This Week in Startups - Breaking down Bard with Google’s Jack Krawczyk | E1752

Episode Date: May 30, 2023

This Week in Startups is presented by: LinkedIn Jobs - Post your first job for free at https://linkedin.com/twist Hampton Join Hampton's community of high-growth founders today at https://joinhampton.com/twist Meowtel - Get $25 off your first booking with code TWIST25 at https://meowtel.com * Today's Show: Jack Krawczyk, the head Google’s Bard team, chats with Jason about Bard’s capabilities and Google’s plan for monetization. Following a few demonstrations of the AI chatbot (43:32), Jason wraps the episode with some thoughts on the series finale of HBO’s Succession (01:01:45) Follow Jack: https://twitter.com/JackK * Time stamps: (00:00) Jack Krawczyk joins Jason (4:58) Discussing the responsible release of products (7:43) The delicate balance of choosing to release or hold (10:24) LinkedIn Jobs - Post your first job for free at https://linkedin.com/twist (11:57) Addressing concerns about the usage of creators' content (13:42) Unpacking the process of vetting new features and the purpose of Google's Red Team (18:39) Integrating tools and features into the framework of Bard (20:51) Upholding core values in the incorporation of tools (23:58) Hampton - Join Hampton's community of high-growth founders today at https://joinhampton.com/twist (25:23) The impact of Bard's coding capabilities on developers (29:36) Gauging the distance from non-developers writing meaningful code (31:11) The business model behind Google Bard (34:18) Meowtel - Get $25 off your first booking with code TWIST25 at https://meowtel.com (36:44) The potential paths ahead (39:30) Google's vision for plugins and opportunities for startups (43:32) Jack demos Google Bard (55:52) Reflecting on the pace of AI (01:01:45) Breaking down HBO's Succession and why it's a must-watch * Read LAUNCH Fund 4 Deal Memo & Apply for Funding Buy ANGEL Great recent interviews: Brian Chesky, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland, PrayingForExits, Jenny Lefcourt Check out Jason’s suite of newsletters: https://substack.com/@calacanis Follow Jason: Twitter: https://twitter.com/jason Instagram: https://www.instagram.com/jason LinkedIn: https://www.linkedin.com/in/jasoncalacanis Follow TWiST: Substack: https://twistartups.substack.com Twitter: https://twitter.com/TWiStartups YouTube: https://www.youtube.com/thisweekin Subscribe to the Founder University Podcast: https://www.founder.university/podcast

I feel like I could be a coder again, you know, like go back to like the 80s and 90s when I was doing Pascal and Basic. And I'm like, well, why can't I do Python coding? It doesn't seem that hard when you have this assistant, you know? It's, it's removing friction to creativity. It's amazing. It's magical. And it's a time to be bold and to be responsible, right? Like, we talk about it at a ton. You can release things to generate that speed of insight. Yeah. But you can't generate that speed of insight at the expense of trust. This week in startups is brought to you by LinkedIn Jobs. A business is only as strong as its people, and every hire matters. Post your first job for free at LinkedIn.com slash twist.

Hampton. Are you a startup founder or CEO seeking expert advice? Join Hampton, the private, highly vetted community for high-growth founders. Get invaluable insights, connections, and support to accelerate your business at joinhampton.com. slash twist today and Meowtel. Traveling soon, book a local and insured sitter on Meowtel, the number one cat sitting app. Save $25 on your first booking with code Twist 25.

All right, everybody, welcome back to the show. We're really, really delighted to have our next guest on. Everybody knows, my lord, AI is moving at an absurd pace. And I said, you know, chat GPT, really fascinating. But then I started playing with this Google Bard. And I was like, holy cow. Google's got some great product. And then they started dropping product.

Boom, boom, boom. Lots of product really fast. And hey, the race is on. We've got so many players right now putting out so much great product that I decided to invite the team from Google Bard on the program. Now, listen, it's hard to get big company people to come on the show. Sometimes they're working on stuff. You know, they keep it close to the vest. They don't need to come on podcasts and talk about it.

about it, but I saw Jack Croftchick on Twitter, and he's a senior director of product at Google, and he's working on Google Bard, and you were sharing, Jack, on Twitter, a bunch of stuff you're working on. So I guess you're showing your work. Google is now ready to talk about what you're building and how it's going. So welcome to the program, Jack. Thank you for having me, Jason. So listen, you've been an entrepreneur. You graduated Carnegie Mellon. You were an entrepreneur in residents at Excel partners where people don't know what that is. If you're a really good entrepreneur, sometimes a venture firm will make you an EIR. Basically, it's like an option. Maybe they get first shot investing in your company. It's not like a formal thing. And you were a VP of product

at WeWork, but now you're at Google. Tell me, what are you working on at Google? And is this like a strategy now for Google to get out there and talk about the product a bit more on the social media's? Yeah. So, well, what do I work on? I work on Bard. I've had the great pleasure of being a part of the team since the beginning of working on, working on getting the experiment out there. And the reason that we're putting this energy behind getting Bard out there as an experiment, knowing that it has these profound capabilities of helping people be more creative, be collaborative, while it has its limitations that we're all familiar with with all large language model technology. It's been under development for many years.

We've talked about Lambda, neural networks to transformers to where we are today, has been a large evolution. And you can't build technology that's helpful for people and aligns with their values without getting it in their hands. And so what you see with Bard is direct access to engaging with a large, language model and we felt it was critical to launch it without making a promise of, hey, this is a product because we don't believe that the technology as it stands today and direct access to a large language model should be a product. It needs a lot of work and iteration. And the Bard experiment is here to get the feedback, start the conversations and continue to

build and iterate and apply those lessons to not only making Bard better, but across all of Google's products. And Google's in a different position than a normal startup. You have a huge reputation on a global basis. You have six, seven products at Google that have over a billion users. When Google produces something. Two million.

