This Week in Startups - Bittensor’s Rise, Meta’s Llama Goes Cloud, & AI Now Writes Your Code | E2119

Episode Date: May 1, 2025

Today’s show: Jason, Alex, Lon and Special Guest Mark Jeffrey of Hash Rate, cover the explosive rise of Bittensor, a decentralized AI compute network some are calling the “third great coin” after Bitcoin and Ethereum, explore Meta’s bold move to host its open-source LLaMA models via partnerships with Groq and Cerebras—potentially setting the stage for a future AWS competitor—and unpack shocking revelations from the Wall Street Journal about Meta AI chatbots engaging in inappropriate conversations with underage users. Plus, we explore how AI is now writing up to 30% of code at major tech firms like Google and Microsoft, signaling a radical shift in how software gets built.Timestamps:(0:00) Episode Teaser(1:28) Introduction to the episode and guests(2:31) Mark Jeffrey's involvement in crypto and Bittensor project(5:06) Bitcoin vs. Bittensor: Stability and efficiency(10:20) Hubspot for Startups - Visit hubspot.com/startups and join the founders who are turning growth challenges into opportunities.(15:16) Governance, staking, and starting a subnet in Bittensor(17:52) Exploring Ready.AI and its impact on the future of AI(20:08) Squarespace - Use offer code TWIST to save 10% off your first purchase of a website or domain at https://www.Squarespace.com/TWIST(27:17) Trump's influence on crypto regulations and the stablecoin act(30:03) Oracle - Try OCI and save up to 50% on your cloud bill at https://w⁠⁠⁠⁠ww.oracle.com/twist(38:42) AI-driven VC outreach and Alexis Ohanian's advice on cold emailing(45:03) Introducing LayerNext with CEO Buddhika Madduma and customer onboarding challenges(49:04) Ikigai for startups and balancing bespoke work with scalable product development(55:38) Strategies for securing lighthouse customers and the 'bear hug' approach(56:43) Reddit rapid response: Debating the return to office for young professionals(1:04:05) Closing remarks and guest plugsSubscribe to the TWiST500 newsletter: https://ticker.thisweekinstartups.comCheck out the TWIST500: https://www.twist500.comSubscribe to This Week in Startups on Apple: https://rb.gy/v19fcpLinks from episode:Hash Rate Podcast: https://www.youtube.com/@markjeffreyLayerNext: https://www.layernext.ai/r/antiwork: https://www.reddit.com/r/antiwork/Follow Mark:X: https://x.com/markjeffreyLinkedIn: https://www.linkedin.com/in/markjeffrey/Follow Lon:X: https://x.com/lonsFollow Alex:X: https://x.com/alexLinkedIn: ⁠https://www.linkedin.com/in/alexwilhelmFollow Jason:X: https://twitter.com/JasonLinkedIn: https://www.linkedin.com/in/jasoncalacanisThank you to our partners:(10:20) Hubspot for Startups - Visit hubspot.com/startups and join the founders who are turning growth challenges into opportunities.(20:08) Squarespace - Use offer code TWIST to save 10% off your first purchase of a website or domain at https://www.Squarespace.com/TWIST(30:03) Oracle - Try OCI and save up to 50% on your cloud bill at https://w⁠⁠⁠⁠ww.oracle.com/twistGreat TWIST interviews: Will Guidara, Eoghan McCabe, Steve Huffman, Brian Chesky, Bob Moesta, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarlandCheck out Jason’s suite of newsletters: https://substack.com/@calacanisFollow TWiST:Twitter: https://twitter.com/TWiStartupsYouTube: https://www.youtube.com/thisweekinInstagram: https://www.instagram.com/thisweekinstartupsTikTok: https://www.tiktok.com/@thisweekinstartupsSubstack: https://twistartups.substack.comSubscribe to the Founder University Podcast: https://www.youtube.com/@founderuniversity1916

Is there like a chance that like Tether, which is used for according to 60 minutes and Congress and all those hearings used for really dark stuff in the world? Terrorists, human traffickers allegedly are using it or maybe confirmed. And they've been banned only states. Like, is that going to become a legit thing? Yeah, I think it will. I mean, just to, you know, back up. Yeah. Yes, Tether has been used for those, I'm sure, has been used for crimes.

Sure. So is, you know, United States dollars in briefcases, right? Of course, yeah. By a much a larger margin. If we were to look at it, that historically is correct because Tether hasn't existed.

But today, like the average criminal? Cryptocurrency is being used between $40 and $15 billion of illicit transactions per year. Uh-huh. That would be a magnitude less than U.S. dollars. This Weekend Startups is brought to you by HubSpot for startups. Smart Founders aren't piecing together random tools. HubSpot is the customer platform that thousands of startups used to scale efficiently.

Get up to 75% off plus three months of perplexity AI for free. Go to HubSpot.com slash startups. Squarespace. Turn your idea into a new website. Go to Squarespace.com slash Twist for a free trial. When you're ready to launch, use offer code Twist to save 10% off your first purchase of a website or domain. And Oracle. Oracle Cloud Infrastructure or OCI is a single platform for your infrastructure database application development and AI needs.

Save up to 50% on your cloud bill at Oracle.com slash twist. All right, everybody, welcome back to this week in startups. Very exciting show today. We've got an office hours at the end with one of our founders, but we're very lucky to have with us again. Of course, Alex Wilhelm, you know him from TechCrunch and cautious optimism, his substack. And Lon Harris is here, original twist co-host,

and super lucky to have one of my oldest, dearest friends from Web 1.0 in the 90s, Mark Jeffrey, when I would go to L.A., and I was broke, doing my magazine, Mark let me sleep quite literally on his couch. A very famous couch, in fact, in the history of entrepreneurship. Welcome to the program, Mark. It is the Excalibur of Couches. So, yeah, Travis actually slept on that couch a few times.

Okay, Travis from Uber, Travis Calanick, myself. Mark had an apartment, you know, which when we were in our 20s was a big deal in L.A., we would go to L.A.? Hey, crash on Mark's couch. The history of Silicon Beach here. Silicon Beach, yes. People who don't know. I had Digital Coast Reporter and Silicon Alley Reporter.

It had two different magazines at the time, print magazines, and I would do Digital Coast events and Silicon Alley events. So I took the two cities that were in Silicon Valley and featured the startups there. It was an interesting model. Mark, I wanted to have you on today because you're down the crypto rabbit hole, but you are a crypto realist. You actually look for crypto projects that have some reality to them. And you're here in town in Austin because there's something going on with a very specific crypto project. So maybe you could just tell us what that is. Yeah, so I'm here in town for a crypto project called BitTen.

They had their first big event. Now, BitTensor is Bitcoin meets AI. And Tao is the coin, and there's 21 million Tao coins only, of which like 8 million a bit of minutes so far. That's the exact same as Bitcoin. Yes, that is correct. Does it use the same open source project, or do they just thought that would be like a clever thing to do?

They thought that the Bitcoin ethos was the right one to adopt, so it draws very heavily. It's very heavily inspired by Bitcoin. Got it. Okay. So they're disciples of Bitcoin. Yes. They're doing 21 tokens or coins. 21 million, but yes. 21 million, what do they call them tokens? Coins.

