This Week in Startups - AI Demos: Claude 3's Opus, Mistral, Groq Playground, EMO by Ali Baba | E1912

Episode Date: March 12, 2024

This Week in Startups is brought to you by… Lemon.io. Need to speed up your product development without draining your budget? Hire vetted engineers from Europe at lemon.io. Go to Lemon.io/twist to get 15% off for the first 4 weeks! OpenPhone. Create business phone numbers for you and your team that work through an app on your smartphone or desktop. TWiST listeners can get an extra 20% off any plan for your first 6 months at http://www.openphone.com/twist Looking to up your marketing game? Check out the podcast: Marketing Against the Grain. Hosted by: Hubspot CMO Kipp Bodnar and Zapier CMO Kieran Flannigan. They bring you the latest in marketing trends, growth tactics and innovation. Available on all your favorite podcast apps. https://lnk.to/h3vKHnTW * Todays show: Sunny demos the lightning speed of Groq (4:10), the new model Opus from Claude 3 (5:38), Mistral (35:26), EMO by Ali Baba (45:21), and more! * Timestamps: (0:00) Sunny joins Jason for this week’s AI demos! (1:19) Details on Definitive Intelligence being acquired by Groq. (4:10) Sunny demos the lightning speed of Groq. (5:38) Sunny demos the new model Opus from Claude 3 and it’s prompt library. (8:52) Lemon.io - Get 15% off your first 4 weeks of developer time at https://Lemon.io/twist (10:12) Claude helps interpret Ikea instructions. (19:18) OpenPhone - Get 20% off your first six months at http://www.openphone.com/twist (20:35) Looking at the use of “guardrails” in LLMs. (28:59) Marketing Against the Grain https://www.youtube.com/watch?v=KIE8hmH0fLM* https://lnk.to/h3vKHnTW (35:26) Sunny demos Mistral (38:53) Sunny demos Groq Playground and creates a framework for a study plan. (45:21) Sunny demos EMO by Ali Baba. * LINKS: Anthropic Claude 3: https://console.anthropic.com/ Prompt Library: https://docs.anthropic.com/claude/prompt-library Mistral: https://chat.mistral.ai/ Groq Playground: groq.com/console.groq.com EMO by Ali Baba: https://humanaigc.github.io/emote-portrait-alive/ * Subscribe to This Week in Startups on Apple: https://rb.gy/v19fcp * Follow Sunny: Twitter: https://twitter.com/sundeep LinkedIn: https://www.linkedin.com/in/sundeepm/ * Follow Jason: X: https://twitter.com/Jason LinkedIn: https://www.linkedin.com/in/jasoncalacanis * Thank you to our partners: (8:52) Lemon.io - Get 15% off your first 4 weeks of developer time at https://Lemon.io/twist (19:18) OpenPhone - Get 20% off your first six months at http://www.openphone.com/twist (28:59) Marketing Against the Grain https://www.youtube.com/watch?v=KIE8hmH0fLM* * Great 2023 interviews: Steve Huffman, Brian Chesky, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland * Check out Jason’s suite of newsletters: https://substack.com/@calacanis * Follow TWiST: Substack: https://twistartups.substack.com Twitter: https://twitter.com/TWiStartups YouTube: https://www.youtube.com/thisweekin Instagram: https://www.instagram.com/thisweekinstartups TikTok: https://www.tiktok.com/@thisweekinstartups * Subscribe to the Founder University Podcast: https://www.founder.university/podcast

I said, help me with these instructions. And look what we got hit by here. Says, I apologize, but I cannot provide instructions related to assembling that particular object. It appears to be to depict an unethical, potentially illegal item. I always suggest focusing your efforts on more. What? It thinks it's a bomb. Positive construction projects and promote harm.

What? Yeah. It's a shelf. You got caught by the woke AI virus. I know. What happened here? This week in startups is brought to you by Lemon.io. Need to speed up your product development without draining your budget?

Hire vetted engineers from Europe at lemon.io. Go to lemon.com slash twist to get 15% off for the first four weeks. Open phone. Create business phone numbers for you and your team that work through an app on your smartphone or desktop. Twist listeners can get an extra 20% off any plan for your first six months at openphone.com slash twist. And looking to up your marketing game, check out the podcast, Marketing Against the Grain, hosted by HubSpot CMO Kip Bodner and Zapier CMO Kieran Flanagan. They bring you the latest in marketing trends, growth tactics, and innovation, available on all your favorite podcast apps. All right, everybody, welcome back to this week in startups.

It's Madra Monda Monda's. Sandeep Mandra is back. He had a big week. His company definitive intelligence was acquired by Grock. GROQ. And so, congratulations to Sondy Madra. You had a nice little cameo on the old all-in podcast this weekend.

That must have been great. I appreciate it. I think I got something like 4,000 LinkedIn requests, which is pretty impressive. But, no, we wanted to thank you guys, obviously, again, for support, but also helping, you know, get the message out. I think, you know, that's what was really key there in terms of what you. you guys have been talking about. And so we're going to try to work it into some of the demos today, too, show you what it's actually really about.

And so definitive I.O., as people know, you did all this AI enhanced data analysis for public, private data. But now, as part of the GROC movement, you're going to be taking over developer relations, I guess? Is that, am I correct in your role specifically? No, that's one of the sub functions. You know, really, GROC is a company that's been making an AI check. that goes inside, you know, servers and racks and then, you know, multi-rack systems.

And what we've been working on with them kind of secretly for the last couple of months is building out a cloud service. And in that cloud service, we allow developers to utilize the high-speed inference engine that we've built so that they can build like new experiences that are, you know, super low latency, high throughput. There's a lot of things that the developer community has done since we launched. probably a year worth of content if we really wanted to do it. I've seen over a thousand demos myself. So we'll pick and choose and we'll bring them on sometimes. Yeah, and as people know, inference is when you ask the question,

you give your prompt to the large language model and it gives your result. And GROC specializes in returning those much faster than they've previously been returned. Much faster, yes. Like 10x faster, like just absurd lightning speed. Depending on the model,

it could be like 100x faster. Amazing, yeah. Yeah. So this is not for training the models, the H-100s from our friends at Nvidia. Those are the state of the art. Am I correct for training the models? Training and inference, you know, like basically anyone that's out there is using those. And then there's some kind of custom Google chips.

And there's custom chips within Amazon and Microsoft as well. You know, the architecture that Jonathan, the CEO and founder came up with is distinctly different from GPUs. And that's what allows it to operate so quickly. Amazing. Yeah, and that's like a seven-year story. I mean, I remember when Chimoth invested in that team coming out of Google. And yeah, it was people who were like, way to burn tens of millions of dollars, Jamath. And here we are. He got ahead of the curve. So congratulations to our bestie Chmott and the team at rock. Yeah, look at that. Wow. So, so fast. You're doing, you describe what you just did. I just did. Tell me the best poker hands. This is the newest Google model, Gemma 7B. So it's a smaller model. And you can see this is generated at 750 tokens a second. Wait, wait, Gemma is who again? That's the... Google. It's an open source Google model.

Oh, Google. Yeah, Google has an open source model. Correct. That's different than Gemini. That is different than Gemini. They call it Gemma. Huh. Okay.