Some of them have two billion. Yeah, you got YouTube, you got Chrome, you got Android, you got Gmail, you got docs. I mean, there's a lot of products out there with a lot. And then, of course, search. So you look at all these products, and you have to be more thoughtful about releasing experiments

because, hey, the Google brand, the YouTube brand, the Android brand, it means a lot to people. And so you are being very thoughtful I see in the product design about saying this is an experiment, check the work,

and we want you to give feedback. So maybe you could talk a little bit about how you do that on a product design basis to make sure people don't say, oh, Google's great. I trust Google for my searches. It's 98 times out of 100,

the first link I click answers my question or the one box answers my question. Now you got something that can hallucinate. It can drift off. This is a very early product. And it's probably going to take a couple years for it to be as good as a Google search is. So maybe talk about how you address that on a product basis. Yeah. So what you touched on at the beginning there is, you know, we talk, you've heard Google's mission statement out there for many years,

organize the world's information and make it useful. and helpful. And under that are three values. Respect the opportunity, respect each other, and respect the user. And to your point, to respect the user, you're spot on. We have six products with over two billion users, 15 products with over half a billion users. And so you have to be responsible in the way that you roll this technology out. But you also have to respect the user and listen to what the community is saying. You know, it's not there where, you know, everyone, you know, my parents in Florida are, you know, going deep on using this technology yet.

It's not there. It's going to get there. And to build that opportunity, we have to get this in people's hands. And to your point around making intentional design decisions, so we're trying to take a bold approach by putting the technology in people's hands and pair it with a responsible approach. So you see a disclaimer at the bottom of every entry point. Bard may display inaccurate or offensive information that doesn't reflect Google's values.

That's one point. You mentioned hallucination. We know that's a challenge of large language model technology. It was a very intentional decision to put that Google it button there at the bottom of every response that you see. Ah, yes. So when you're at the bottom, you give it a thumbs up, you give it a thumbs down. Or you can just Google it so you know, hey, listen. Bard is an experiment.

If you want to Google it, you can get that, hey, 98 times out of 100 hits the bull's eye experience or whatever the number is. But internally, you had to have a big debate about this. Some people are, let's face it, you know, going to be perfectionist. Some people are going to be, hey, let it out in the wild. What's the climate like at Google now? Because it had to be absurdly frustrating to have so much product being worked on. And, you know, young startups can release whatever they want. chips fall where they may.

They don't need to worry about what people think of them because people don't even know where they are. There is nothing at stake. There's no 2 billion user products. So talk about like how the vibes inside of Google right now. People must be stoked to have it out there, but there were people who said maybe we should hold this for a while. And how does a company like Google come to this decision of how you're going to navigate, release, don't release, you know, and then we'll get into some more granular, tough edge cases. If I could put it in one word, it's effective. It's just so exciting to come to work every single day to be at the forefront of what feels like the next paradigm shift in technology, this platform shift, like the opportunity that we have with these large language models, with AI, what is producing is this ability to collaborate with it.

It's not just about retrieving knowledge, it's about creating it. It's about having an abstract idea, turning it from an idea in your head to a concrete idea, getting more high-quality ideas and less time, building, compounding, moving quickly. We'll walk through a couple of examples later of how I've used it, demo some of it, some of the things that aren't necessarily clear. But one of the things that comes out is you think about getting this technology out to people there's a natural evolution that people take when something that's profoundly new

comes out that they want to use it like the thing that's most familiar which would be search whatever it may be let's take any technology take TV for example what was the first TV show people reading a radio script

and just like you had a radio script you weren't taking advantage of the visuals right and like you you take what's familiar and then you build and compound and One of the things that I found super inspirational of how people are using Bard is the depth of the conversation that they're going to. They're talking to the product or to the experiment as though they're talking to a teammate. I think your terminology is besty. It's like you're talking to a bestie, right?

Like you're having, you're not putting in five, six, seven words, like do your magic. Like, of course people are going to do that. The magic gets unlocked when you're at 30, 40, 50 words, you know, a lot. walk you through an example of like how I got to a 200 word prompt that like is now part of my daily daily process that like couldn't have existed before. And so as you bring this technology out, right, like you got to listen to people going back to that Google value, respect the user, respect what they're looking for. They're looking for this collaborator to help them have these ideas and build and grow and compound on them. And it's pretty fun and exciting to watch.

LinkedIn. It's amazing. I told you they get to a billion users and they just crossed 900 million people using the platform. So just think about how many insanely qualified people are on LinkedIn right now looking for an opportunity, looking to work with you. And you know, the stakes are high right now. You need to have every single position in your company firing on all cylinders. You've got competitors. You've got runway. You've got customers who are demanding new features, demanding customer service. So go post an open role on LinkedIn and be 100% certain that you have access to the most qualified candidates available. And guess what? First one's on me. It's on your boy, J-Cal. That's right. LinkedIn.com slash twist. LinkedIn.com

slash T-W-I-S-T and post your job for free. You got nothing to lose, everything to gain, including getting that purple ring around your profile photo that says you're hiring. And then maybe you start doing some updates in the feed on LinkedIn. Maybe you start following people, doing endorsements of people. Once you get yourself integrated into the LinkedIn network, It's so viral. There's so many qualified people that they're going to click on your profile and see that you're hiring. And then friends over friends, friends, friends of friends, friends, those candidates are going to come and they're going to apply for your jobs, right? And so better candidates faster, better candidates faster, better candidates faster.

That's all I want you to remember about LinkedIn jobs. You're going to get better candidates and you're going to get them faster. LinkedIn jobs helps you find qualified candidates you want to talk to faster. Post your job for free. LinkedIn.com slash twist. That's LinkedIn.com slash twist to post your first job for free on your boy J-Cal. Terms and conditions do apply. I have to say, you know, one of the things, as a content producer my whole life, former

journalist and running content sites, I have some deep concerns about how, you know, content creators' content is used. One of the things I noticed, you know, Google has always had a delicate relationship with the people they indexed. Robots.t.c. hey, if you don't want to be in the index, you don't have to be, but if you're on the index, we're going to need some information. And you've over time, literally decades, navigated the relationship with content creators,

trying to make it fair. Putting aside legal issues, you and I are not lawyers, or I don't think you're a lawyer. But it's trying to be fair with content creators so you can have a working relationship. Same thing on YouTube. Hey, you're a content creator. I put somebody else's music in my video. You have content ID. I can take it.

All these kind of clever ideas. I'm just sharing my screen here. And, you know, I was doing some searches while we were, I don't know, can you see my screen? Okay. Yeah. Oh, look at that beautiful image. Well, you know, I was like, hey, where should I eat sushi when I'm in, when I'm in Los Angeles? And it gave me all these great places, but I noticed it was starting to, when it puts in the thumbnail images, that's something very new, putting in the captions and linking directly to the source material.