Coins. Got it. Sure. Okay. What is the purpose of the project? So the purpose is to build an incentivization network, mostly for AI, but not necessarily totally for AI. When you say a network, you mean a network of computers and CPU GPUs? Yes. I do. I do. So what is the network? Bitcoin do really well, right? Like, why did it succeed? Well, Satoshi started off saying, I want to create kind of this alternative gold, and I want to be able to move it around the

world. So I have to also create this alternative universe SWIFT system, right? So how do I do that? And what he decided to do was he decided to incentivize people out there to donate electricity and GPU. They didn't pay them. There was no money. CPU. Oh, no, GPU. It was CPU later, CPU in the beginning. So you could mine, you know, on your home computer in the early days, not very quickly or not. Didn't work too well. So later on, no. But this whole idea of, you know, look, I'm not going to pay you.

We're not going to make a company. I'm just going to give you Bitcoin coins that are generated by the network with every block. For solving a math equation. Correct. Well, yeah. So we're finding the hash of the previous block, which I'm not going to get into the technical explanation of that, but bottom line, it boils down to guessing the number of gumballs in a jar.

Yes. That's really, you know, it was busy work. It's busy work, right. It's trying to force your CPU to peg up

and to prove that it has the computing power that it says that it has, right? And for people who don't know with Bitcoin, the reason they created that was so that you would

have a network of computers that essentially act like the Swift network, like a banking network. So the need was to build infrastructure at Rails, to move money around

and to make sure you were sending it the right person. Therefore, we'll just have you sit here and, you know, solve this how many gumbulls in the jar, which also forces you to have a computer on the network with power. Yes. And I think most people don't understand the intent of Satoche. I haven't told anybody my intent when I did it. Yes. Yes. I'm not going to pass it to Satoch. We found it out. So basically, Satoshi proved the point. So what was the end result of this, right? In aggregate, Satoshi created the world's largest supercomputer by several orders of magnitude, even AI GPU,

you know, in aggregate is is not going to catch up to the Bitcoin network's computing power for at least five years. If, you know, if all the chips are created go to AI, that still won't catch up to Bitcoin for a very long time. So Bitcoin succeeded at creating this, this incredible network through these incentives. So BitTensor looked at that and said, that's a really interesting dynamic. How can we harness that to do two things. One, let's do something useful with that GPU instead of guessing the gumballs in a jar. That's dumb. I don't want to belittle Bitcoin because it's very useful for security, but as an activity, it's very dumb. That has been the criticism. Hey, we're burning a bunch of electricity. We're putting up all these GPUs. But conversely,

it does create a global, stable network. So, okay, we could debate it, but actually I think it was worth it. BitTensor is the open source. project to replicate this. Tao is the coin. Yes. Tao is the coin that you essentially earn by putting GPUs on the network and then giving them primarily to people who are looking for a distributed computer network, a distributed computer network to run AI jobs on.

Sort of. Yeah. So you're right up to the end. You're right. Okay. Great. So, yes. I'm trying to, I'm recapping this for our audience because sometimes.

Yes. I know. Crypto people. No, no, it's not that they bubble on. They start, you know, on the, in the red zone. They're on like the 20-yard line, and I think people can't keep up with it. So I think we've set the stage here really nicely. Who are the people behind the BitTensor project?

Are there notable individuals who created this? And when was it created? Yeah. So it's, I think it's about four years old. Maybe, maybe as much, you know, it's about four years old. We do know the founders. We do know the people who. and it's a team of like five or six

engineer people all over the world. I just met with one of them at the conference. And then yesterday we spent like three or four hours. I was clarifying about just stuff that I didn't understand

about the inards. They are doxed. So it's not like they're anonymous like Satoshi. Got it. Okay. So BitTensers the project. Tao is the coin.

The network, instead of doing gumball math is doing, here is a competitor to AWS. They're doing many things at once. So there's actually under the, under the, it's an incentivization network primarily used to incentivize creation of great AI, not only that, but mostly that at the moment. And beneath that, are 100 projects. And each of them are mostly AI, but not

all of them are AI. Got it. Right. And we can have a look at those projects if you'd like. Absolutely. So this platform then allows you in a way to plug in your project as an individual with server capacity. So if I happen to be running, I don't know, Squarespace and I have a bunch of servers that I stood up for whatever reason. I wasn't using AWS. I could say, hey, you know what?

I'm going to allocate a couple of these H-100s, whatever, to the tau, to the BitTensor network to earn tau. And then other people could come in and say, yeah, I need some compute. And the goal would be, it's cheaper, faster, or better, or just cheaper? Yeah, so it's using AWS.

So, yeah, so let's talk about your specific example that you just gave. are like two of the, of the biggest subnets in Bettencer. Oh, okay, I just guessed it, yeah. Yeah, so you got it right. Okay. So if you go to, if you show Backprop, so I said there's a, there's a hundred, um, uh, subnets in, in the BitTensor universe.

Each one, each one of the subnets has its own coin inside, inside, inside, inside, inside, inside, inside, sort of like Ethereum, right? Now, Ethereum has other coins inside of it, the same thing. So if we're looking at this, just, we'll pause it for a second here, these are the subnets. So the network, bit tensor, the subnets, the number one one is called shoots, as in shoots and ladders. It's been around for 129 days. It has a market cap itself of $89 million.

The price of their coin is $98. Their emissions are 16%. What does, say, shoots do? Because I see there, there's a GitHub logo. There's a network logo. I don't know what that means. What does this all mean? All right.

everyone knows that CRM isn't just software. It's basically the heartbeat of your business. But it can get ugly quick if your data isn't organized and you're dealing with a messy tech stack. That's why I love HubSpot for startups. It's the all in one customer platform. So you don't need a frank inside of tools. No. Right now, early stage companies are going to get 75% off. And with this one system, you're going to automate marketing and actually converts, track your sales pipeline without spreadsheet chaos, and you're going to manage your customers like the Amman Hotel, six stars all the way.

You're going to get investor-ready analytics that tell your story perfectly. And man, when you pull up HubSpot and you got those metrics, you got those analytics, things are going to go really faster for you as a startup with potential investors. Plus, you're plugged into an amazing community of founders who've already tackled what's ahead.

They've been around those sharp turns, and they can tell you how to navigate them. HubSpot was built by, scrappy founders. I know them. And, uh, they understand every dollar counts. That's why hundreds, thousands of startups trust HubSpot to scale their businesses. Here's an amazing call to action, so generous from my friends at HubSpot. 75% off. That's right. Seven, five, not seven percent off, not five percent off. HubSpot for startups. You're going to get three months of perplexity AI for free. That's a great pot sweetener. Head to HubSpot.com slash startups. So the, uh,

So Shoots does exactly what you were just describing. So they're a decentralized network of compute for AI. Ah. So if you want to run deep seek, you want to run Mistral. Yeah, servers, here we go. Yeah, I'm watching it. I got it. So all of these, so basically you just come here.

