's a big question. First, I want to say to give you some time to think.
I think, CB, you probably have a bigger chance of nobody noticing or caring about your project and nobody helping you than you probably do somebody noticing your project, stealing your idea, and monetizing it before you get a chance. First of all, you're the one with the vision. So they would just be seeing your code, but they don't know what your plan is. They don't know what the vision is.
They don't have the passion. So I don't think it's a huge risk factor, and I don't know if I would let that prevent you from open sourcing it. Plus... There's copywriting and things like that you can do. But I would consider that the bigger risk, is that nobody cares and notices that so many people notice that you get ripped off.
Yeah, isn't that always the case? Especially starting out a project like this, maybe it's your first of your own, I don't know. You're more likely to languish in obscurity than to get Sherlocked. Yeah. I mean, I still lean towards permissive licenses. I've heard some good arguments and I've seen some successful projects.
Now this is, I might be messed up here, but the LGPL4 has like a, it could be the LGPL4 or the LGPL3. I always screw them up. But has a way where you could make it like kind of copy left so that you can't just get Sherlocked. Again, I think that's not a huge risk in most cases. but it gives you the ability to do proprietary-ish things, which you might want to do.
I tend to still, I know this is lame, I still tend to use, I think the last thing I did was like Apache license.
There's always a good way to go. I mean, this could be, again, now this is something ChatGPT could probably help you kick around. Or, you know, insert your LLM of choice that's probably read through all these licenses. But I think it's a great idea. I've been excited about it since you first told me about it, CB. And please keep us posted.
Like, if you do get it up there and whatnot, we'll link it out and let everybody know. And thank you for the boost. User 55820033 came in with two rows of ducks, which is... First time booster here in the UK. Hey, hey! Well done. Thank you. It's an extra set of steps, I think, from the UK.
He says, I listened to the Financial Times daily news briefing podcast with a mild animus when their AI correspondent did her best access journalism to describe OpenAI's exploration of advertising. It seems like the logical next step. I dare to say maybe a final destination for OpenAI, but I still hate it. I'd love to hear your thoughts on this.
Advertising in the OpenAI results, in the chat GPT results is coming, Mike. For the free accounts. I think it's coming. I think our booster's right. Yeah, that's a slam dunk. Yeah, and then, of course, at your $200 level, I bet you they do the Elon thing. Because you know on Elon, on the Elon X, if you become a member or whatever, if you pay for the X premium or whatever, you still get ads. Yep.
You still get ads. So they're going to do the same thing, except for you $200 ballers. They won't get ads, but us plebs at $20 a month, we're going to see ads. And I think it's gross. I know there's something about like it goes from being like a library of information to a Walmart of information.
It has the Google problem, right? How many pages do you have to go before you see a non-Google property on most searches now?
Yeah. Yeah, exactly. Yes, that's it. It's going to become more of that, I think. I hate to say it, but I'm not very enthusiastic about it, obviously. I also want to say thank you for the reminder about the Advent of Codes. I wrote a text for that. He's doing it with Go. Great. Well, thank you very much for the boost.
You can go into your fountain profile and set your name and let us know who you is. And we'll give you a shout out. DG at PTC came in with a Jar Jar Boost 5000 sats. You're so boost. When it comes to kids and medications, all the most experienced professionals who see my child, the pediatrician, the pharmacist, psychologist, neurologist, they say no.
It's the teachers who are not remotely qualified to be discussing medications at all. They are all the most aggressive pill pushers. Parents really need to stand up to our children, especially when the faculty tries to get your child started on amphetamines.
Thank God we have sensible, older, experienced professionals to help disqualify the wild, uninformed, and reckless opinions our school has been slinging lately. Wow. Damn. Love the fire there.
Yeah, no. Well, it's his kid, right? Her kid.
I don't know if it's he or she. Coming in hot with the booze. Yeah. So they definitely did that with my son a couple of times. Really, really for enough where like as a family, we had to have the conversation. Are we going to do this? Are we going to put him on medication? And as a kid who went through late elementary and 70, 75 percent of high school on ADD medication, I was against it.
I did not like the experience. So I tend to lean towards DJ at PTC there.
Thank you for the boost. Was it because he was just like being slightly disruptive?
Yeah, I mean, he was being disruptive. He was bored. He was not behaving all the time. You know, he can sometimes, when somebody tells him to do something, he can ask why. Those kinds of things that aren't great in group settings. That's like kid stuff, though. I know. He's gotten better at it, too.