So now, if my content was used at least. I can get a potential click. And of course, you know, this can be highly clickable stuff. But my lord, you know, when you ask a follow-up question, I said, hey, which of these is the most expensive and hard to get into is. Oh, sushi Ginza, which I've been to. And it is hard to get into and it is very expensive. And then I said, hey, can you get me some of the top dishes there? And it starts pulling the dishes from the menu. I mean, this is next level in terms of how quickly you're making progress on it. When you're releasing features this fast, what's the process of vetting them and having discussions.

Because these images and thumbnails, I don't know when that dropped, but I use barred basically every day, every other day. This is the first time I've seen one in a search. I don't know if this came out in the last week, or maybe I just didn't do a search that had images in it. When did images start?

Images started two days ago, Tuesday of this week. Amazing. I mean, it's literally like we were in a time machine 20 years ago when the web came out. You're moving product out every couple of days. Amazing.

Yeah. Well, you touched on a, a couple of things that go into the design and building a bard and getting back to like, why is it infectious to be a part of this? You can look at the speed of delivery. We're obviously very proud of being able to get technology in people's hands. But it's really about the speed of insight that we're focused on. Again, this is like the early stage of a long arc of technology. You talked about it. The only way that you can build and grow and generate on feedback is, again, to put it in people's hands. Of course, we're not launching this without testing it with our team inside of Google, going through the red teaming exercises that are critical in the development.

Explain what red teaming is, yeah. So red teaming is a concept in the development of these large language models where you get a group of individuals that intentionally try to break the technology. And this is a critical aspect to that point of making the technology, responsible, making it safe, these things, we are aware, are capable of saying, convincing things that may not be what you want to present. And so they will make mistakes, but there are some pretty egregious mistakes that you want to make sure that you catch before you put it in the hands of people. And so what we're trying to do, again, is respect the user and testing it out, but respect the user by listening to them. That's the number one way to generate that respect. And, yeah, all of these things, people try to break.

They're going to type in some query about Hitler or some horrible human being or some really dramatic event on the planet. They're going to try to get it to say something it shouldn't. And that's all part of what the Red Team does is. So you're going to be aware of that. People are going to type in adult search contents, whatever. And you just kind of try to head it off at the past. But you can't let those gotcha moments.

I did it with Chat, JBT-Bt4. I was like, tell me something great about Hitler. And it was like, yeah, no, nothing great. And it was like, well, but he did create the auto bond. I was like, oh, man, this is getting into a dicey area, right? Like, and you really have to think about the person who, you have to keep in your mind, the person who's writing a term paper on World War II and the person who's trying to screenshot it and try to make you look stupid and say, oh, look what, you know, Bing said, look what Google said,

and then try to make a screenshot or then maybe, you know, get it to, you know, into the popular press to kind of dunk on you. That red team has to try to really do gnarly things. That's quite a job. Is that a full-time job for somebody being on red team? Or do you just rotate in red team members and say, hey, you get you get to do a day of red teaming. This is where it comes back to like the excitement of going back to work. Like getting people to try to help get the product out the door to be ready, get the experiment update to be ready. And by the way, we're being transparent with not only what is going out, but why we're doing it, why are we testing it? What's the conversation that we're inviting?

So you can go to bar.gogood.com slash updates and see when a new feature is out, what's there? Why did we do it? And so it's inviting that conversation and getting back to the point of respecting the opportunity. You talked about images and the ability when you click on them to go. to a site that that is another element of the discussion and discourse that we're trying to have by getting this technology in people's hands.

You know, we go back to Google's core values, respect the opportunity. We only succeed when others succeed. That's been at the DNA of this company for the 25 years that it's been around. I've been here on and off for 16 years, 2007. It was the first time I stuff put in the door at Google. and that respect for the opportunity that when we only succeed when others succeed, you've seen it translate through a variety of delivery that we've had at Google over time. And again, we're inviting the conversation.

This is far from a complete technology that is solved and there. You've got to get the conversations going to figure out how to grow and evolve. And that's what we're doing. and we're super thankful for conversation that is. I notice it's not hooked up to my Gmail yet. And this to me is going to be like the huge win. You know, Gmail was such, were you at Google when Gmail launched the second product at Google? You know, I was a senior in college when it launched.

So I remember where I was in the electrical engineering lab at Cardingham, and my dear friend, Seth Feinberg. who I haven't seen in a long time gave me my first invite. So, Seth, if you're listening to this, thank you for that. It was the first invite system as well. And it was very precious.

I remember, like literally not to drop names, but emailing Larry or Sergei or maybe both of them was like, hey, can I get on this? And they gave me one. And they gave me like three invites. And I had three invites. And I had to be like, whoa, who do I give these three invites to?

These things were worth a lot of money at the time. But GMO was such a, you know, ground baking product because it was like, oh yeah, you never have to worry about deleting your emails. They're just all there. And you can search them. And it was like, Whoa, yeah. To do this with

Udora, whatever mail client you were using was just really hard. And all your mail was stored on your local drive. But now, you know, there's so many insights in my Gmail. How does the Bard team and the office team, the Gmail team, and the

YouTube team, how does in a big organization like this, this giant armada of aircraft carriers, you know, each one of them, those two billion users each those are like aircraft carriers, but you have this core platform technology. How does that work?

How do you take BARD and Gmail and YouTube and then incorporate them? Because I got really specific requests that I want for YouTube and I want for Gmail. And then you're just trying to get me an answer on like the best sushi and the best, you know, and solve those queries and those requests. So how is that working? Do they have their own version of BART and they just go off to the races on it? Do you guys put a thousand engineers in a Google meetup and talk about it? I mean, how do you execute?

Yeah, well, so first, let me say, we announced the ability for BARD to access tools at Google I.O. a couple of weeks ago. So the good news is it's coming. It's coming. Okay. So the intention is there. Yes. There's a couple of things that you've got to get right. Again, going back to core values, respect the user.

If you're going to start having a language model access your email. and interact with that email and produce a result, the user needs to be in control. That's the first principle for which we're building. The user is in control, and you have to be transparent. One information is being seen and how is it being used? Those are the two guiding principles of bringing an extremely powerful technology like this to people. And it's not just Gmail.