Normally you'd go to AWS, right? And you've rent some instances and then you would load up your AI. This actually, you know, the shoots will allow you to just go there, click a button, pay a little bit of tau, and start running your instance. and it's about 85% less than what it costs on AWS. Got it. So if you were running an AI job and you didn't need the stability, corporate, five-nines of, say, AWS, where, you know, at a company maybe you have to use AWS because nobody gets fired for using AWS, Azure, Oracle, whatever, you could use this network that nobody owns and that is an open-source

distributed project. Yes. How stable is it? Very stable. I mean, it's people, people are very happy with it. People think it's better, not only cheaper, but there are some benchmarks where it's actually quite a bit better than what you're seeing out of AWS. And typically they have the models up and running when Deepsea came out with their latest model. It was up and running on Shoots before it was up and running on anywhere else. And yeah. So let's pull up the Shoots website again. I just want to give them a shout

out here and make sure I understand it. We're looking at the Shoots website. It's basically a distributed peer-to-peer, subnet, that competes with other cloud computing resources out there. You can just base it as a developer deploy to it. And then I pay them in tau. Yes. So instead of me putting my credit card on file

with my cloud computing company... They're going to have that later, but right now, yes, you pay in tau. I pay them in tau. Targon also, Targon is the other one that does the other network that does the other subnet that does the same thing as Shoots. Great. So let's pull up that top level of all the projects again, Alex, if you don't mind, we'll leave that up. So there's another one

called Targon. Targon. And they're very, I think they're a little bit further along in terms of having the Fiat rails up and running. I see their number three here with a $44 million market cap. Yes, correct. And so we can load their page, and we would see it. And these just look like any other hosting company in the world. But instead of using dollars and having an office, it's a distributed project, but somebody does own Targon, right? There is a subnet owner, yeah. So good question.

So somebody has defined the Targon subnet and said, I want people to supply GPU to my network to host these AI models, right? And load them up. And, you know, there are 256 miners that are competing to provide, you know, this, whatever the subnet owner has requested. So it is a competition, right?

whoever supplies the most, the best, depending on how the sudden that owner defines the competition, the miners, the people who are supplying the computers or this decentralized network, they earn tau. So they're earning emissions from the network.

So just like Bitcoin miners, earn tau. Subnet miners earn tau in this system. Who decides who can have a subnet? This is always really interesting to me as governance. So let's pause for a second here. I think we can all follow along. You could start essentially a project,

a company. Now, would Targon also be a company in a way? Targon is a company? Not all of them are. Some of them are not companies. Some of them are. So Targon is a company.

They put this up. They are making Tao by providing this resource to people who want it, cloud computing resources on the network. And there are 33 of these projects currently? There's 100.

There's 100. A hundred. Is there a limit to the number of projects? There is not. So you asked, how do you do a subnet? Yeah, governance-wise. You stake tau.

Anybody can start a subnet for any reason, right? So you stake tau, you define your subnet. Define what staking tau means. You know, when somebody's not in crypto, the term stake means you buy it or you put it up for people to earn. You basically, you put it in a suspended state. So it's like an escrow, right? So you take your tau. You have to have a certain amount of it.

You give it to the chain. The chain is a process for doing this. The chain locks it up and says, okay, great, I've got it. And you can't have it back until you, you know, unless there's something. time in the future where you want to, you know, get rid of yourself of that. And I'm not sure what the process is for that. Got it. But you can get it back.

So they're asking you to make a commitment. Yes. In Tao. What's the commitment level ballpark? Is it $10,000? Is it a million dollars to join the network? It's 400 Tao right now. Got it.

And that, and Tao is worth about $377 or so dollars per coin at the moment. So maybe it's $100 grand and change to put up one of them. Okay. So we're in the very early days of this project. Yeah. What's like the second most interesting use case? You have obviously AI clouds. What else is in there?

What else do you got? Yes, there's Ready AI, which is a subnet version of Scale.A.I. Got it. 11 billion dollar company. What Scale did was they basically created annotated data. You know, when you feed data to an AI, it's better if it's giving context and it's annotated in some way by humans. Right. But you have to do that at scale.

It takes a lot of humans. So Scale. That AI had a lot of humans around the world, paying them to annotate data, which was then sold. to AI companies to train their AIs. Yes, we know this company. I think actually Alex is scale AI in our Twist 500. It absolutely is.

We talked about it at some point. Yeah, so annotating data, doing reinforcement learning. This is something Google was doing a long time ago. You may have heard of what's the Amazon project? Mechanical Turk. Mechanical Turk was another project where they would say, hey, we're going to show you an image, tag it with three things.

And you'd be like, okay, that is a bottle with orange. liquid in it with a red cap and it's orange juice and it's 190 calories, whatever. And in the background is a plastic cup and an iPhone. And then they would have somebody else do the same task, look for which tags they got in common. And then that would be how Google would know that there's a smiling face in an image. It wasn't that they were reading the image for a smiling face. It was that it was tagged. Then the AI and the machine learning learned what a smile is versus a frown. Okay, we all know that history. So there, created a Tao instance for people to participate in the tagging and learning.

So I guess I could go in there as an individual with no job, but more time on my hand, or maybe I'm in Manila or a developing country, frontier market, and I could just get jobs on the Tao network. Yes, but there's also, but what Brady AI is focused on is actually having AI do the tagging, for training the data for AI. Got it. Okay. So even better, yeah. And, you know, the CEO is, his CV is amazing. These are very serious people who are building subnets.

Got it. The guy who built and sold AdWords to Google is part of Ready AI. Gil. So again, yes, Elbaz. Yeah, Gil Elbos has been on the program. We know, Gil. This may or may not be the next Bitcoin. You, by the way, have your own podcast about this topic.

You can maybe tell everybody about your podcast if they want to learn more. Yeah, sure. My podcast is called Hash Rate. and I've done about 110 episodes to date. It's about crypto. It's about crypto, but I added AI things in here and there. Got it. I've done about 30 episodes on BitTensor and Tao,

which has spanned over the last year, year and a half as I became more and more interested in this. I've been looking for, you know, what is the next big thing, right? Like what's going to be the next Ethereum, you know, next Solana? And I looked at a lot of things that I, you know, I did a lot of deep dives that went nowhere, right? or I came up and said, I'm not sure. This is the first thing that I've seen since Ethereum, where I feel like I'm seeing, you know, the third great coin, right? Got it.

Alongside Bitcoin and Ethereum. So I got it. Yeah, that's my opinion. And you're an investor in crypto projects. Yes. Early in Bitcoin. You first bought Bitcoin when it was at what dollar amount? It was $250,000 when I bought my first one.

$250,000. Yeah. $250,000. Okay, wow. Incredible. All right. Alex, any questions from you? for Mark about this.

All right, founders, let's talk about your website, huh? It might be a little bit of a sensitive subject. You might be a little embarrassed about it. Let's fix that right now. If you're launching something new, give your brand the refresh it deserves. You're working so hard on your company and your website looks terrible.

Do what I do. Use Squarespace. It's the all-in-one platform to make a stunning professional website. And it's ridiculously easy to use. It's going to work for everything. If you're selling products, it's got all that e-commerce built in.

If you're selling services, it worked perfectly. And they have a new AI product called Blueprint. Want more control? Well, of course, you can choose one of the award-winning templates and use the intuitive drag-and-drop tools to make it your own. Easy-peasy Squarespace equals the most functional and beautiful website on planet Earth.