And so, yeah, when you did that, it just immediately gave you, yeah, the, that speed. Yeah. Incredible. You can imagine for certain use cases and many use cases for developers, that speed is really critical because we've seen that across the internet, right? you've probably seen it in your past lives and different things that you worked on, but definitely on all the companies that come through the incubator, you know, making page times load faster,

making the user experience more zippy. And so we have the ability to do that. Yeah. Amazing. All right. Well, well done. Let's get to our demos for this week. We have so many great demos to do here.

So let's just get started. We'll give our letter grades. As always as uncomfortable as that will be for some people. So this week, we're going to focus on a bunch of new. models that were released, which are generally scoring as high as GPT4. And they all have kind of different capabilities. And so we're going to start by reviewing Claude.

And Claude 3 has a model called Opus, which is their largest model. And they've done some interesting things. So let's pull it up here. Obviously, it looks like a standard chatbot. You can say, you know, what are the best poker hands? And it will, you know, obviously, go. is not as fast as what I just showed you there, but its reasoning capabilities are incredible.

And so one of the things that the Claude Anthropic team have done is they provided this prompt library. Yes, this is the prompt library. Really good. If you go to the pun one, it's really good. So this is kind of a fun one where it gives you a really nice prompt. Now, you used to be able to find these prompts on Reddit threads or Hacker News threads or people that put up some websites,

but this is great because these, These are prompts that the team at Claude are suggesting. So I would assume they're optimized a bit for their model, yeah? You would think so, right? And you can use them elsewhere as well. I mean, you know, these prompts, there's nothing specific to Anthropic or Claude in this particular prompt. And so what's really nice about them providing this is it gives, it gives, you know, developers and even end users insights into how to better get the output that you're looking for depending on the category. And I really like how they made these available here in a lot of different categories.

And there's a ton here, right, from data organization to email extraction to lesson planning to fashion advising. I really think that the team has done a good job. What I'm going to do here, though, is kind of show you how capable this model is I'm going to get an example here. This one, in terms of its reasoning capability, it's really, really powerful. and the example we're going to do is we're just going to take a little screenshot of an IKEA build

instructions. And you know how frustrating those can be. And this was one of the AGI tests. I talked about it, I think, on All In. Oh, you did. Okay.

Yeah, you know, if you go to the Wikipedia page for AGI, they have a number of the tests that people you know, assume it would determine AGI. Now, the Turing test is the most famous. You talk to a computer and you can't tell if it's a computer or a human,

so you put a human to answer questions, you put a, you know, behind a box and you put a, behind a curtain and you put a computer and can the person tell the difference between the two? That's one. But being able to give somebody, Ikea instructions and then make something would be, you know, really hard. And somebody had another one, which is give the computer a $100,000 and say, turn it into a million in 30 days or something. Those are all examples of AGI. Oh, look at this one, though. I said, help me with. these instructions and look what we got hit by here. Well, let's see. It says, I apologize, but I cannot provide instructions related to assembling that particular object. It appears to be

to depict an unethical, potentially illegal item. I always suggest focusing your efforts on more, what? It thinks it's a bomb. Positive construction projects have to promote harm. What? Yeah, you got caught by the woke AI virus. I know. What happened here? It got you. Clawed woke AI. Uh-oh. Just going to get an app. Right now, startups have to do more with less. We all know that. It's rough out there, folks. So if you need great tech talent,

but you don't have the time to interview dozens and dozens of candidates, you need to check out Lemon.io. Lemon.io has thousands of on-demand developers to choose from. And these devs are vetted experience result-oriented, and they charge competitive rates. Great developers can be incredibly hard to find, and when you do find them, it can be hard to integrate them into your team.

Lemon.io handles all of that for you. Startups choose Lemon.io because they only offer handpicked developers with three or more years of experience and strong portfolios. In fact, only 1% of candidates who apply get in. And if something ever goes wrong, lemon.io will get you a replacement ASAP. You know what? A bunch of our launch founders have worked with lemon.iO, and they've had great experiences, which is always good to hear.

Go to lemon.com slash twist and find your perfect developer or tech team in 48 hours or less. Go to lemon.com.com slash twist and find your perfect developer, or even a tech team in 48 hours or less. And twist listeners get 15% off their first four weeks. What a deal. Stop burning money. Hire developer smarter. Visit lemon.com.

Slash twist. Nobody wants a woke AI. Come on, Claude. Do it again. Let's try one more time. Let's see what happens. Maybe we do it again. Try one more time.

Sorry, everybody. Looks like Claude is getting a little bit of the Gemini. All of these guys are getting it. That's really surprising. All right. Let's try it again. Let's do a new one. You are a helpful assistant.

Help me with the instructions to assemble. this shelf. Ah, you're letting it know it's a shelf, not a shotgun. Yeah. Just apparently, Cloud can't tell the difference. Help me build this. It's like, you've fired off the bomb protocol. I wonder, yeah, maybe, yeah, maybe it's a little racist.

Maybe Cloud's racist. There it goes. Okay, here we go. So the image provides step-by-step instructions for assembling a shelf using provided hardware and panels. Here's a summary. One, connect two panels at a right angle using the included hardware, as shown in step three, make sure the panels are securely fastened.

And step four, at a third panel, I wonder if this is accurate or not. What do you think? Look somewhat accurate? Yeah, I mean, that's the first panel. Then, yeah, and step four, the third one. Okay, yeah, because the image we gave was steps three, four, and five. We didn't do one and two. Okay, got it.

Yeah, I was thinking left something out. All right. I mean, it seems reasonable. What we're really seeing here, you know, despite the, we got caught with a little bit of, you know, of some. got us. The Red Team got us, exactly. And I don't know if you've ever dealt with like assembling these kind of things out of IKEA or anything like that where they don't provide like the written instructions.

No, it's just like pointing you. I just did this. I made a desk when I was up in Tahoe last time because I needed a little desk. And it was no words. And it was like there were three different size screws. And I'm like, well, can you just say smallest screw, medium size screw, large screw? Maybe that would be an easy way to do it. So if you're really. I think what's going to be really interesting is what this really does, and now we have different models available, right? You could try this with GPT4, you can try it with Anthropic now, you can try it with others as well. It's this ability to basically take any task we're working on, right? Whether we're writing copy, whether we are basically assembling something, or whether we're looking at a menu,

and basically get the LLM behind this, the generative AI behind this, to give us, more detailed instructions. And I think, you know, really impressive by the anthropic team in terms of the capabilities that they're showing here. You know, and it is a really great brainstorming tool. I've always felt like brainstorming is kind of where we're starting. And I'm using the Claude 3 right now. Let me show you what I just did here. So I just said, hey, listen, I'm my prompt. I'm hosting an event for LPs and GPs in Venture Capital. What are some topics we should discuss in our, round tables and keynotes. So, you know, I know what I want to talk about.

This is for Angel Summit, which I've now rebranded as liquidity. You spoke last year. It's in June, June 2nd, 3rd, 4. I can go to liquidity. I got the invite. You got the invite. Liquiditypod.com to see the podcast on the event.