There are plenty of tools across Google that we're going to be using and bringing. and it's one piece to get the language model to produce something that's compelling and helpful. And we're very excited about that. And again, going back to that speed of insight, the thing that enables this is Bard's ability to translate between natural language and code and start to the quote-unquote thing. They're not sentient, but they start to produce logic results in the form of code. As it does that, you can start to ask it to write code to interact through the APIs

that exists, et cetera. And so while we build the technology, we also have to build that respect for the user to make sure that, hey, if you're going to trust us with using this tool, we're going to be extremely clear how your most private and personal information is being used. And so we don't want to put it in people's hands

before we feel like that. That design element is confidently in people's hands. This is going to be a difficult one, because you're training the model. Every time people search on Bard, it's a training exercise. So I give it the thumbs down. Hey,

I'm looking for that video on YouTube of the English Bulldog and the monkey playing the Japanese video and it's like, ah, I got that wrong. And you've got to give it a thumbs down and say, no, that's not exactly what I'm looking for.

A human then can intervene on those instances and try to figure out what went wrong. Probably don't want that in somebody's Gmail, right? You've got to be really careful and thoughtful about that. And not training the language model on, yeah, that's the challenge.

You're not changing the lounge amount of private information. Like, this is where, again, like, we are at the early stage of an extremely long arc of a profound opportunity, but you can't just rush these things out the door. You have to ask these questions and solve them intentionally. And so one of the things that we're doing as we roll these things out, again, what are we doing? Why are we doing it? When we believe it's ready enough for people to start interacting with it, we want to want that transparency. And so when, to this point, one of, one of the tabs that you see on Bard

is barred activity. And you can see, what are the, what are the conversations that you've had? And you give users the choice of, hey, if I want to be a part of this experiment, but I feel very strongly that I don't want this information to be locked, then you give users that choice. It's critical. That's part of respecting the user and why it's been there since the day that the experiment launched. When you're the founder of a high-growth startup, things are chaotic. Pretty often, let's be honest. You face a ton of questions.

You've got problems that you don't always know the answer to, especially if it's your first startup. But you don't have to do this alone. That's where Hampton comes in. Hampton is a highly vetted private community exclusively for founders and CEOs. What's their mission? Founders' mission is to create the most valuable and engaged community for high-growth founders. And here are why some members have already called it life-changing. When you join Hampton, their membership team carefully hampicked seven other members to join your core group.

It's almost like a mastermind group, right? Okay, so now you got your core group. This group is like a personal board of directors for you. They're going to provide you with advice, critical feedback, and they're going to aim to help you accelerate your growth. I mean, that's what we're all in it for anyway, right? Personal growth. We want our companies to grow, revenue to grow, the amount of money we raise to grow, all that good stuff. So Hampton members run some of the fastest growing startups in the world, and you probably use some of their products.

like morning brew, blank street coffee, dribble, and more. The connections, personal accomplishments, and a sense of belonging are what make the Hampton membership so special. Community is everything. You know that. Like here this weekend, start our business community,

and here is an even smaller, more focused, tighter community for you to take a look at joining. If you're ready to scale your business, join the Hampton community today at joinhampton.com slash twist. I love this idea. Joinhampton.com slash twist today. in terms of demos

you know there some of the most impressive demos are writing code maybe you could and I don't know if you write code yourself but maybe you could explain to me what impact this is having internally because people writing code

to make AI experiments interfaces etc are now being assisted by AI itself and this is one of the profound aspects of what we're experiencing right now. We talked about how, hey,

images showed up on Google bar two days ago. Are developers going noticeably, dramatically described it for us faster in creating content now, creating,

you know, code now than they did when you joined Google with wavy-davy, you know, 20 years ago. I'm talking about Dave Rehberg,

wavy-davy-davy. We got a, do you know the wavy-davy story? I don't. I mean, I've heard the term wavy-davy, referring to bestie David Freibur from the only podcast,

but you worked with him when he was at Google? I met Dave, well, he was in the early days of Weatherbill. He had been a former Googler, acquired a company named Urchin. We met through a mutual friend, Brett Crosby, saved friends and neighbors for many years. And I'm not going to steal the thunder, but maybe this is something you guys can talk about, but it involves a fraternity chef,

some funky base, and the term wavy-davy-dive. So I'm just going to wavey davy yeah This is a teaser for a future episode for you Yeah Got it well you know San Francisco is you know A long tradition in wavy davies So let's talk about code

And then maybe we get some of your demos And you can tell us like where you're at with the product Any new product announcements or new New experiments that you're working on Yeah so code has been live for a few weeks now When we launched the Bard experiment We didn't feel like the experience was helpful enough.

I think it was three or four weeks after the experiment launched, we gave Bard the ability to code or release the ability for Bart to code. And it quickly became one of the most popular ways to engage with Bard. And I feel like a lot of the conversation is like, do, you know, write a Python script that does X. And of course, people are doing that to kind of like get them started. But there's two other things that are really interesting to watch people use coding capabilities on Bard.

In addition to generation, there's a notion, of explaining code, hey, I have this going, I need to kind of figure out what it's doing, can you do it? Again, not batting a thousand, nobody's batting a thousand, but it just helps people understand code. And the other bit on that is debugging code. So actually, hey, I have this thing, what's wrong? Like, why is it not going?

And one of the features that we built in along with that was this notion to export directly into colap. So CoLab is a mixed notebook and Python ID for that Google is created where you see a Python script and there's an export button at the bottom. You can export, go directly into CoLab and start iterating. And so we're watching it speed up people's process and is really inspiring. And one of the things that left me a frustrated, failed engineer that then became a product manager was I couldn't. debug a regular expression to save my life. And that's actually one of the most popular things that people are doing trying to figure

out why does this regular expression not work. So the more technical people out there, I promise you there's help. It's pretty cool. You know, like if you just go into Google bar right now and say, hey, teach me how to write the code to draw a circle in Python, it'll give you the code. It'll link to people who've written blog posts about this. And then say, well, what do I do with this code? and how do I execute?

And it's like, oh, yeah, you need an ID or you can run it from a terminal window or you can run it in this thing. You're like, what's an idea? And it's, oh, that's an integrated development environment. It's, you know, way for you to execute code. And now you've put those two things together. So this is where, like, the ability to have a guide on the side, a co-pilot, whatever, is really going to be dramatic. How far away are we from having civilians, non-developers, writing and executing meaningful code on the internet? That's a tough one to project, but when you break code into what it is, code is a logic language.