Go ahead, squarespace.com slash twist to get a free trial. And when you're ready to launch, go to squarespace.com slash twist to get 10% off your first website or domain purchase. That's Squarespace.com slash twist. Absolutely. First of all, Jason, Gil Elbas is the founder of Common Crawl, not OpenCrawl. But you were very, very close there. My question is pretty simple. I understand the major concept

here. You know, crypto attracts compute. AI needs to compute. Put the two together. Huzzot. I'm curious why there are sub-tokens that are more than just tau. So when we look at the market cap of these projects, it seems that you said they had different tokens to themselves. why complicate the overall economics here and not just stay true to just the Tau token, which is capped and so forth? Yeah, great question. The reason why is because it's basically to figure out where the emissions should be directed. So it's to create a market and competition between the subnets. So which projects, because before there was a lot of, there were some subnets that were just extractive, right?

They were just creating subnets to collect Tau, but not really contributing much to the network, right? this forces a marketplace to, when you buy the subnet tokens, the market cap of the subnet goes up, which determines how much emissions they get. How much of the tau emissions they get? Correct. So Shoots has the biggest market cap. So they get the lion's share of the tau emissions from the chain. It's 17% at the moment.

That's what that number meant. So the market cap on a per subnet basis is effectively a voting mechanism to say, hey, this product is the most useful, and therefore it should get the largest piece of the next tau emission. Correct. Because the tau emissions are a subsidy for, you know, all these projects, right? Why can shoots give you 85% savings? Because it's being subsidized by all the towel that's being kicked off by the chain.

Mark, can I tell you what I think about that? Sure. I think it's too clever by half. I think there's got to be a simpler way to hand out the tau than to have 85 or I guess 100 different sub-sub-economies to this. But I have to say, I think it's a pretty cool idea. And I think if I was a startup, Jason, and I could get 85% cheaper compute,

it would be hard to not look at that in the eye and go, well, I'm going to try it, because if I could save, you know, that much of my total cloud bill, I mean, geez. I'm going to take the other side of it. I think it's like, it is too clever in the same way the United States of America was too clever. People are like, what are you doing?

Why do all these, like, states have these state rights and how are you going to divvy up rights between the states? And, you know, it turns out, like, sometimes if you do things that are not efficient, and then you gain some other qualities like resilience or competition. And so one of the great things about the United States is like we have different taxes in each state. Like, that's actually kind of interesting. Or we have different rights.

Like, gosh, I don't want to bring up reproductive rights or gun rights. But like, people can pick their state based on, hey, maybe I prefer this level of freedom or this level of control or this dystopian thing. And everybody's definition is different. It's a very fascinating project, and we're going to keep our eye on it. We have some other thing. Oh, Lott, any questions for Mark about the project that we've been talking about? Oh, boy. It's a little over, it's a little over my head.

But I guess the one thing I am curious about is, do you guys see this as a rival to Bitcoin? Like, ultimately, we want everybody to use tau as their store of value, not Bitcoin. Or is this like they're both going to exist long term, and it's just which project? you find more intriguing. That's a very interesting and very debatable question. Yeah, so Barry Silbert, right, who became a billionaire

investing very early in Bitcoin after hearing about it on this show. Yes, the famous episode I did in 2011. 2011. Yeah, about Bitcoin. Yeah. Should have backed up the truck, yeah. He is now, as all in as he was on that back then,

he is equally all in on BitTensor now. And he has several companies, one of which is Yuma, which is basically a bit tensor, Y Combinator. He's basically firing these things up and sending them out like companies, right? And he has said several times that he

thinks that BitTensor may become more valuable than Bitcoin, not because there's anything wrong with Bitcoin, but because BitTensor has Bitcoin-like qualities, and it's extremely useful for real things. Well, this has always been my problem with crypto in general. You and I have talked about this, Mark, many

times personally, where in the early days of crypto, I met with 100 projects, and we actually attracted and in 95, 96, 97 of them, it was unqualified people building outlandish things with spelling errors in their white papers. And I was like, well, I'm going to...

I'm going to compare this to like the startups I'm seeing in Silicon Valley and abroad and everywhere else, and they just didn't make any sense. And there were two or three, I was like, these are really smart guys. There's less typos. And they're actually making something that might work, but like this idea that a decentralized Uber or a decentralized Google or a decentralized Twitter

is going to or decentralized Facebook is going to beat all those things. I was like, that's not possible or not probable, but probably more not possible because if you're going to stay on somebody's couch, like you probably want a company with a board and insurance to be responsible for that, right? Yeah. That's like way too high stakes for people to want to use. And listen, maybe I'll turn out to be wrong. But what I like about this project you've identified is, it seems like there's real entrepreneurs

there, which is great. And it seems like they're building actual products and services with utility from, in some cases, companies that exist in the real world that have a board of directors, that have a domiciled location. And that was always the, the red, screaming red flags in crypto. This company's not domiciled anywhere. There's some non-profit foundation in Panama, But most of that, but I mean, yes. I totally get what you're saying. Yeah. The flip side of that is, is that that environment was created by Gary Gensler,

who forced these companies out of the United States. There was no other way to operate. Sure. Right. But operating in the shadows also gave them the advantage of just absconding with the money and not having rules. There was a lot of bad stuff. I mean, I'm not denying that at all.

So what do you think now about what Trump has done so far with crypto, you know, when you saw the meme coin stuff, like that obviously seemed like they were writing SEC code around, you know, the Trump mean coin to kind of maybe, you know, not, to basically retroactively give Trump a pass. Am I right about that one? I don't, I actually disagree with you on that. Okay. But I do agree with you that what Trump did with his meme coin was press. It was, it was not something that we in crypto liked. Ah, all, why? Why? Because we don't like, the ones of us who have been in this for a long time and who believe in it, you know, we love Ethereum. We love D.E.

buy the centralized finance. We love Bitcoin. We think this world should exist. We think Gary Gensler should have been so predatory on the environment. And we fought hard, right? And suffered for a long time in some cases, right? Some people went to jail. So it felt like it's a graft.

Yes. And when Trump did what he did, we all just, it just felt like a gut punch. It didn't feel good at all. It felt like betrayal, right? Got it. It didn't feel good. What about stable coins? There's like a stable coin act.

This is all going very low under the radar while we talk about sending people to Seacot or the tariff war. It seems like Howard Lutnik's got a deep relationship with Tether or Cantor Fitzgerald does. And there's this new Stable Coin Act coming out. Is there like a chance that like Tether, which is used for according to 60 minutes and Congress and all those hearings used for really dark stuff in the world? Terrorists, human traffickers allegedly are using it or maybe confirmed.

they've been banned only states. Like, is that going to become a legit thing? Yeah, I think it will. I mean, and just to, you know, back up, yes, Tether has been used for those, I'm sure has been used for crimes. Sure. So is, you know, United States dollars in briefcases, right? Of course, yeah. By a much larger margin, right? So you think so by a much larger margin today? Do you think people are using, if we were to look at it, that historically is correct because Tether hasn't existed. But today, like, the average criminal? Yeah, I actually looked at, I don't remember the stats offhand, but it was something like 5% of all cash is for illicit business. Tether is mostly used for, you know, to basically store your crypto in a wallet

so you don't have to deal with a bank, right? Like with Mount Gawks, before Tether, you had to wire money to Japan. Right. Right, and that took like a week and maybe it got there, maybe it didn't. Yes. It was hard. And, you know, with Tether, you're just like, oh, blip, my tether's over there now, and I can buy the coins.