Now it doesn't come up anymore. He's a kid. I don't know. Yeah, I mean, I have no direct experience with this with my child, but I know one of his pals is on it now, and the poor kid's a zombie.
Yeah.
It's really, the change is just shocking. Yeah.
Yeah. Yeah. It's funny. I remember because when they first started me, I was overmedicated. And so I was a total zombie. It's like he's stoned. And that. Yeah. Oh, yeah. And the teachers loved it. They love they went on about how great I how well behaved I was, what an improvement has been, what a great student I am. Great, let's sedate 25 kids in the class and see where we go.
Yeah, I mean, I suppose I might want to do it too, but as a teacher, I'll tell you what, I've spent times in my life, can I just sedate these three kids I've got? I kid. All right, our next boost, I'm going to say it's... It's Glavata. And they come in with a 5,000-sat boost, which is a Jar Jar boost. It's a boost. And they are a first-time booster. Two first-timers this episode. That is awesome.
Thank you, everybody. I know that can be a hike, that first journey to get that all set up. I'm glad you found Fountain. I think that probably is the easiest path. He says, thanks for the fun shows and keep it up. Thank you. Thank you for boosting, Mr. G. Jordan Bravo rounds us out with a row of ducks. 2,222 sats. Honest question. Have we already hit peak AI hype? Or is it still ahead of us?
I hope it's the former, but I fear it's the latter. So, you know, Mike, I believe that the reality is we have kind of plateaued on the next wave of innovations for a bit. It's going to be a lot of building, you know, kind of breakthroughs like XAI is doing with how to move the data around. It's going to take, you know, it's going to take like another wave of figuring this stuff out. But,
The tech industry can't let that stop, right? The train has to keep moving. Like if you look at the S&P 500, look at the seven companies that make up all of the performance of the S&P 500. Those seven companies almost all have some kind of AI play or almost all AI money. It's the money moving the market right now. If this money train ends, it's devastating.
I mean we'd be looking at massive layoffs. I just saw a story right now about – in fact, we'll cover it – about a story about all the commercial real estate getting bought up because of AI. Like it's a huge deal right now. So I don't think they can afford it. to let the hype train stop. I mean, look at Apple.
They haven't even fully shipped their AI products yet, and they're probably even only at, like, what, a tiny 5% or something of maybe more, probably less of their user base. We're at the very beginning. We are at the very, very beginning. And they're going to push this autocomplete as far as it can go, I think. Do you agree, Mike?
We're just at the beginning, or do you think we're kind of over the hype cycle?
Um... I think we're going to get more practical applications for it, which might be less sexy for the non-techie media.
Yeah, I can see that.
So, I'm trying to think of how to put this correctly. It's going to be actually more useful, but maybe less, you know, Jim Cramer's, well, Jim Cramer will yell about anything, but, you know, CNBC is not going to be freaking out about it as much. I do worry a little that the massive expenditures that particularly OpenAI is taking, I mean, that's a big old nut to crack.
So, I mean, I guess they're going to get there 200 at a time, right? I don't know.
Yeah, what you might see is... Okay, I'm following your lead here, and I'm thinking what you might see is the big companies kind of slow down and the features become more iterative, but you start seeing what has been developed adopted further down the market stack. So maybe companies start putting chat GPD embeds for support instead of those stupid, crappy, live pop-up things that they have.
You start seeing implementations... of this stuff that just you know make tiny small improvements to a small business maybe that small business doesn't have to hire one person something really minor but you magnify it across you know a million small businesses and it starts to actually make a difference i mean i could see that kind of thing it's a good question We'll just have to wait and see.
Thank you, everybody who does boost in. We had 21 of you stream sats, and we stacked a humble 16,939 sats from you sat streamers. Thank you very much. And when you combine that with those of you who boosted in, we had a very humble 164,895 sats, which thank you very much. That gets split between myself, Mr. Dominic, Editor Drew, and the network.
And if you'd like to support the show via a boost, the easiest way is Fountain FM and then load it up with Strike, which is available in like 110 countries. But if you're in Canada, I think Bitcoin Well is probably a better way to go. And there's a whole bunch of solutions at podcastapps.com. I think Fountain is probably the easiest way to get started, but there's a lot for you to choose from.