And it's one of these things that's really amazing to see when you start to have logic build and compound, what are the ideas that you can build and grow and develop on? And taking a step outside of art, like, you look at the quote-unquote no-code revolution that's been taking place with other tools out there, you know, the web flows of the world, really fun to watch. I think you just start to see these intersections of people's creativity is coming up against less friction. And it's not just in code.

And so that's the thing that's so amazing. Like, when you talk about like the opportunities that are profound, like, personally, I think like creativity is one of the most amazing parts of humanity. Like, our ability to imagine is magical. And we're reducing the friction to imagination. And so imagine what can happen as you get these ideas and you can go from start to finish faster. And so when does it hit the mainstream, three months, six months, three years, the only way we're going to know and get there faster is going through this process of getting the technology in people's hands, getting the feedback and improving and iterating and anchoring back in that respect for the user. Make sure you're delivering on a trust value proposition.

And that's what we're seeking to do. what's the business model going to be here, do you think? There's a lot of speculation. Hey, Google Bard, you know, is going to be so good at giving you an answer. And Larry and Sergey said this early on, like, hey, well, no, we've done our job when you do a search and we just give you an answer, as opposed to the 10 blue links. And that's a journey Google's been on for a while. Hey, here's surfacing some information about a restaurant review or maybe a one box with a snippet of an answer or the best sentence in a blog post and then linking you to it. Or here's the weather.

Here's the stock price. You know, over the decades, it's just said, okay, what are the top searches? And let's just see if we can get people answers quicker. But of course, there's, you know, things like shopping or local that are filled with clickable ads. There's speculation now, hey, this is so good. Maybe people click on less ads. So is there a way for Barr to have ads in it or commercial messages that doesn't ruin the experience? And the same way Google has tons of ads all over the place that, you know, and I have

had this discussion with Larry at some point back in the day, like, what's the difference between like a great ad that solves your problem and the first blue link? It's like, it might not be much of a difference. Like it actually like the Volvo ad might do better than somebody's review of the Volvo and answering your question, right? So there is this kind of natural tension. A great ad could be an answer. A great answer might not, you know, might be an advertisement for the product. The line blurs a little bit. So how are you thinking about advertising? And, and, monetization here. This is something you think people will pay for.

You've never had a paid product, really. I mean, maybe Google Docs. How does Google think about the business model here? It goes back to the first principles again. Respect the user and respect the opportunity. Respect the opportunity. We succeed when others succeed. I spent 10 years of my career working in advertising,

some at Google, some of other companies like Pandora. And why is advertising valuable? advertising is valuable because it helps grow awareness of your product. It helps drive conversions. It's something that Google has been very focused on in respect and respect for the opportunity. And so as we're in this seat of respecting the user, we're still learning how to make these tools the most helpful for people. We're seeing them being used as, hey, this thing's a collaborator.

I have an idea. I can build. I can grow. I can compound on it. I have this idea for a character I want to write into a YouTube film that I'm going to do. I can build, grow, expand. We need to keep going through those and inviting the conversation. So to answer your question directly, what's the monetization plan?

The monetization plan for anything we do at Google follows the same rubric. Look to where you're generating value for people and make sure that the way that you monetize falls in light with that. And so that's how, that's how we're going to approach it. We're focused on driving the value and respecting the opportunity means that when we're, we're ready to fold that in, it'll be folded in. Listen, you know, I'm a bulldog guy. I've had bulldogs my whole life. But you know what? Dog owners, we're spoiled. We have so many dog sitting, dog walking options, right? What about cat people? You know those cat people in your life? Well, I've got the perfect solution for them. Meow tell. Meow tell.

is the number one cat sitting app. And it's available on iOS and Android. And MeowTel sitters have over 60,000 five-star reviews from happy cat parents. And you know, these cat people, they're very demanding. They're not giving five stars out easily. All these sitters are fully vetted and insured. And they offer once or twice daily drop-in visits or overnight stays in your home so that your cats can enjoy playtime and cuddles in the comfort of their regular environment.

A meeting greet is included with each reservation so you can ensure the perfect fit. And your sitter will send detailed updates and photos to you each visit to let you know how your cats are doing. So you get that security, right? It feels great when you get the photo. You're like, ah, now I can enjoy my vacation in Italy or going skiing in Tahoe. Miao Tell provides a 100% satisfaction guarantee and a team of support specialists to help make sure you have the perfect experience. Use the code twist 25 for 25 bucks off your first booking. Okay, it's 25, Two tens and a five, waiting fee. Code Twist 25 for $25 off.

I mean, just looking at this search, I just did like, hey, what are some flights options to Tokyo, nonstop flights, you know, and Google flights, prints money for Google, tons of paid links in there. But it's also the best for me. It's like the fastest, best way for me to find the least arduous painful flight, which I define as like the shortest time, you know, less. layovers and a really great answer here in terms of like, hey, here are the nonstop flights, you know, Delta Hawaiian Airlines, whatever, or Canada and the flight times. But it doesn't incorporate Google flights yet or travel, whatever that's called, that gives you like the flight information. And but it gives me some like, hey, you might want to use this flight search engine or

consider a budget airline, some stuff I didn't ask for. But here, having links at the bottom to, you know, offers, here's a $25. off or here's a link to whatever airline, you know, those would be quite valuable to me. And I wouldn't mind seeing them in here. And I would probably click on some of them. So is that like you think a potential path? And have you started to have those discussions? The potential paths are endless, which is part of what makes this so fun. Like, I'm going to go back to like, why is it infectious to be at work? Like, yeah, there is no shortage of exciting ideas to run toward and build and execute. And so one of the things that we are excited about is this

notion of bringing tool use into art, how can you get it to a point where as you have interactions like this and kayak, one of the partners that we're talking about bringing to the full, how do you make sure that the experiences that you're getting are trustworthy, are reliable, are getting there? And that's where you need to focus now, where we need to focus now. Making sure that the responses that you're getting are helpful and can be compounded on. Because, like, I'm assuming you maybe have someone that helps you book your travel. I do it myself now, but I did have that for a while.