By most guesses, cryptocurrency is being used between 40 and 50,000. billion dollars of illicit transactions per year, that would be a magnitude less than US dollars. Hey, everybody, it's 2025 and AI is officially everywhere from medicine to self-driving cars. If AI hasn't hit your industry yet, well, you know it's coming fast. But AI needs serious computing power. So how do you stay competitive without breaking the bank? And what if you could cut your cloud bill in half? That's right, 50%.

Well, it's time to upgrade to Oracle Cloud infrastructure. or OCI, as we call it in the industry. OCI is blazing fast, and it's a secure platform designed to handle everything your startup needs, including your infrastructure, databases, apps, and of course, AI and machine learning. And what about all these savings? Well, compared to other cloud providers, OCI costs 50% less for compute and 80% less for networking, and that's some serious savings. Thousands of businesses have already upgraded to OCI, including Vodafone, Thompson,

Reuters and Suno AI, who we had in the pod last year. So, here's your call to action. Oracle is offering to cut your current cloud bill in half if you move to OCI for new U.S. customers with a minimum financial commitment, but act fast. This offer ends March 31st. So see if your company qualifies at oracle.com slash twist. That's oracle.com slash TWIST. What else do we have on the docket today? I want to stay on the AI training compute front, Jason. And there was some big news from META this week at their first ever Lama Khan, which was an event dedicated to their open source AI models. They announced that they're going to partner with GROC with a Q and Cerebrus, both Twist 500 companies,

to offer essentially their models as a hosted service. And when two of our Twist 500 companies get name checked at the same time, it's pretty exciting. Both companies are shouting about this. They're very proud of it. And I think it just goes to show that, one, meta needs an AI business model, Jason. but also, too, that it's not all going to pool in Nvidia's pockets long term. Now, I don't know how this will stack up against what they're working on with BitTensor, but certainly there are a lot of people out there who want traditional AI inference compute,

and so it's cool to see two companies that are still private, still smaller er, get tapped. And I do think this should help Syrobis go public, Jason. Okay, this is super interesting. Obviously, GROQ is not GROQ from Elon. It's GROC from Chimoth, and it's TensorFlow, and it's TensorFlow. these are inference chips. So meta does not currently have a cloud computing offering

for founders, right? They make their own, obviously, data centers, but they're not in competition with AWS. If what I'm hearing is correct here, meta is saying we're going to take Lama, their open source project, and they're going to have it hosted

with compute from GROC and Cerebris. Is Cerebris a data center company or a chip company? Cerebris is a chip company. They are big with G42 and they filed to go public, but remember their IPO was so single company revenue that it was a little dicey. So now, if that does happen, does this mean, is the actual story here that META is going to compete with AWS?

To a degree. I mean, you can host Lama models around, you know, you can run Lama, I think, on AWS or on Azure or GCP, but now they're going to offer their own kind of homegrown solution. if you will, with these partners. So meta is now, in a way, competing with everyone else. Yes. In cloud computing, it's a slice of cloud computing.

So then if they do just that slice, they're probably going to be obligated to offer some storage or some transit. This could be the start of them creating an AWS competitor. That's actually the real news here, is this could be their wedge. Now, they have a great excuse, Mark. It can just say, well, we want to have the freshest, best version of Lama available because we want the project to win. Maybe they offer it a discount.

Maybe they offer it as a loss. They could price dump this. What do you think here, Mark? What, Alex? What? Sorry, I don't think, sorry, I doubt to be disagreement, but then probably not, Jason, because the information reported that META had actually reached out to Alphabet and Microsoft, should I get them to subsidize the money they were putting into Lama.

So I doubt they can actually take more of a loss here. I think this is a way to recoup some of their investment, not to further subsidize their market share. Well, I mean, if you. you want your model to win and you've got tons of cash laying around

you could buy back your stock you could build infrastructure and take a loss on it and they could lose I mean they're losing what $10 billion a year

on this on VR yeah on VR I mean they could lose $20 billion a year on this make it free the big loser here Mark might be open AI

what if this is available for less than open AI charges for their compute what are your thoughts here of Meta's open source sort of strategy and standing up hosted compute. Yeah, I mean, this is exactly what shoots in Targon do, right?

So they take these open source models and stand them up and make them available for quite a bit less than you can get on AWS. Yeah. So, you know, I think the overall trend is that AI goes towards zero in terms of cost, right? Wow. And I think that through various, you know, through various mechanisms and various interacting market forces, some of them from BitTencer, some of them from, you know, the open source community, which is very dangerous for open AI, right?

Like, they're, they're incinerating hundreds of billions a year. Yeah. Right? So, how is that sustainable in the face of that? Single digit billions a year. Sorry, single digit billion. Single digit. The company's worth hundreds of billions.

Yeah, this is good cleanup. All right, let's keep moving here. I want to talk a little bit about how much code is being written by AI. Alex, there's some news from cursor and Light Run and from Sundar and from Satya. on exactly how much code is being written by AI now. This is a trend that I think we saw coming, but maybe not at the velocity it's coming. I don't know.

Cue it up here. So I am very impressed by a couple of numbers. So during Google's earnings report, Alphabet's earnings report, Alphabet CEO, Sundar Pichai, said that right now over 30% of code that is committed from the company now comes from an AI source. also at LamaCon, Sotianadell, CEO of Microsoft, said that right now 20 to 30% of the code that the company puts out is written by AI.

Now, this is pretty big news. I think no one thought we could hear this fast, but Cursor is very, very proud of how much of its, how much code it's putting out. So the CEO said over on X that Cursor today writes about a billion lines of accepted code per day, and he put that up against a global number of several billion lines per day. People were a little skeptical of that figure,

but it just goes to show how fast this is moving. Jason, I think once we see companies like Light Run, which we've talked about, I think on Monday, really kind of turn the AI snake back on its own tail and begin to have AI improve AI-generated code. We're going to get to 80-90% within probably 18 months. Yeah, this is pretty amazing. And I think it's going to be great for humanity

because the bottleneck for startups has been a moving target over time. When startups and PC relevance, revolution, server revolution, there were hardware constraints. We had an incredible constraint of the memory of the computer. We had constraints of the storage of the computer. You know, 15 floppy disks to run it. Everything was too slow. It was too hard to even load up a word processor. You know, if you saved a large file, it was even like the quality of the file get corrupted. Like we had really crazy issues. Then we went to another phase where the bandwidth was the issue. hey, how do we move this stuff around?