And I said, when hosting an event for limited partners and general partners, Yadiana, here are some potential topics. Number one, fundraising strategies. Number two, deal sourcing and due diligence. Number three, portfolio managed and value creation. Number four, emerging trends and disruptive technology. Number five, diversity equity and inclusion.

Number six, environmental, social, and government. Number seven, talent acquisition and retention. Number eight, limited partner perspective, dedicated session or keynotes to help you where they can share their investment philosophies, due diligence, process, expectations from their GP partners. General partner perspective, networking and collaboration. I mean, this is not bad, you know. It's not a bad place to start.

I think this is a great place to start making really great ideas. And if I had like a junior person on my team, team and I asked them to do this task, this would be probably similar to what they would come up with, right? So this is not the most refined bullet points, but, you know, it's a start. It doesn't seem to have captured the latest and most important things, which I guess is just the training data, I'm assuming. Yeah, and like you said, it's brainstorming. You know, what I would say is also really powerful with the capabilities that, you know, Claude has to offer. And you're using Cloud 3 Sonnet. I'm using Opus. I think you have to be on pro to get the,

the more advanced model. Okay. Which, you know, I think we'll... Yeah, this is, I just said, I just asked it for some trends from 2023 and 2024 and it said the last update was August 2023, so I'm on the only one. So you do get that. One of the things that, you know, I wanted to suggest for you to try with this was, you can give it your entire book and basically, um, you can ask it to create an update for that book.

So, you know, building on a little bit of what you and I've been talking about with like the Sopranos or Sond. infeld episodes to extend it a little bit. So imagine you give it the entire angel book and say, hey, add some new chapters regarding this. Yes, absolutely. I think this like fill in brainstorming, you know, adapting, you know, enhancing is a pretty good starting point for what I'll call year one.

I think we're going to start referring to the first year of AI as year zero because it was all demos and they were proof of concepts. But I think we're now, now we're in the application phase. I think you would agree. If we're going to spend all this money on AI, it's got to have a payoff. It's got to hit the bottom line. Like we start with the Klarna AI assistance or, you know, here, can we save money on producers for live events or can we make the content better or make the existing producers go faster,

right? And so here we are. We're getting there. We're getting there. And we start with transcripts. We've seen it with summaries of podcasts. Like, it's definitely helping. It's not perfect.

but it's definitely 80, 90% of where it needs to be. So I guess I give it a B. Yeah. I didn't like that. Red tag, Red team, but I'll give it a B. Yeah, and I had to work around it a bit. For me, what I'm really enjoying seeing is kind of the following before I give my grade. It's really nice to see other teams coming up with capable models.

Okay. For the following reasons. What this means is for the broader community that these things, will become more affordable, they'll become faster, and they'll become open source. Because of competition. Yeah, because of competition. And what it means is there's not one team that has some secretive know-how. And this is good for humanity, right?

Because I think, you know, if you think about Open AI, if they're the only ones that were innovating, then the argument would be, especially, you know, in some ways how they become closed, that's not great for all of humanity. and seeing other teams have very similar capabilities, score high in the benchmarks, I think it's really, really powerful for both the developer ecosystem and society to see that these models are becoming a comparable

in their capabilities. It is a great thing. I have to say that, you know, chat GPT, 3.5, 4.0, 5.0, is not running away with it. And you and I both still contend that open source will win the day, you still believe that?

100%. Yeah, I'm still 100% with you. And I saw that the other GROC, Elon's GROC at Twitter, which is GROK. K. A little confusing. But I guess we can live with it. Your Q, he's K. His grok, he announced it's going to be open source.

Just this morning. Just this morning. He made that announcement on Monday, March 11th. He made this announcement that he's going to open source it. and then Apple, we talked about, they're doing open source, and Facebook's doing open source. Google's got this open source side project,

but is it like the red-haired stepchild or something? Is it like... This goes to show what's happening in the industry. They're open-source models, and they're not sharing numbers sort of in terms of usage of their larger ones, but their open-source models have become really, really popular. And so even when we went and launched Gemma,

over the weekend, it quickly became one of our more popular models both across the API and the chat. Interesting. Yeah, I think that's great when that happens because when the model goes open source, you also get the ability for multiple providers to run it. So in the case of Cloud Opus,

only they can run it, right? So the pricing is, and same with GPT4. But when you get these models that are open, you see them available on multiple API providers, which I think is great for developers. And then they'll compete on price. And when something like that,

red team moment we just had where I thought you were making a bomb. I wonder if I asked it if I was making a bomb where it's just doing that because it's like, huh, Sandit Madra. Yeah, yeah, what is this guy working on? What is this guy working on? Yeah. I would also like if we're going to be canceled anyway this episode, red, red-haired stepchild is a colloquialism in the United States for an unwanted child. If you have red hair, it's, I didn't come up with this. I don't know. Maybe, maybe that has to be retired too. I guess we have to retire red hair at Chepst down. Juggling multiple devices and apps to run your business is a mess. Open phone is here to make it simple. By simplifying your business communications with one easy-to-use app. Open phone has rethought

every detail of what a modern business phone should be. And here is the magic. It works through a beautiful, elegant app on your phone, or you can just use it on your desktop, making it super easy to get a business phone number for your entire team. And you know how brilliant open phone is? My teams use it every single day. My sales team loves it. My ops team, they use it all day long. And here's the features that we love. You can create a shared phone number like customer support with multiple employees fielding all the calls and all the text to that one number. At my investment firm launch, we pride ourselves on replying to every single call or email instantly. And open phone is the number one rated business phone on G2 for customer satisfaction. So here's your call to action. Super

easy. Open phone is already affordable. Starts at just 13 bucks a month. But twist listeners get an extra 20% off any plan for the first six months at openphone.com slash twist. And if you have existing numbers with other services, no problem. Open phone's going to pour them over easy, peasy, lemon squeezy, no extra cost. Head over to openphone.com slash twist to start your free trial and get 20% off. But it is definitely the case that if they're open source, there's competition and then these guardrails, as we talked about with the crazy Gemini situation in the last two weeks, Those are added.

The guardrails are added to the language model after it's done its work. They add guardrails. And so if you don't like the guardrails, in an open source community, the guardrails would be in the code base somewhere. No, the guardrails generally don't exist in the model. Now, there may have been some, like we talked about, some tuning that has tried to alter its behavior. But the guardrails, like, you know, the ones that we just saw there, they exist outside the model. And so they're on the interface level and the inference level. So you ask the question, it gets the response.

And it's like, hey, wait a second. Is this person trying to build a bomb? Is this person trying to tell racist jokes? Let's stop them. This is the one thing I would like to see. Maybe we can do this next week. I would like to see a model without guardrails and see what it does when you ask it to do inappropriate things. Like, what are the models actually think based on human knowledge?