And thank you, everybody, who boosts. Boost! Now, before we get out of here, this was fascinating. So I talked about this earlier. San Francisco offices were a disaster scenario, like getting discounted down 70% until certain businesses realized that return to office could be a perk, I guess.
There has been somewhat of a demand for in-person work, at least according to CNBC, who says Gen Z professionals are seeking more collaborative in-person work environments, reflecting a shift in attitude after years of remote work. And so they cite a couple examples of these new AI startups that offer four-day in-office schedules. They think that's just a little bit more engaging.
So you still get some remote work, but you do four days in office. Others are saying they're going to prioritize office-first cultures as a strategic choice in hiring. The market demand for this in San Francisco has been kind of a big deal. San Francisco vacancies hit 34.9%, but AI startups like OpenAI and Sierra AI, they're snapping it up.
In fact, they now account for a significant share of the entire real estate market in San Francisco. 62% of AI-related leases in 2023 were subleases, reflecting the trend of sharing office space among multiple startups instead of renting entire floors, too. I guess it's a good price right now. That's probably why.
But I wanted to get your thoughts on a perk and the company focusing this as a perk is like you're going to be in the office. Some of it's like you're going to be here a few days and some of it's like, no, it's 100 percent office. And they think that this is something that Gen Z craves. Young professionals value connection and energy of in-person work, they say. OK.
It's a back to basic shift, they say. Something about it. I don't know. I mean, I guess. Here's what I do think is smart, is if you are going to seriously lean in on being an in-office workplace, you should totally present that way and appeal to candidates that are looking for that. I think that part makes sense. I don't buy the Gen Z in particular for some reason. I don't see why Gen Z...
Would be more into remote work than Gen X or the millennials, right? Or less into remote work, I guess. That part doesn't really... I guess I don't understand why it would be Gen Z that's so into going back into the office or whatever.
It just seems like a strange... I mean, are they implying that because they want to meet people, potential dating, I guess?
No, no, it's collaborative. No, it's that collaborative energy, man. Yeah, okay. They want to have that collaborative in-person... Energy of sitting in a cubicle and then walking past people that you want to avoid most of the day.
I don't know. I don't know. I can't see a case where I want to go back to the office.
Yeah. Well, maybe it's because you're an old man. And, you know, if you were a Gen Z looking at an AI startup, you'd want to go back to the office. It's true.
If I was cool.
Honestly. Honestly. If you're going to be living in California, living in San Francisco, you might as well go into the office. I don't know. To me, it feels like we really are stuck on this stupid – we're really trying to put everything into a box. For me, there are some days I work from home. There are some days I come into the studio.
It kind of just depends on what my workload is, and I don't try to label it as hybrid work or four days in office and two days home or whatever. I don't even try to put it in a box like that. It's just whatever the job needs is what I'll do. That just seems natural.
And corporate America, and especially the media, always wants to just put everything in a category, give everything a label, and then they even want to slice it up by age group and types of work. And I just don't think it actually works out like that. It's just a convenient narrative. But it is interesting that the AI startup companies are taking a lot of that VC money and snapping up
subleases on a bunch of San Francisco real estate. That's what they're doing with that money. Because they're all just plugging into the OpenAI API anyways, right? That's right. Anyways, we'll see. Maybe I'm wrong. Maybe I'm wrong. Boost didn't tell me what I got wrong, but just the way it looks to me. Mr. Dominic, is there anywhere you'd like to send people this week before we scoot?
Go check out alice.dev, and you can find me at Dumanuku on all the socials. I have not been posting much, so it's been great.
All right. We'll forgive you. It's probably better that way. Keep it a little bit of mystery. If they want to know what's going on with Mike, they got to tune into the show. They can't cheat on the show by not listening and just getting on social. I guess I don't post much either, but if you want to experiment with Nostra, you can find me over there, chrislas.com. Also on WeaponX.
And you can email the show. There is a contact page and all that goodness over there. In fact, speaking of that web page, it's got links to all kinds of stuff. Our RSS feed, our live page, our matrix chat room, all the good stuff that's going on all the time.
Plus, you can go to coder.show slash 599er and you can get links to everything we talked about, including links to that weird real estate article. I don't know. Maybe I'm wrong. Maybe Gen Z does all want to work in the office. You tell me. Boost in, because for God's sake, we could use the support and tell me what you think. Are they all just touchy-feely? Is that what's going on?
What's going on over there? You tell me. All right. Thank you so much for listening on this week's episode. See you right back here next week.