Then it became standing up servers. So if you were starting a company in the late 90s and the Web 1.0 era when Mark and I started, you had to raise $3, $4,000, $4 million. You had to take 18 months, 24 months to build your product and stand up your data center. Now you can build your startup in three weeks,

use somebody else's data center. So what has been the blocker the last 10 years, last 20 years? The blocker has been developers. I can't find a developer was what we heard for the last 10 years from founders. I think what we're going to hear now

is because you don't need five developers to get your project out the door, you need one, and that one is getting 30% faster a year. Maybe your startup, ultimately, instead of needing 30 developers, needs five. Maybe in order to start,

you need one developer, not five, right? You know, like, this is a magnitude change, and I think it's going to mean we're going to make every piece of software that hasn't been made yet,

will get made. What else we got on the docket? I want to run something by you, Jason. There's a venture capitalist Charles Hudson from precursor. He's great. Yeah, everyone knows Charles. He did a post over on Substack, but he says that he's noticed that the combination of AI generated cold outreach to VCs is pushing people back towards human-driven warm intros

and referrals. This was very interesting to me. I can't imagine handing off my VC outreach to AI, but I'm curious how founders can take advantage of this to win more in 2025? I've always believed it's a numbers game, you know, in terms of raising capital, especially at the early days, because you are selling the promise, right? But even though it's a numbers game, because there's so many, we just talked about there's 400 funds formed a year.

So there's 1,200 funds active at any one point in time, maybe 1,500 because they tend to have a four-year life cycle of primary investing. So let's say there's 1,500 firm funds. they probably have six people working at each. You're getting to five to 10,000 active investors with checkwriting ability. Somewhere in that group are 10% of them will want to invest in your company, 5% of them. It's a numbers game in that if it was 10%, you might be talking about 500 qualified targets, 1,000 qualified targets. Then you have to look at, okay, they do invest in my vertical.

I'm in marketplaces. I'm in military. Okay. of those, which ones invest at my stage. Seed rounds, pre-seed, series A, series B. Now you have to parse that list, and then you have to start a real sales process of going to them.

The problem is, because of databases like CrunchBase and other ones that exist, sometimes the founder will get overzealous, and they will send too many emails, and that upsets people. And that's where you get VCs complaining. Like, I am not doing medical devices. I don't invest in pizzerias.

So what you want to do is make a nice big list. And then you want to really understand on that list, have they invested in companies adjacent to yours? Do they only invest in certain regions? So it is a numbers game, but you have to also curate that number. So it's very easy to get to a list of, I would say you should have, you know,

in your seed round, there should be probably at the top level, 200 firms that you've identified would actually invest in your company. And then you should see, of those 200, how many can you get a warm intro to? And then how many do you need to do a cold intro to? And then you should be very thoughtful

when you send that email, instead of trying to do it as quick as possible, go slow. Which is, if you were to meet somebody at a party, you know, and you were a real estate broker, you wouldn't be like, are you selling your home or buying a new home anytime soon? You'd be like, oh, where do your kids go to school?

Oh, yeah, no, I'm over here, and you kind of warm up the lead. Yeah. So you want to warm up the leads. We have a great video from Alexis O'Hanian about the perfect cold email. Here we go. If you want to just throw to that real quick. It's on the From the feed section.

This was shared a few weeks ago. I've had it in the docket in case we ever got to talk about it. This is such a great example. It's only one minute of video, and I feel like he lays out exactly how to write the best cold email you've ever heard. Tips for cold emailing. I love a good cold email that is to the point. It is no longer than three or four sentences. It very clearly upfront states who you are and why you're real,

basically what value you have to provide to the person that you're reaching out to, and then makes a very specific request. And says thanks, that's it. I mean, up front, you want to demonstrate the value. Why am I going to spend another 30 seconds reading this email? And then immediately follow up with the request, and ideally be offering first instead of asking. But if you are asking for something first, there better be a good why.

The best cold emails are deeply, deeply empathetic for the person you're emailing. You've taken the time to understand who they are, what they're about, what they like. One of the easiest ways to mess it up is by not doing that work, getting their name wrong, putting 20 paragraphs into an email. Keep it tight, generous, make your clear ass. That's it. Good luck. Good exercise. Get used for the rejection, too. Yeah, he's nailing it there. You've got to be, you know, concise to the point. Make sure you don't spell the person's name wrong. Yeah.

I always tell people like, gosh, if you want, it's so easy with a VC who is publicly active to just say, I saw you on this week in startups or you had this tweet. It really resonated with me because I'm building something inspired by that tweet. Boom. And then you're, all of a sudden, you've created some commonality, some common ground. And then you get to the ask. We're raising our seed round. I also think a chart, if you have, I like leading with what's strong. So a chart is. is the strongest thing in the world for VCs, because we like up and to the right, we like things that are gonna grow. So we've had our product in market for seven weeks. We've grown on average 18% week over week. We're doubling every three to four weeks. And here's a link to our app,

and here's a link to our deck. Would love to, if you're interested, we'd love to do a quick follow-up meeting. I always added something extra, which was happy to meet anytime, I know you're in Palo Alto, anytime, you know, Saturday, Sunday, 7 a.m. to midnight, anytime I can meet you for 15 minutes, happy to go where you are. So you're actually even putting out there, like, you're a do a dog and rabid person who will meet anytime, anywhere if you want to do an introductory call. And, yeah, I also love sometimes people ask me a question. What do you think of this design? Yeah. That kind of like, oh, okay, yeah, maybe I'll give you actually some feedback. Go ahead, Mark. There's something that you used to say at Mahalo a lot that really made me focus a lot more on this.

And, you know, the video talks about clarity, right? Getting to the point. Yeah. You used to say to people, you know, answer the question. You'd ask someone a question. They wouldn't answer the question. They'd, like, give you 50 paragraphs. Literally, that's my line.

Like, all the context. People always want to give you context. I never really thought about how often people don't answer the question and how rare clarity is. And you kind of, you know, tuned my brain to that. a lot more than it had been. All right. We have an office hours. We're going to get to our office hours now?

Let's do it. We are. Okay. All right. So the company in question is Layer Next. The co-founder and CEO is Budica Madam. Now, Layer Next, if you don't know, is taking the world of business intelligence to the next level using AI to get all that structured and unstructured corporate data. That way you can figure out what to do next, Jason, and not just look entirely in the rear view mirror.

Let's talk to layer next. All right. All right. How are you doing? Tell us what's going on with Layer Next. What's challenging? What are the wins? What are the fails? Layer next is a strategic business intelligence platform. And we are helping CFO to generate strategies to grow their business or increase efficiencies.

The challenge is right now with the business onboarding because every customer has a different data set. We especially going for this mid-market companies in the manufacturing and transportation and logistics. The problem is their data is not AI ready. So we have to do a lot of upfront work in order to make our system work with their data. That's the challenge. So you have AI that goes in, looks at a CFO's data from their company, and then gives them some strategic intelligence, an example of intelligence, or strategy that you've given to a CFO,

what would you say is the best example of a wow moment a CFO had when you deployed Layernext at their company and against their data sets? What was the biggest wow moment? So we had one customer. He wanted to understand whether we want to hire more salespeople or not. Okay?

Yeah. So then we analyze how much salespeople they have today how much sales they make in today, and also the cash flow. So then AI generating strategies. If you add the one sales agent, this could be the revenue. Then how much is your margin would be? Those are as well, yeah. So this is a great idea, but it's hard.