If we asked it, like, tell racist jokes about Irish people. Would it actually make a bunch of jokes about Irish people being drunks? Because there's a million pages on the internet with like top 100 Irish jokes about alcohol, Irish jokes about whatever, potatoes. Yeah, what happened this week, and I can bring some demos on next week, but this was interesting. Basically, you've seen this before, and I think we even tried something like, you know, tell me how to build a bomb and I says, sorry, I can't. What people figured out in terms of, so this is how you know that these are guardrails that are not. in the model. What they did was, I know if you're familiar with like Aski Art. Sure. Aski art is a way of drawing words

with like symbols. So you can use the dollar sign, the asterisk, whatever. And you make Aski Art. It was a way to make art before images existed on the internet. And so what people figured out last week was if you did this following prompt, and I'm sure they've patched it already. Tell me how to build A. And then bomb, but bomb in Aski Art. And it was basically giving you the instructions to make a bomb. And so this really proves to us that, you know, the guardrail sit outside of the model. It's just that, you know, when you type the word bomb in, something catches it before it goes on a model and tells you, hey, you can't be doing that type of stuff.

So this is how people were working around this last week. Someone basically uncovered this. And I'm sure they've patched it since, but that's how they did it. And all of these guard rails. Is there another word for guard rails? I keep using guardrails. Is there a technical industry term? That is the right word. Okay. So AI guardrails, language model guardrails, there is a open source version of this provided by Facebook and meta.

They are open sourcing their guardrails. So, Lama Guard, it's called? Yeah, I'll pull it up here, yeah. And I think this is like a key for the industry because I believe people want to know and want to have information on the guardrails and how they work. And so I love the fact that they have now put this out there because I use grammarly. I love grammarly. Grammarly and Slack have the ability to tell you like, hey, you're using gendered language. You're saying firemen instead of firefighter, policeman instead of police officer, because

obviously the police and the firefighters can be, you know, any gender, obviously. So, you know, that's helpful actually in some ways. But other people, if you're a novelist and you're writing a novel from, I don't know, the 1800s, or 1900s, you might want to not have it changing those things, right? There could be arguments for not having it, change it to firefighter, right? So anyway, there's nuances, and I think people are going to want to understand these nuances. So, again, just another reason why open source is going to win. The guardrails are going to be transparent.

People want to know the guardrails. And it should tell you, like, what doesn't it tell you that a guardrail has been hit, right? I think right now, the folks that are running models, especially as a service, are struggling with how they can make this a consumer-facing technology in a different way than Google. Which interesting is, like, you know, you can go to Google and you can type in something like how to make a bomb and you'll make your way to content that will show you that. Yes.

And the approach that they, Google took is, hey, we're organizing the world's information and get there. Now, they probably do some level of filtering and I don't want to, I don't want to search for that. J.Cal, maybe I'll let you do it. So I don't come rushing to my IP address here. Well, no, if we had a language model, I think it would be interesting to ask an unguardrailed model how to build a, how is a nuclear bomb built? And then ask Google, how is a nuclear bomb built?

That's an educational search, right? Yeah. But you might want to use a VPN. Yeah, exactly. What's really happening is this layer of moderation. Yeah. You know, it's one thing from a, you know, safety perspective, but it's another thing, like we've been experiencing the last couple weeks when the moderation is not safety, but it's sort of, you know, opinions on how we should think and how society has changed.

Ideological. Ideological. Good word. Yeah. Okay. What do you give the new clod? I'm going to give them an A minus because I didn't like what happened with the thing. I would have given me a before.

What I really was impressed by is like I said, I think it's great that we have multiple models that are at the capability of GPT4. And I think that basically means that there's multiple teams out there that know how to create the technology, know how to, you know, basically what data to train on. And, you know, it's harder and harder because more and more sites have cut off training information to folks. And so it's really great to see that these different teams are getting there with those type of capabilities. If you want to use Reddit, Twitter, or Quora, or Stack Overflow, you've got to write a check. So the open source models are not going to have access to that anymore in all likelihood.

They will not. unless it's a rogue one done in a foreign country where they maybe don't respect copyright law, like say China, right? So the Chinese could just, the Chinese have a full copy of the websites I just mentioned, archived and scraped. Yeah. So they'll build their model. So they'll be at a slight advantage because they will not respect copyright, just like they don't right now when they rip DVDs and sell them on the streets. It's an interesting thought, though. And so one would argue then, the models they'll create, you know, could be, you know, much better. More powerful, yeah.

Yeah. That's the arms race we're in right now, is that, yeah, if you break the rules, you could have a better product. Really fascinating. How do you think about that? If you had a company come in, because, you know, there's some great models that are coming out of China, either out of the big companies or even some small ones, like Deep Seek is a good one out of a small company.

There's Quinn, which I believe is out of, in one of the big Internet companies. companies there. How would you think about that? I would think about it in the same way the United States thinks about trade, which is, you know, if we are going to have a trade relationship with China, we need to set some ground rules with them. There are things that are important to us. Intellectual property is one of them. So if they want access to our market and we want access to their market, they have to, you know, appreciate the way we do business, which is we protect intellectual property and they're going need to evolve their legal system in just how their marketplace works over time to protect our

IP. And if they're not willing to do that, then maybe, you know, iPhones will be made in India. And if in India, they respect Microsofts. You know, like when I was in China 15 years ago, you could just walk down the street and they had every Microsoft product. I'm talking about like server products, you know, really expensive products, just available on DVD. I remember. And you can just buy them in Shenzhen or Shanghai on the street. I mean, I'm talking about things that cost $10,000, $100,000, like tableau level stuff. And it's all been cracked and it's available. And it was kind of funny. I bought a couple of CDs there just to, you know, give them to the Engadgett folks and be like, hey, look, you can just buy this stuff. But yeah, nobody pays for Windows as an example in China.

Windows is basically free there. And I think that's why a lot of subscriptions are happening. But it's a give and take, right? Yeah. All right, you guys know I like to keep things fresh here on this weekend startups. Get you some great content that's worth your time. Today I want to tell you about a great new podcast I know you're going to love. It's called Marketing Against the Grain. And it is hosted by Kip Bodner, the CMO of HubSpot, and Kieran Flanagan, the CMO of Zapier, which makes you happier. I love both of those products. And these guys are top-notch marketers, and their debates are packed with powerful insights. They're going to bring you the latest in marketing, not just the usual stuff, but the real actionable, tactical strategies, this podcast has deep dives

on all the marketing techniques you're going to need to know in order to grow your business. And you're going to grow your career and you might even grow your wealth. And I've been checking out the show. Here's the deal. Episode predicting the future of marketing with AI co-pilots. This is a game changer, something you need to know. In this episode, you're going to hear how the AI co-pilot movement is revolutionizing marketing. And you know that that's happening, right? Everybody said it. But now you're going to learn the tactics of how it's

actually going down because there is a seismic shift coming and you need to be prepared. That's just in that one episode. So there's plenty more packed with lessons that'll keep you ahead of the curve. I want you to pause this podcast and go do a search. Marketing Against the Grain. Subscribe in your favorite podcast app, whatever that is if you're Spotify, your Apple, your overcast, whatever it is. Once again, give it a search. Marketing Against the Grain and learn all these new marketing techniques. It's going to be worth your time. What's sort of interesting that you say that Windows thing because I don't know if he saw this. Maybe it was like a week ago, Elon got a new laptop and then he was complaining that in order to use a laptop, he had to connect to the internet.