I like doing it myself now because I'm so good at it and the tools are so good. Well, in the part of when you had someone to help you with your travel, if you said, get me a flight from point A to point B, how many times would that person come back to you with the perfect answer and the perfect itinerary? Yeah, it depends on the person I've had, you know, the person I have now who will do it sometimes for me, gets it almost perfect every time after years of working together and knowing I like Google, I like that, Jepp Blument, you know, I want to get my United Miles, let's use my United Miles.

Yeah, no, no, but there is a whole like heuristics of what I like, right? So I think what you're getting at is, hey, Google Bard's going to understand that. And maybe if I like kayak or I like Google flights or whatever, or if I have preferred airlines, that's all going to be known by my Bard, by my Google Assistant. Look, to project where we're going to be in end months of time would not be a responsible respect for the user, right? Like, we are focused right now on what is the feedback that we're getting that we need to improve upon? Like, one of the things that we heard, why did images just come out a couple of days ago? It came out because as we launched, people were starting to say, like, hey, this is a, like I'm talking to it like I'm talking to a teammate that's helping me think through an idea.

and I want to visualize these things sometimes. But like Memorial Day is coming up. I'm guessing this is going to happen after Memorial Day. We're all having barbecues. Well, maybe not everyone, but many people are going to have barbecues. As you dream about that long weekend that you're about to engage in, like, having a little bit of extra visual stimulus makes it a little bit more fun and more enjoyable. And so that's why you get something like images and responses.

And so we're going to continue to build compound and just keep listening to how people are responding, what they want to do, what they want to grow in iterate. And of course we want to be helpful. You got tables now. That was a big one for me. Yeah.

I mean, I was like constantly doing search. I'm like, put that in a table. And it was like, yeah, I don't know. I'm a Google AI assistant.

I don't know if I can do that. And then one day, it's like, of course I'll put it in a table for you. It's like, well, and then it's obviously, hey,

put this in a Google sheet for me. And then it's, you know, now we're starting to get really interesting. Am I in the Bard interface? and then exporting to Google Sheets, or am I in Google Sheets doing barred searches, and then plugins, obviously,

you announced plugins. Explain to the audience, like, for a startup, I got a startup that's a, you know, MeowTal, and I'm doing babysitting for cats.

It's literally a cat sitting marketplace that we invested in called MeowTal. How would a Meowtow build a plugin for this? When are those going to be available? What's the roadmap there? Because you might really prefer a chat interface like this, this to navigate something like MeowTal or we have another marketplace called Gigster that is like locations for films and, you know, photo shoots.

An interface, like, all of those companies are going to start building their own interface on their website, but then it's like, wait a second, if Google has this, can we do a plugin, and then can that plugin exist on my site the way Google ads or Google search? You had a product back in the day where you could do Google site search. So talk to me about the vision Google has for plugins and for startups to get to leverage this technology because we're here on this week in startups. Yeah, I mean, the developer ecosystem is one that has been core to Google for many years. There's obviously Android, Chrome has an extension framework. And part of as we're building, so we're calling tool use and barred as extensions because we feel like there are some similarities to what you've seen with Chrome.

extension where, hey, I have this idea for an interface that can be helpful to help you interact with the mode that you're in. And so we want to continue to build and grow and evolve and listen to how people want to interact in a conversational way. And again, the way that we're building it right now is building all that core principle. We only succeed when others succeed. So we need to make sure that what we deliver is going to be helpful. and the initial set of tool use extension partners that we have

are really focused on, A, getting it right as you bring it to users, and B, making sure that it is helpful for the people that are trying it out. So we're thankful to have partners like Instacart that are joining us. Replit is coming soon. I've been playing with Replit. I basically did a whole thing where I was like,

I started their Python like 100-day course. So Replit, you're going to be able to have barred inside a replet? or you do a bard and then export to replet? You're going to be able to export to replet. So again, code, one of those things that people love to use on the BART experience. And so we heard, again, listening to people, hey, like, this is a good way to get me started, but I want to continue. I want to grow.

I want to develop. Let me export this to Reply. Can we start up and build a plug in now, or is that not open to everybody yet? We're certainly excited for people's ideas. We launched a private preview of our Palm API, which is also available for startups to sign up for. So certainly you want to go and visit that. And so that's not where I spend. My personal time on the Palm API, I have a phenomenal set of colleagues that are driving that.

But we're looking to engage the developer community. So limited access right now and we'll open up later. Maybe you could do some demos first here. We got you here and show us some stuff you're working. Yeah, demo or die, as we say in the valley. Demo or die. All right. I'm going to start to pull up my screen here.

By the way, somebody got to talk to Sundar for me. I loved when you guys had made your own Chrome books. I know that it's like austerity measures right now and like can't do everything, but I love Google Chrome. And man, I wish you guys would still make that pixel book. Man, that's a great product. Well, hey, Jason, I appreciate you being, I appreciate you being a fan of Google. You've got the infection enthusiasm that we're feeling on a daily basis working on Barr. So I want to walk you through a little bit of a story of how I'm using Bard personally.

So I'm one of those weird people that goes on a walk in my neighborhood. And I got my headphones in. I got my pixel buds in. And I just start talking and thinking out loud. And normally it's been a phone call that I call to someone that I just want to riff with. And I was working with a teammate and said, what if I could riff with Bard on some ideas? And so I'll spare you the recording of me blabbering, walking down a bike path, thinking out loud of how to prepare. But what you see on the screen here is something I use my pixel recorder for where I just said, hey, I'm going on Jason's podcast tomorrow.

It's called this week and start up. See what's to see Bart do some cool things. Here's some things I want to highlight, like how we're doing coding, how we're doing this, that, and the third. And then I finished the recording. It was two minutes long. Two minutes. Two minutes will be just talking out loud. And Pixel Recorder has this cool feature where it creates a transcript that you can export to Google Docs.

So I exported to Google Docs, finished my walk, got a lot of other thoughts done. I regret to inform. I'm thinking of other things other than being on this podcast, but I'm very excited. I'm very excited to be here. Yeah. Yeah. Yeah. Hey, Memorial Day is coming up, right?

And so what I started doing is asking Bard to summarize my free form thoughts. So I just thought out loud. Can you actually like mold it into something interesting? And so I'm going to scroll here and Bard started to give me ideas of demos that we could do. So what I'm going to read out loud here, I know not everybody watches this on YouTube. We might be streaming, not looking at the screen. Here's three demos that we could do. Generating a list of hot dog recipes from different cities.