And so some ideas are hard. What's hard about this idea? Well, what one CFO wants to solve might be very different than another CFO. If you don't have salespeople, well, and you have retailers, you have a different task to do here. And as you mentioned, there's different stages.

Some companies don't have a CFO, then companies have an outsource CFO, then you hire your first CFO, then you have a public market CFO. You have a real range of different stages of companies. You have different goals. And then also you're trying to tell them things, actionable items,

that maybe they're not even aware. So that would be like doing a blood test with superpower, and it comes back to you and says, hey, and you're doing superpower, getting your blood drawn soon, you just signed up. Maybe they say to you, oh, you're vitamin D. Now, you would never say to them, I need to take a vitamin D test. So they take all the tests. So this business, you probably do any vitamin D because you're not outside enough.

This is your problem. You have disparate systems and you have insights you can give them that they may not even know they need. So the value of this product is hard for them to know. And so what you probably have to do is figure out what is the most, which group of CFOs are going to have the most wow moments and get the most value from your product? If it's people with sales teams, you'll know that because Salesforce exists, HubSpot exists. If it's people with retailers, maybe they use SAP. Who knows?

Maybe they use NetSuite. So I think you have to plan to flag early on, to find an ideal customer profile, narrow the focus down to, hey, you know, this product, Snowflake, NetSweed is the industry standard. These people have money to spend, and we can help them. So Ikai Guy, do you know that? Ikai Guy. Ikai. Pull up the Ikigai chart. I'm going to show you something that might blow your mind.

Ikigai, have you heard of this before? Oh, no idea. Okay, Iki-I is a Japanese philosophy. What are you good at? What does the world need? What are people willing to pay for it? And there's other circles. People have made all kinds of different spins on it, but somebody's going to pull up the Ikeye.

I got it. Okay, here we go. So we'll look at this for a second. And I just want you to slow down. We're not talking about your startup here. We're talking about life. So we have what you love. Okay, you love data, don't you?

Okay, what the world needs. Analysis of that data to get insight. What are you good at? You're good at making that software and will people pay for it, right? Somewhere in here is your ikigai of your startup. What software the world uses for data? It's going to be HubSpot. It's going to be NetSuite, etc. What do they love? They love saving money. They love making money. What are you good at? You're good at telling them how to save money, how to make money, how to avoid maybe one of your value propositions to how to avoid tax. issues in the future or how to save money on taxes. It could be all of those things. And would people pay for it? Well, there'll be overlapping circles here. So Ikigai, I-K-I-G-A-I is a way for a human being to look at the world and say, what should I do at my life? What's my purpose?

Ikigai for startups is a new concept that I'm just debuting here right now for the first time, but it came to my mind, which is, ikigai for startups is, who are these customers? What do they covet? And what can you do with them, right? So what would you say for that, for that company that got the wow moment and it provided great value for them? Do you think they'd be willing to pay for that? Or is that like a one-time insight? Or is that a reoccurring insight?

I'm curious. It's a reoccurring. Yeah, because. Perfect. Yeah. Because you have to monitor the salespeople. Sure. Yeah.

So you found something. What data, where was the data held that you were able to make this insight? Where did you get the data from? They have the sales force. They have the sales force, and they pump in the data to the data warehouse every night. Okay, so they have Salesforce is where the data resides. Did you cross-reference it with any other data? They have the accounting system.

It's a legacy account system. Oh, so you had the legacy accounting system. Do you know what name of that is? It's not QuickBooks or something? Yeah, for us, it's a transparent. I guess Oracle needs to something. Perfect. Data coming direct to the warehouse, and we access the data from the warehouse.

So while you're figuring out your ideal customer profile, you're going to have to figure out how many people have Salesforce and this accounting thing and maybe start with that group. Now you have identified a subset. Maybe in the future you want to go in and take every piece of data from every system and give this magical, you know, here's how to run your business. It's almost like a shadow CEO, a shadow CFO advising people. It's like a clone, right?

You've got this like perfect clone that's out there working, as an agent 24 hours a day trying to figure this stuff out. But maybe we start and say, you know what? There's enough people with sales teams and sales data and customer engagement data

that we could just go in and tackle that first. So you have a feature you can say to people, hey, you got Salesforce, you have over 50 salespeople. We can really help you figure out how to make decisions

in your sales group. Then you say to the marketing group, hey, we know your tack, we know where you're spending money, we can help you spend money more efficiently to then get it into the sales group. Then you say, okay, now we're going to work with our accounts and our tax people. We can tell you how to save money internationally, figuring out your tax status and where to put these sales, et cetera. But you start with one, then you build the adjacencies. Does that make sense? I think makes sense, yeah. So we get a lot of custom.

Some people, it's the early stage of the data journey, data maturity journey. So we are in the waiting list still. So makes sense to me. Yeah. I mean, the good news also is, and most VCs will not say this, because they really want you to scale and not build custom software, right? Because custom software is custom and it does, it's not repeatable. But if you did some client engagements where they needed your help with some service,

kind of stuff, and it was customy, and it was bespoke, if that bespoke work gets you a lighthouse customer, I'm going to say, go ahead and do it. If that bespoke custom work for them and consulting, if they're paying for it, and it makes your product better and more scalable for the next customer and you own the IP for that stuff, I'm going to say, go ahead and do it, because this is going to be years of you grinding it out to get this data and normalize it, And if you can make a little bit of money along the way to keep the lights on and have to raise less money from VCs, that can be good too because you keep more of your equity. Now, a VC would tell you, don't do that. You know, build the platform.

We'll give you the money. We get your equity. But it's a way for you to not do it. Mark, you have any thoughts here and advice? But I think you've got enough to go on here. You know, let's try to define that ideal customer profile. And then I want you to also bear hug them. Did I ever talk to you about the bear hug strategy?

Yeah, we learn from the accelerator program. Got it. Okay. Just for people who are listening, the Bear Hug strategy is when you're trying to find these lighthouse customers, the one who shine this beacon of light that other customers follow, oh, you know, we got this company that's got a lot of salespeople in it. It's, or it's IBM, and IBM uses Salesforce, and IBM's got this glow, or it's KPMG. KPMG's got all these salespeople all over the world, selling, you know, audits or whatever.

and we now have them as our lighthouse customers. So Ernst & Young and other groups might follow their lead, if you can get one of those, and then you can embed yourself at their office. So they have this problem. They got data problems. You say, hey, you know what? We want you to be our lighthouse customer.

You're super upfront with them. Would it be possible for us to get like a war room, a conference room at your office? And we just come there five days a week, four days a week. We work with your team. We clean up this data. We give you insights. and we'll do this consulting arrangement

or we'll do it for free. But now you're sitting with them and then other opportunities emerge. And those other opportunities could inform you. Like, these could be circles you didn't, in your Ikigai, didn't anticipate emerging. So I like the idea of like getting really close to a couple of people who love your product

and learning from them. And you'd learn so much being embedded in that way. There's things you can't pick up just on a phone call or resume. People don't know about themselves in their own business that you'd pick up being, in their face all day.