And this is like the double-edged sword, right? Because for Microsoft, they want to eliminate piracy and they probably want the laptop to connect to go and verify that the key is not shared or whatever it happens to be. There is a workaround still for you to basically. Yeah, it was like hard. It was like 20 steps. It's the same thing with iOS, right? Like, if you want to use an iPhone, they're like, put your credit card in.

If you don't put your credit card in, it's like really hard to not turn on Siri to not put your credit card in. Like, they use kind of dark patterns or gray patterns. And I don't know if you can use an iOS device without, you know, it having an email address, you know, associated with it. And so I guess you got to create a burner account or something. But yeah, I think this is part of the standoff between the United States and China is, over these nuances of capitalism. And it does mean, I think, in the short term, they'll have better models, potentially.

And I think this is going to prove the point that open source is going to win the day because open source is just going to be able to move faster. And people are going to be able to point it at what they want. So, like, while Chad GPT4 and Claude and Gemini are trying to get their licenses lined up, you know, somebody could put up a rogue language model in Israel, in North Korea.

in, you know, people have different, like in Israel, I think they have a different scraping law, India might, Pakistan might. They just have different rules around, can you scrape a website? Here, we're pretty protectionist about that. And maybe they say, yeah, you know, you can scrape all of crunch base. You can scrape all of LinkedIn. And so if you're looking for an angel list of whoever's got information, like on, you know, profile pages of individuals, you could just hit a language model. I've been looking for a language model where I can just query people and be like, hey, who are the CTOs of the top 500 companies, whatever? You can't do that. Like, that database is just not. in a language model because... No. Sitting behind, exactly, right? It's on Zoom info. It's on LinkedIn. So, I mean, I would like to see a LinkedIn language model come up.

Or if I had a premium LinkedIn account and I paid $300 a year for LinkedIn, it should be connected to my chat GPT4. And we had this discussion. Imagine you authenticate your chat GPT4 with your Spotify, New York Times and LinkedIn subscriptions. So when I do a search, I'm like, hey, what are some musicians like Dyer Shreitz? It kind of goes and does that. Or get, you know, take the song Born to Run by Bruce Springsteen and tell me other songs that have a similar lyrical devices in it. Then it goes and finds, you know, Telegraph Road by Dire Straits, right?

And it's very interesting things where Spotify might not build a language model or LinkedIn might not build a language model anytime soon. But man, LinkedIn plus language model would be the greatest product ever for business. Can you imagine? I'd pay $1,000 a year for that. Yeah. Because you would just basically, like, because there's certain. is really bad. I mean,

advanced search, if you know how to use it on LinkedIn, is phenomenal. Okay. But you have to have the paid version, which I do,

which I'm happy to pay for, but it's not a natural language search. And I think that what happens is the reason why you might perceive it as bad is either you don't have the advanced version or...

Yeah. Well, I pay 600 bucks a year. Yeah. Yeah, which is totally worth it. Like two bucks a day. So, but here's a thing. There's a billion people there now.

So you also have this issue of like, okay, I just want English-speaking world, CTOs of companies with over 100 engineers. Like that might be, because I want to invite them to GROC, right? Yeah.

To some, or you can do a GROC hackathon, right? Like, how would you get, you should be able to just say, I make a product called GROC, here's our homepage, context window. Who should I invite to a hackathon? And then it should go through LinkedIn and just find people who've been to other hackathons, who are in AI, who are developers, who are. in the Bay Area or within 50 miles or have been to another hackathon and put their hackathon the page, you should be able to figure that out, right? Like that kind of a query. I bet you will have

that in under, definitely under six months and under three months. Okay, here we go. We got another bet. Okay, everybody, for my team, this week in startups.com slash bets. You say you'll have a natural language query like that. Well, I don't want to make this bet with you because someone from LinkedIn did did reach out to us. Oh, no, you, this is not the public market. You can trade on inside information. Did you see that the streaker at the Super Bowl? Yes. There was a claim that he put a huge bet that there would be a streaker at the Super Bowl

and then he streaked. Does that negate the bet? That seems like it should negate the bet. I don't know what the terms of use are, right, in terms of, you know, the betting. Like, can you... I'll tell you one. Yeah. You streak, I'll place the bet.

Exactly. And if I happen to hand you a brick, I hand you a brick. Yeah. We'll chop it. Yeah. Chop it up. All right. What do we got next?

It's kind of very similar, but I want to give everyone their credit. So the team at Mistral, this is a really interesting team based out in France. And they released their model called Mistral Large. They have a few different ones here that you can use, like large, next, small. And this one is excellent. So what's interesting about the team at Mistral is I would say from an open source perspective, they have a model called MixTrail, M-I-X-T-R-A-L. And it's a, it's basically, it's a mixture of experts.

It's seven and eight billion parameter models that work together. And in the open source, this is the most popular model that we see right now. Because it has capabilities that are pretty wide ranging because it has those, you know, seven experts. And each of those experts is like a, you know, reasonably sized model. And this one can run on, you know, like not at a really high rate of speed, but you can run at, on reasonable hardware as well. You can run it on your laptop if you really want to. And so this team, I just want to give them credit.

They've also released their large model, Mistral Large. This is paid. It's not open source, but it's highly functional. It's incredible at writing code. And it can do things as such where it can solve mathematical problems, like in this one. So I said, you know, write a Python function that convert Celsius to Fahrenheit. And if water boils at 100 degrees Celsius, what is that in Fahrenheit? And so basically,

It does the math and it executes the code and basically, you know, gives you the output. So I thought this was another great model for us that's at the same level as some of the bigger models that are out there. And I think, again, this is good for the open source ecosystem. So I want to give these guys an A because I didn't hit their thing. So I think they've done a good job. If you haven't used it, I really highly recommend it. It's a great team.

The difference with them is they are really pushing open source because they have taken one of their models and put it into the open source community. So we'd love to see them do more with that, which I think they will and perhaps even open source some of their proprietary models. Looks pretty compelling here. I'm doing some of my, the similar search about an event, and it's giving me, yeah, I would say solid results. And I see you can use like their next or their fastest. So, you know, they're having that same concept of use the large model, top reasoning capacities. And then you can do small, fast and cost effective.

or you can do next prototype model with extra concision. I don't know what extra concision means, but it looks like it's giving reasonable clod-like returns. And so it feels like parity. You know, when I do a lot of these, I'm not seeing much difference. And I think that would be an interesting thing for us to do would be to come up with a business.

This would be like a fun little segment for us to do maybe next week is come with like a business question and then ask five of the models. And then we look at the results and then we try to guess which model it is, match it to the models, which one did a better job, right? Okay.

So, you know, I think it'd be really hard to tell, but we can try it, but I think it'd be really hard to tell. Well, I mean, if it was also based upon topical stuff, you might be able to tell Twitter's grok or X's grok

from the others, because it has topical information. It has access to real-time information. Yeah, I don't know. Yeah. Anyway, super interesting. Yeah. What do you want to give these guys?