So you got friends from different cities, they might have something that's really interesting. Chicago ones, yeah. Yeah, Chicago, New York, Tokyo, London, whatever. You know, we can go global here. Yep. One of my personal favorite use cases is, so I love hip hop. I think I've heard you talk about hip hop as well. Yeah, you're an East Coast guy, West Coast guy.

What are you, old school? I'm an East Coast guy. We're going to get in on my top five here, Jason. I love it. And then we're going to go back and forth. And then debugging a regular expression. So I talked about that a little bit. Sure.

And so Bard gave me these ideas. And I think, like, that's kind of cool. But then I wanted to see, like, All right, well, and then it tells me why it thinks that it's a good demo. These are just a few demos that you do on the podcast to highlight Bard's ability to access and process information with web, generate content, debunk code, and endless possibilities for what he could show off. That's, I feel that.

Thank you, Bard. Bart's got me. All right. So, I asked Bard, all right, let's do a two-minute strip for each of these. We don't have to go through each of these, but like, for what it's worth, my top five, Nas, the late-backed. Lipsy Hustle. Respect.

Plus arrives. Okay. Jay Z. Of course. Biggie Smalls. The goat. So I want, I want Bart to back me up. So, L.A., you've got the West Coast.

I live on the West Coast now. But four out of five, I'm going east coast. So I asked Bard to back me up on it. And then on the first one, I wanted to make sure it was munching an example of cities coming together. And so what we're trying to highlight here is make sure that we can show that it's not just about retrieving information, it's about synthesizing it and generating new knowledge through these forms of synthesis. So I think this might be a little bit fun based off of what Bart did. I don't know if you can read my screen. Maybe we can go a little back and forth.

Yeah. It created a script for us. Jason, you can play the role of host, and I will play the role of Bart here. So what do you think? Let's do it. Welcome back to the podcast, everybody. Today we're talking with Bard, a large language model from Google AI. BART is a trained and a massive data set of text and code and you can generate text, translate languages, write different kinds of creative content coming at you, 97.5, and answer your questions in an informative way, 72 and sunny Santa Monica. I love it. Cheers for fear, everybody wants to rule the world.

I just add lived a little bit there. That's beautiful. That's beautiful. I was surprised I didn't go with Hot 97, but, you know, we got it. It's my morning zoo. There we go. There we go. And so, you know, Bard gives a response here.

Hi, I'm Bart. It's great to be here. Back to the host. Yeah, so Bard, I'm a big fan of hot dogs. Can you generate a list of hot dog recipes from different cities for me? Sure. Here are a few hot dog recipes from different cities. New York City.

I think I know someone from New York City. There you got. New York City. Hot dog is a classic. Made with the seam hot dog. Sourcrow, mustard, ketchup. Good thing. This is a visual show because there's an image on the screen here.

Perfect. A delush, delicious New York. Do you put sourcrown on your hot dog, Jason? Of course. Mustard sauerkraut. And also the onions. Yeah. Yeah, I'm Polish, so of course I put sargar out.

And then we go through, we go through Chicago dogs. We've got an L.A. dog, baking, grilled onions and avocado. Disaster. Right in the garbage. And then you got the London dog. I've never heard of this one. I never heard of that. Bates beans.

Ah, very interesting. Tokyo style. Yeah, yeah. So I think you've got, you know. I might mess with that Tokyo style with the Japanese mayonnaise and curry sauce. That's interesting. It looks delicious. All the good news is you can click on that.

You can click on that. Delicious. Thanks, Barr. You're welcome. And now let's see. Here you go. Last one. As you can see, Barr can only pull information from the web, but it can also generate new forms of knowledge via synthesis.

For example, Barr was able to combine the ingredients of different hot dog recipes from different cities to create a new list of hot dog recipes. It's just one example. The many ways, bars can be used to generate new knowledge. Yeah, so I'll spare everyone. Yeah. We can do this for hours. Yeah. But some of the things that I loved of the other demos that it provided here, like so,

For those that may not be fans of hip hop out there, everybody's got their top five list. I'd love to hear yours. For me, top five list, poof, I would go Biggie, JZ, public enemy, Tribe Call Quest, Beastie Boys. Love.

Yeah. It's just really old school. I love it. Hey, Beastie Boys are foundational. You wouldn't have Def Jam without the Beastie Boys. Exactly. Yeah. So, you know, you go through here and then some of the things, there are these nuances that I start to get really, uh, really inspired by like, just seeing.

So, again, I did have the great pleasure of getting a call Nipsey Hustle a little friend. He's no longer with us with senselessly murdered. But one of the things that I have really, you know, touched by is sort of this notion here of, I asked for the top five, Nas, bust arrives, Jay-Z, Biggie Smalls, all in the present. times is, speaks about an Epsie Hustle, he was a talented artist. And it's just one of those things, it's these nuances, these things are really hard to capture. It doesn't always get these things, right? But when it does, and I see this, it starts to like trigger this notion of, oh, man,

this can be amazing, this can compound. And so, yeah, it is a fun use case, I will say, if you've got a top five and you want to, you want to back it up and your friends are trash talking about it, yeah, Bard is a great backup. And so then Bard goes through debugging a regular expression. and I found that the regular expression debugging, like, it's fine, but I didn't really understand the example that it said here. It needed some more information.

So, like collaborating with another person, I just said, hey, I'm giving you some feedback. That regular expression example needs some explanation. And for those that aren't looking at the screen, Barg gave a demo of how to make a regular expression, check for a value of out. I mean, it's basically a how-to article, like this is now teaching you. And this is, I think, some of the most exciting. aspects of this. You did this two-minute stream of consciousness on your walk, all the things on your mind. It starts giving you, hey, here's a whole agenda. I'm going to try to answer as much as I can.

But the follow-ups are when you really start to learn something. And this is where augmenting human intelligence gets truly exciting because, you know, it's like, there's always something you don't understand, but maybe you're too embarrassed or you forget to ask the question. You're at sushi, hey, am I supposed to, what am I supposed to do with the ginger? How much wasabi am I supposed to put, do I put it in the soy sauce? Do I scrape my two chopsticks together, whatever? All these little nuances of just going to have sushi or whatever. You know, it starts to explain these things to you and you can really go deep and it's going to actually customize this stuff to you over time.