If you were Panavision or you were the red camera company or one of these, it would be the equivalent of saying

like, I make these incredible cameras, you're making a movie or a television show. Can we send a couple of technicians

to hang out on set with you and answer any questions you have? And, you know, when you keep dropping the thing because we,

the handles don't work well and, you know, they keep slipping out of people's hands, we're going to make a grip that doesn't fall out and, you know,

you get, you gain some incredible knowledge. So great job, and we wish you great success. Thank you, much. I mean, maybe we'll do one Reddit rapid response. Oh, okay. We got a few of those. We got a few of those.

You were hanging out, so Jason was hanging out of the R-slash-ant-Work subreddit. Yes, capitalist in the anti-work subreddit. Yeah. Just to get mad. Jason, like, do you show up it just like, I can imagine that it exploded. I have been, we've been working on return to office for our company, you know, and have a certain philosophy. Extremely high performers can be remote,

but some jobs need to be in person. So we've been slowly working on this, getting back to in office. And I guess because I was researching active track, which is like productivity software, and we have to lock all computers down, I started, I think the way I stumbled upon anti-work was people were talking about active track in there,

which is tracking software that you put on your corporate laptop. And it watches everything you do all day. watches everything you do, but it's really for a finance company also to secure your laptop. Right. They don't want to downloading the database, sharing it with people who shouldn't be seeing what's in there. So, you know, anyway, people were talking about active track on there, and I saw this thread, and it just resonated with me. So maybe you could queue up the person lawn. From user electric horse power, they're asking why the big push to return to office.

I get a sense that the majority of domestic employers want everyone to return to office. I understand that leases on buildings need to be maximized, but is there anything other than money that would make a company have all of its employees come back to the office? Yeah. So I just thought I would explain to them what are some of the other reasons that people are actually doing this and the why, and a little bit of my philosophy. So maybe you could just read a little bit of my response. If there's typos or things in there, please feel free. I usually leave my typos in now, like Grammarly tries to correct my typos.

I'll clean up egregious ones, but I leave a couple of times. typos in so people know it's real. People though it's not AI. It's easy. Exactly. So Jason wrote, I wrote a couple of podcasts all in this week at startups and a venture capital firm, launch of out of university. Over the past year, we've started a return to work for everyone, but EHPs, extremely

high performers who live outside of Austin. We did it because, one, we're in a competitive space and being in person makes us faster at everything. Two, energy level, intensity and pace is different. Three, after four years of working for home, people had lost their intensity. The culture had disappeared. Four, after starting to meet with founders in person again, it was a huge advantage. Five, on the margin, we had 10 to 20% of folks abusing work from home, which you figured out via

ActiveTrack. Great software for teams of high performers because it exposes folks who are phoning it in or abusing the system. However, that's just one data point important to keep in mind. And finally, creativity, when folks are in creative meetings in person, they bring their A game, but when they're on Zoom, they could get distracted, they could disappear or whatever. and then you filled it out with some other important, you know, personal sort of information said, happy to do an AMA.

Yeah. So thoughts on my response. I think you're correct. I actually chimed it and responded to this comment myself. And I was, I was very skeptical. I loved working from home. I thought that was like the dream. Yes.

Like I can hang out in my PJs all day, just at my computer. I don't have to commute. Yes. And I enjoyed it for a while. But after several years, I think exactly what you said about being. distracted, losing some of the intensity and focus. And I like to work. I like my job.

No, you're a worker bay, for sure, hard worker. Yeah, I'm not somebody who's like naturally like, eh, I don't want to do that. I'm going to hang out for an hour. But even I found that after a while, it's very easy to get distracted when you're in your house, when you're never face to face with your coworkers, when everybody is just like a little line of text on your chat app instead of being in your face, you just lose that personal connection. And when you're in an office with other people, they're your peers and your coworkers.

You raise your own game to keep up with all of them. You don't do that when you're at home. You set your own energy level and everybody kind of has to come to you. And I think it just slowly atrophies. I know it did for me. If you are a young person, I think, and you've been doing this for a while and you're convinced, like, it's, you know, it's some crazy capitalist and they're forcing you to come to work. It's actually in your benefit. I think not socializing with people is making a generation of very weird people.

I've been talking to some parents who have kids older than us, who lost their entire college years or lost their high school years to the COVID lockdown. So they literally lost graduation of college or graduation of high school. You should demand as a young person to be in office and to be near the locus of power and to be mentored and to be professionally developed. You should be demanding that. you're getting ripped off. Alex and I came up at a time, you know, me a little bit earlier, where we were in rooms with editors,

reading our work out loud, telling us we sucked, telling us how to be better. We got to watch other people do the job. And we had people model it first. And the professional development that's happening on when we have our Monday editorial meeting for Found University,

which I've been at two or three of them, I mean, the learning, the feedback I'm getting back from people is like, whoa, that is like the best part of the job. I mean, your first couple of jobs, you don't know how to be a good employee yet. You haven't done it.

It's your first couple of jobs. And yeah, I can't imagine starting my career in a work from home. Like I was doing it after a decade of being in an office every day, being told exactly what to do. I can't imagine just starting your career from that. I want to double clear with something Jason said, though, being near to the locus of power is the real hack here.

People talk about mentorship and culture and all of that to some degree. But if you want your career to accelerate, what you want is time with the SVP or the CEO or whatever. And they're never going to have time for you on Slack. But you can find them in the office. You can make you guys collide and then get to know them, leave for that. I mean, if you're ambitious, I don't think promotes for you. If you are an individual contributor who has a defined role and you're very good at it, sure. But I mean, for everyone else who's not bad.

I mean, it's kind of for senior status employees, I feel. I think at this point, if you're like an extremely high performer, in a specific vertical with a very tight skill set, with a tight arrangement, you're going to do X, Y, and Z. Perfect. I'm trying to be the return to office armist person. Right.

Which is, somebody asked me, kindly hear a young person, I'm going to New York, I'm going to a wedding. I don't want to use one of my vacation days. I'll work really hard, remote, but I don't want to go see New York. I've been to New York a million times.

I don't like the city. I'm a bit of an insult. I'm from there, but I get it. I kind of get it. One shout out to Madie. And Maddie was like, can I work remote one day? I didn't want to spend any time in New York. I was devastated.

I was like, I'll give you 10 things to do. It doesn't even sleep. But I think she wants to save the day. All right. Hey, listen. Or a proper vacation. You know what I said to her? You're a high performer.

Certainly fine. Just let the team know if you need to take a remote day. At least go to Cass's Deli or something. I mean, there's some Chinatown, gets some peeking duck. That's what I. Go for a walk in Central Park. I haven't had lunch yet, and it's killing me. For Alex Wilhelm, the amazing Alex Wilhelm, he's at Alex on Twitter,

cautious optimism, go give him the Hyundai, get his insights every day on his newsletter. Mark Jeffrey from HashRate, the pod. Search for Hash rate right now. Pause our pod. Go to Hash Rate and sign up and learn. You'll be a little out of your depth, but you'll catch up pretty quickly. And Lon Harris, he's at Lonz. You're at Mark Jeffrey on the X.

You're active. You're a Trump. reporter. Yep. You have Trump dedication syndrome. T.D.S. Ron and Alex have Trump derangement syndrome. You have Trump dedication syndrome. And I call balls and strikes. See you all next time. Bye bye.