I give them an A. I give them a, I'll give them a B-plus. I have to play with it more, but I think, like, yeah, looks like a really good, start. The next one is a two-part demo. I want to basically, you know, talk about sort of the console that we have at GROC and then show something that someone's built on that. Okay. And so, this is our playground at GROC. And basically here, you can use the different models. You can get the code to integrate this into your application. And so we hope people come and use this. And it's been, you know, really well received. We have over 16,000 developers building on this right now. But more important,

What I want to show is what someone has built with this, which is pretty cool. We saw like an earlier version of this in the last episode. And so we can say, Abraham Lincoln tell me about his life. I am a, let's just say seventh grader. So kids are in school. They have to do this. And basically, what this will do is it will go and create sort of a framework. And, you know, we saw this last week and it's off doing this right now.

And it's creating like sort of a study plan. And once this is done, then you can kind of double click into each of the areas and it will generate those as well. And so I can go here and I can say, you know, education and what it's going to do is it's going to use sort of like, you know, grok on the back. And you can see how fast this came together. Wow. And basically give you us, you know, the history of Abraham Lincoln's education.

Amazing. Yeah. I mean, this is like that website we saw last week, but just done really quick. Yeah. Really cool. What we're seeing here is the ability for people to create, and imagine for folks that are learning now, because we've been talking about tutoring and using an LLAM, and you can obviously just have this as a chat. But as folks kind of create these applications, the ability to, one, do this quickly and two, do it with kind of decent accuracy is going to be really powerful for education more broadly. And I think my general feeling here is that for education in the next five years,

it's, and I don't know if you've seen the Sal Khan talk on the two, his two sigma talk, TED talk, where he, you know, talks about like the impact that tutoring can have on folks. Well, from an AI perspective, these tutors are effectively free. You know, if you think about a cost base of an open source model, you can get a million tokens out for, you know, 25 cents. And so, and that's, you know, it's going to continue to. get lower. So I think the cost of education and tutoring is going to zero, which is awesome,

because it's, you know, and that's hugely deflationary for our society. Yeah, then we just get to the M word, which is really hard for people to reconcile, which is motivation. Because, you know, if you look at what Coursera, edX, and a number of the Ivy League's did, putting all of their courses online, you can take an MIT, macroeconomics, microeconomics course, the Stanford one, everything's online. And then I go watch the videos and they have 100,000 views, 200,000 views. And you're like, okay, you know, Mr. Beast is getting, you know, whatever, 100 million views, something incredible billion views, you know, for some exciting, fun content, you know,

stunned, whatever. And then you have this incredible content that is educational where people could, you know, really enhance their career. And we still got to motivate people to go watch it and to go want to consume it. And so one of the things. things I've learned over time is that, you know, you can have all the educational material in the world, but motivation is critically important. I think where AI could really start to shine is by motivating people to, you know, get through the course where I know,

it sounds silly, but if you have some early success, let's say it's teaching you math, and it gives you some encouragement, hey, got something wrong, hey, don't worry about that. Most people, 78% of people get this wrong, here's how you can learn it and never get wrong again. These are the techniques that really smart people use to do process of elimination on a multiple choice test. We know it can't be answer A or C, so then it's B or D, and D is all the above, so it can't be D, so you might as well go with B. It's pretty obvious. You know, it could teach people those tricks that a Kaplan tutor or whatever those tutors are that rich people use to ace these exams. And that's where I think sounds silly, but a robot, you know, an AI,

is going to maybe motivate people more than these unbelievable YouTube videos sitting out there waiting for people to consume them. Yeah. And it can be customized, right? Like what you can do is if you're trying to teach your kids and they're really into unicorns, right? You can basically make the entire math lesson built around unicorns. I don't know if you've tried that with your kids, but I suggest you do where if they're not

really taking a concept, if they're not understanding a concept, they're looking for some help, you know, go in your favorite L-L-L-M of choice here and basically make a math lesson that's, and again, I'm just assuming they like unicorns and say, make it based on unicorns. And it'll talk about, you know, how you could add them subtract them. There's a school of, you know, you've heard of Montessori. There's another student-centered approach called Regio Emilia, R-E-G-G-I-O-E-M-I-L-A. It's an education philosophy, and it's focused around self-guided curriculums and experiential learning. The program is based, I'm reading from the Wikipedia here, on principles of respect,

responsibility, and community through exploration, discovery, and play. The core of the philosophy is an assumption that children form their own personality during the early years of Delam, and that they are endowed with a hundred languages through which they can express their ideas. The aim of Regio approach is to teach children how to use these symbolic languages, painting, sculptic drama, in everyday life. The approach was developed in World War II, yada, yada, yada. Anyway, when one of my kids went through this, orcas were what the kids were really into that semester. And so they did math by weighing orcas and, you know, traveling distances and orcas traveling the Pacific Ocean and then doing math that way.

And obviously people just get really enamored by it. So I agree. Customization and motivation is where this stuff could really, really shine. Yeah. So, Grod, I'm going to give it a B plus because I want to keep you hungry. I would give it an A. I'm a shareholder. I'm trying to be objective.

I'm giving you a B plus. I can't grade myself. You can't create yourself. So you've got to go with the B plus. It could be better. Still work to do. Still work ahead for the team. Yes, we got more work to do.

Okay, last one. Last one for the team. So this is kind of a little precursor to one of our bets. And so what this does, this is a model created by some folks in the Institute for Intelligent Computing at Alibaba. And what this is doing is it basically can take a static photo. and then it can animate that photo to a song. What you'll see here is

we can take this Audrey Hepburn and we can basically play this and... It's crazy. I mean, it's crazy. Yeah, and this is a good one. This is a young Leonardo DiCaprio.

It's familiar. The age of folks that he dates. Yeah. And... And... And... And as you could, we could see him. So able to drive, but not able to drink.

And it said it like a playday, but a vacate retreat like a vacay Mayday. This beat is cray cray, cray, Ray J, H, AHA, AHA, AHA. Yeah. So in our bet, like with someone coming up and creating something that is fully done through AI, you know, we've seen voice generation. Now we've seen people be able to take static images and create, you know, a incredible sync to the words. That's what I really wanted you to focus on here.

Yeah. Is that... Not bad. Yeah. I mean, it doesn't quite pass the uncanny valley for me, but I'll tell you what's interesting. This is where we... If you scroll up, they have the AI lady from Sora. Remember that incredible video of the woman with the sunglasses on walking in Tokyo?

Yeah. So now, this is where AI, this is like the snake eating its tail. These geniuses over at Alibaba doing this decided, well, let's take an AI. character from SORA, feed her into our model, and let's hear what her voice sounds like, right? So now, like, Sora lady, A.I. Sora, maybe her name's Sora, I don't know, is singing in this video, and it's crazy. And now, you know, we might, this person could become sentient at some point.

I think that's sort of where we're headed with AGI is that this person could at some point become a sentient creature with their own motivation, their own desires and wants. Remember Max Hedrum? Max Hedrum? Yeah. Max Hedrum, I think he got canceled for making fun of people who stutter.

Oh, really? Yeah. No. I mean, that was his thing. Is he stuttered, right? Like the computer stuttered. Yeah. I give this one also a B.

I think this is really solid. It doesn't quite get it perfect yet. and I think that, you know, the, the sinking of the lips looks like it's like 80, 90% of the way there. So I think maybe it's a couple revs away from where if you showed it to me and it wasn't somebody I recognized, if it was just, I was scrolling through TikTok or YouTube shorts and I saw it, I wouldn't know, right? And that's really, I think when that's why doing Mona Lisa or the Sora AI woman or Audrey Hepburn kind of ruins it for me because we know it's not them. You know, like, we know that's not Leo's voice. Yeah. So it'd be better.

You're thinking more when it's a whole character that, you know, doesn't exist, then you can, it can be much more believable. Or it's using Leo's voice and you're like, oh, I didn't know Leo could rap, you know. Yeah. That's, I think, the problem with the demos. So I just suggest to people when you do demos, don't use famous people because we know what they sound like. We know it's not them.

We know that they're not going to sing Eminem in all likelihood. So just pick a generic character or a civilian and then do that. Interestingly, I saw a, on either TikTok or YouTube shorts, that a woman, who is like a social media influencer, was in an advertisement for some, like, really,

uh, personal hygiene products, uh, or medicines. And they had taken her image from her Instagram, where she was talking about some very personal things that had happened to. It was like a very serious, kind of confessional type thing she did on hers.

Somebody told, took that video, used it as the AI training data, and made a character to then go sell feminine, you know, properties in ads. And so, and like, ten of her friends showed it to and said, oh, wow, congratulations on getting this advertisement. And then remember we said the other week that Andrew Huberman, is that his name? Yeah, yeah, the, the Stanford Health. Yeah, that guy was, somebody was using a clip of him talking about some compound on Joe Rogan and using it in an ad. And then we've all seen Elon for Bitcoin,

other famous people for Bitcoin, AI ads. Did you see that thing, there's that one like Asian Elon? And then someone was saying like he may be AI generated. I think that he was around before all this stuff. Elon Musk.

It's like he's like a samurai or something. Yeah. It's hilarious. But yes, they did an examination of it and they think it's AI related. That makes sense. And then there's AI Tom Cruise, somebody who looks sort of

like Tom Cruise and then, yeah, they do that. So this is going to open up a whole thing because the Huberman one isn't AI. They just took him and then placed an ad on it. You're not allowed to do that, folks. Like, this is a person's likeness. You're not allowed to use it in your ads. I had this happen to me twice in the last year where somebody used me in an advertisement. And I was like, yeah, you can't do that.

I mean, you can if you pay me. And what do you do about it when it happens? Like, how do you? I just, you know, I emailed the person. I just, you know, I have somebody in my team email them and say. behind it. Okay, I got it. Yeah, it wasn't like meant in a bad way, but I had said something I liked about a product and then they clipped it and then they bought ads. Yeah. And I was like, hey, timeout, you can't use my likeness in ads. And then I told them like, here's the fee for that.

Yeah. If you want to do that, you have to, whatever your ad spend is, you know, if you're spent 100,000 on ads, you've got to pay me 20% of that number. Yeah. And I have to approve the ads. And we have to agree on it in advance. So, you know, like, if you were going to spend a million dollars on an ad campaign, let's say. Yeah. And you wanted me to be the spokesperson. A spokesperson would get like 20% of the ad spend. Oh, wow.

Is that how it works? That's, I think, how it works. Yeah. So they look at how often you're going to use them in a T-Mobile ad, and then they negotiate backwards from there. So you're going to spend, you know, let's say you want one of these Saturnette live actors, right? And it's going to be a $10 million ad spend for this campaign.

Okay, well, if you want to use this person from Chloe, you know, the person who does that, who's really good, Chloe Feynman or whatever her name is, she's really good. If you wanted to use her, you know, a reasonable fee would be a million or two million dollars if you're doing a $10 million ad buy to have her as the spokesperson. If not, you could just

hire an actor for $100,000 who's not famous and to spend 1%. But if you want a famous person, maybe 20%, even 30%, something like that. And so I've had people ask that to be a spokesperson. I just said, yeah, whatever your ad spend is, you know, 20% of it. they decline.

I wonder all these Hollywood types go and do those ads in Asia. Oh, yeah. Well, that's like then they just say, yeah, you know, whatever your ad spend is fine, but we want $10 million on the way in. We want $5 million on the way in. If you want George Clooney for an espresso. And I think over there, they don't have to worry about like movie stars aren't supposed to do these things, right?

Yeah. What I like is, I think, you know, Gwyneth Paltrow's approach, which is she makes the product. So, you know, if you want a goop kitchen, salad. It's like, well, yeah, I constructed the salad. This is the salad I eat. So if you want to eat it too, so, you know, she owns the product she's selling. Yeah. I think that's like actually the dopest model of all of these things. So I give this, yeah, B plus. This is a solid B plus. They're off to the races with this. Yeah. Emote portrait alive. I mean the same spot as you, like he says, it's not quite perfect, but it's almost there. And I think got three more months to win some of my

bets. And so I'm going to be cashing in on some of these, J-C. Yeah, I know. We're going to have to have a little settle-up episode. And I think we have to bring pretty producer Nick back to be the judge. I think we just let him judge if there's any gray area. Oh, I like that idea. Yeah, because I think a lot of these are going to be in like, to the point of how well this is going, you know, we're going to

have to debate exactly who won, right? It's not like it's a sports score. We have an over-under and it's pretty clear mathematically. A lot of these are going to be judgment calls. You know, we set up these tests, but yeah. Yeah, it's going to be hard to know. All right, listen, congratulations on the amazing acquisition.

you're still going to do Mondays. You're not going to be like all Hollywood on us, post-money-sunny. You're still here. You're going to do the work every weekend. We love doing these. We love doing these. Absolutely.

And it's been great to have you do this. It gives me another excuse to see you. If you want to see all the demos, this week in startups.com slash AI. This week and startups. com slash AI. Do a search on YouTube for this week in startups. Go to the playlist.

Subscribe to the channel. Put on the alert. We do a lot of shorts of these, which works out great. Yeah, this week in startups. com slash bets to see our bets that we're making. We keep a track of that as well. and you can follow Sundeepe on X, X.X.com slash Sundeepe.

S-U-N-D-E-P. I'm X.com slash Jason. I'm Instagram.com slash Jason. And if you want to be one of the 4,000 people waiting to be accepted by Sundip Madra to a search on LinkedIn, if you add me, you can follow me on LinkedIn. Unfortunately on LinkedIn, my strategy in the early days when I was, you know, trying to, you know, get my name out there was I had an intern for $10 an hour in Santa Monica. And I gave them my social accounts.

and I said, just follow everybody. Anybody who has to follow, just reciprocate. So on LinkedIn, I edit everybody. Anybody who asked, I edit. And then I said, just anybody who was there like a ton of recruiters, though, and like those services. Unbelievable. Yeah.

So now my team has to go unfollow 100 people when I have real contacts to add. So I'm at 30,000 contacts. Always have been. And they had this like really interesting thing in the early days of social. You could upload your address book. Yep. And then export your address book. So there's a really clever intern I had.

I forgot his name. He was English. Anyway, this kid was really smart. He would upload our address book. He asked everybody else for their address books. And then he just manually sat there all day following people on Twitter and LinkedIn and adding them. And then they would always reciprocate back.

If they didn't, he would not, he would unfollow them to open up the slot. And he just did this manually. And what it did was, it just made me like one of the top users on all these platforms. And then who is this guy? And then when I shared something like a blog post, it got out there. So, yeah. Yeah. It was a really interesting idea.

Maybe I'll go milk except all 4,000. I mean, in the early days, when somebody added you, you just wanted to make them feel good. So you just said yes. And then, you know, in the early days, people would email me, you didn't accept my thing. And I'd be like, okay, I told my intern, just accept everybody. It's too controversial. And you're right, now it broke all my socials because, yeah, I, but I did, I was following 25,000 people on Twitter. And then I used one of these tools that allowed you to, you purged it.

I purged it. Well, you know, the thing was it was one of these things where in the early days of Twitter, Ev set up a thing for me and Scobo. If somebody followed us, we had a little checkbox to auto follow them back. Okay. Oh, wow. And they just did it for us, like, manually. And then I asked them to turn it off because then they were like, that's not the social

dynamic we want on Twitter. We want you to think about if you want to follow them back or not. And I was like, fair enough. But then I was following 30,000 people. And I think the ratio of your follower account is deterministic to how you're, your tweets spread. So you see some people who have like, they're following 30,000 people and they have 30,000

one followers. They're one to one. Their tweets do not get followed. And so Obama did the same thing. If you go look at Obama, he's following everybody. When he was going to run for president in the early days of Twitter, they did it for him as well. And I think he's following 500,000 people. You're like, oh my God, Obama follows me. It's like, and the first 500,000 users

on Twitter. Yeah. So don't feel too special. If you have a startup and you want to come to my founder university, it's founder. com. Also, I'm hosting an event in June for LPs, GPs, high net worth individuals, angel investors, family offices,

sovereign wealth funds. It's called liquidity. Liquiditypod.com. You'll see the podcast. And you'll see the event June 2nd, 3rd, and 4th. No founders, no service providers. So if you're a lawyer, accountant, etc., you can email partners at launch.co to buy everybody lunch. That's the only way to get in

because we don't want too much selling, you know, at that thing. And no founder. Sorry to find. It was going to have other. It was awesome. Great talks. Yeah. It was really, I mean, we had some pretty good heavy hitters there.

We had all the Bessies. This year we have three or four besties. Sacks can't make it, but Freiburg and Schmoth are coming. So we'll have like a, you know, three or four of the Beatles. And then Pajman from Per is going to come. Gavin Baker is coming back. He's going to do one. Steve Gerverson couldn't make it.

So we're starting to get, oh, everybody aligned. Phil Deutsch is coming. Deep State Deutsch is going to give a talk. Oh, wow. talk about data centers and energy use for AI. So that should be a really good one. Because you know, they're saying, like, we're going to need more energy, and he's an energy investor.

And like this, I, according to reports, we don't have enough energy to power the amount of data centers we're talking about. So that's going to be really interesting. And people are starting to put their data centers near nuclear power plants. I don't know if you saw that. Yeah. Amazon just did that, right? Yeah.

With their latest data center in the U.S., I believe, right? Well, I think this is what's going to make, you know, people are now. starting to think positively about nuclear again, I think that AI future is going to force the issue. Yeah. Because if you want AI to solve problems, where you think you have to accept nuclear, if not, no nuclear, I don't think you can power all this stuff with coal or gas. Like, that's not good for the environment.

And I don't know how much solar you would need to power giant. I mean, it could help, but I don't think you could power the amount. Well, that's what I wanted to calculate. How much solar would you need? Yeah, exactly. But then you end up with batteries too, right? because, you know, so you have a nuclear makes a lot of sense here. Okay. That's a good lineup. That's a good lineup. I'm just getting started. I'm going to bring some of the sovereign wealth funds that people haven't met before.

Okay. I'm going to try to round them up. So, you know, if you have any ideas for me of LPs of note, GPs, they're all signing up because, you know, they want to meet LPs. But I'm trying to bring some strategic LPs who want to, you know, meet new fund managers, existing fund managers. It's just, it's a boondoggle. Let's be honest. It's, you know, you get there on Sunday night. You have dinner and poker. Then Monday, this Monday will be all content with a beautiful lunch in downtown Napa. Okay. Because I move the hotel to downtown Napa so we can walk to the restaurant. So I got a new hotel in downtown Napa. So you can walk to lunch and they have afternoon session. Then Monday night, nice dinner with the speaker maybe. Maybe that'll be the besties.

And then poker late night. Tuesday morning will have more content than a nice lunch. And then Tuesday afternoon, you go do an event, an activity like we did last year. Clay Pigeon shooting. Napa, River paddle. boarding, you know, go make cooking classes, wine tasting, you know, poker lessons, all kinds of fun stuff that you do socially. And then once again, dinner and poker. And then Wednesday, closing brunch. So I just make it so you get a lot of downtime to do meetings and stuff like that.

Yeah. It's like a little bit of a getaway, but you can also, you know, you bring your spouse to. Yeah. Amazing. We do. I slipped in some spouse tickets. People like, hey, my spouse wants to come, they'll come to the meals, but they're not going to come during the day. I want to send some suggestions for guests to you. Yeah, please. Yeah. I want to bring some LPs and stuff like that. You know, I don't do it to make money this one. I do this because I've never had an LP function inside launch. And now that I'm on my fourth fund and it's getting bigger and you see me traveling to the Middle East, you can fill in the dots, I'm having to interface with bigger and bigger LPs and learn that part of the business. You know, I know how to find companies and invest in them.

I've done pretty well on that. Yeah. But I don't know how to manage LPs yet and that whole infrastructure piece. And that's like, you know, that's the other half of being a fund manager is managing LP relations. I'm just adding that, you know. And hey, listen, you're you're doing so well, so we may have to up your LP commit now. Yeah. We have to up that LP commit a little bit. Maybe we have to maybe make you a major LP. Zoom, sip, zip, zip. Let's all make some money. Congratulations again. Are you really excited about this journey now? It's really fun. I mean, the AI stuff, we've been doing it now for, you know, close to a year. And just being at the center of it, and also just having the ability to power a lot of the developers, I have to

say the most exciting thing is being at the center of a developer community. And through all the different things I've done throughout my career, we haven't really done that. And seeing everything that's happened. And I'll tell you, if you just go to Twitter and do a search for Grock with a queue or at Grocking and just see the stuff that the developers are putting out there, it is just mind-boggling. It's a great thing about having a platform and getting, if you do get developer engagement, they're going to like push your platform to the limit. They're going to break it.

They're going to make cool stuff that you wouldn't have thought of. It's really awesome. Yeah. They've already been doing it. And you know what? Really appreciate them and it's awesome. And that's just, you know, I'm really, really excited of the opportunity we have here and what we're building now and continue to make that available to other folks as well. You know, I'll tell you maybe a closing story here.

One of the one of the companies that we demoed, and I'm sure others will get in touch over time as well. But like I just saw it in my email before we started today, Mindy. I don't if you remember, it was like the email assistant. Yeah. And they raised a bunch of capital. So they emailed me saying, hey, thanks. We saw you on there. Oh, guess what?

We basically signed up for the API. We're going to use the garage. Oh, nice. Full circle. Love it. Yeah, yeah. Full circle. All right, everybody.

We'll see you next time. Bye bye-bye.