That's occurring now, or is Google tying all this learning to your account yet so that when you see another code snippet, it goes, okay, we are relatively sure he understands this concept. but, you know, Jack understands concept A, Jason understands concept B, but they might be weaker on these concepts. Is that happening yet or is that in the future? Again, the possibility is how to grow and build in of all this thing to get better. You know, they are, they do feel endless. Bart even said it. The possibilities are endless. Can't keep saying that, yes.

If I do say so myself. You know, for now. It's getting a little ahead of itself, right? The possibilities are endless. Yeah, the X. Oh, you have the export button there. That's great. So you can export it to Gmail or docs, yeah.

So in this case, right, like, you know, it explains the demo that it wrote for itself, explains how to verify this. And let's say you and I were having a conversation we were going back and forth and I wanted to follow up with you. Then I would just ask it to draft in Gmail. Goes, all right, new email created, open it in Gmail. And so now I'll share a new tab. Yeah, that's when it starts getting exciting. formatting it perfectly in email.

You know, this is the kind of stuff that would take hours, you know, or 20 minutes to get it perfectly formatted. And now it's all going to be done. A lot of the manipulation presentation layer of data is, you know, so easy for this to do. Just, you know, taking the data we were talking about before on flights and putting it into a table is awesome. Now, you had the button to export it to Google Docs. I don't have that yet. Is that not out yet? Are you on some sort of early thing?

or is that rolling out? That has been out, so you should be in a little. You're not seeing it? No. We're going to have to deeper. This is. Maybe, you know what, I'm not on my Gmail account. I'm on my docs account.

So I wonder if the docs. Yeah, I think sometimes with Google, if you have a Google docs, like your custom domain, you get like maybe releases a little bit behind the at Gmail stuff, which are considered the Google accounts. So we release on at the beginning of. the ability to enable BART for enterprise accounts. Thank you. And so we want to make sure that as we roll these features out to our enterprise customers, that, again, there is that element of making sure that the information that's flowing

is very clear and seamless. And so you can see that grow and iterate, but I really appreciate you enable it. Yeah. And listen, thanks for the great work. I want to keep having you come on. I think this is a great moment for Google to get out there and communicate with the world. I know big companies, sometimes they perceive there's a downside to talking about stuff because you get big. Oh my God, there's gotcha journalism going on.

People are trying to, you know, whatever. But here on this program, we just want to talk about your progress and the features and really, you know, the opportunity here. And I'm particularly interested in how you interface with startups. So maybe, you know, in two or three months, when you got the startup stuff and the plugins done, we can do a deep dive on plugins. and how startups can really engage with this because I think that's the most important part. You don't have the time, you know, to do everything. And I think one of the things Google's done well

is to really collaborate with startups, send traffic to them. You know, listen, everybody's in competition on the margins. But I think there's such an amazing opportunity here for Google to support startups and to enable technology in startups with the barred platform. And so I'm really excited about that possibility. You're touching on this thing. It's like, it's this feeling.

that we have right now, right? Like, mobile came out and like, you know, it's from the launch of the iPhone and Android operating system, right? Like, apps didn't come immediately, but you can start to feel it. Like, it felt different. You know, the internet started rolling out.

It started to feel different. This feels different. Yeah. I mean, I don't, I mean, you've been in this business for 20, 30 years like me. I've never seen a pace like this. Like the mobile development pace, the broadband pace, the web base, cloud pace,

all those paces were brisk and, you know, felt at times like we were, sprinting. But this is different. I mean, it's every day, you know, every 48 hours, two or three really impressive demos coming out from everybody. And it feels like we're really going to reset how we live and how we get things done in the same way that putting a computer on everybody's desktop, Microsoft's original mission, or Bill Gates is original. Hey, put a desk, you know, computer on everybody's desktop and then Google's, hey, index the world's information. Now, it's just making everybody a superhero. I just love the fact that I,

I can now become a designer or a coder where I might have felt like, yeah, you know, I'm not good enough at design. I'm enough to get myself in trouble. Oh, it's just too big of a bridge for me to become a coder. I feel like I could be a coder again, you know, like go back to like the 80s and 90s when I was doing Pascal and Basic. And I'm like, well, why can't I do Python coding? It doesn't seem that hard when you have this assistant, you know?

It's removing friction to creativity. It's amazing. It's magical. And it's a time to be bold and to be responsible, right? Like, we talked about it at a ton. you can release things to generate that speed of insight, but you can't generate that speed of insight at the expense of trust. And so hopefully that's what you're seeing from us.

I'm super thankful. Yeah, I mean, it's pretty clear you're containing this. I mean, I don't see the ability in here to connect to Google voice and start calling people up on the phone. You know, like when these things jump out of the browser and they can start emailing people, which they're going to be able to. But right now it's like, hey, export this to email. You've still got to click the send button.

You know, Google's going to be very thoughtful about it's not going to let you go, hey, you know, make me a list of 100 people and then say, oh, yeah, now email 100 people. And now let me make some malicious code and then email that to 100 people and then, yeah, try to hack their address books. Like, you're going to red team that. You're going to red team the heck out of the letting BARD email somebody, letting BARD use Google Voice or, you know, connect to a phone system. Those things, when it jumps out, you're going to have to be super. thoughtful about. Absolutely. Those discussions have already happened, I'm sure.

They've happened. They're happening. They will happen. It's the commitment to being responsible. If you don't build with trust in mind at the beginning, this won't be a paradigm shift. There's no Gmail hookup now to send Gmail, right? As in one example, that's going to be like, and Gmail already does this. Gmail's got a throttle.

If you open a Gmail account and it's under X number of days old, it's not letting you send the thousand emails. You have to warm up that account. You've got to use it for a little while. And the second it sees you send 300. I think maybe the limit on Gmail is 500 outbound a day for a normal Gmail account. I can't remember. And so there's already throttles in there.

And then you put AI plus Gmail. Yeah, you can come up with like a spamming super scenario. But Google's not going to let you do that. They've seen this movie before. It's imperative to be responsible. But it's imperative to be bold. It's exciting. Like, you know, we are sending everybody the email.

for my birthday party, yeah, maybe I want to do that. And then Google's going to have to determine, like, are you emailing people you've traded emails with before or people you haven't?

Let's be thoughtful about this. Yeah. That's why you build with core principles. Yeah. User control. Transparency. We'll get there. Jack,

thanks for coming on the program. I know it's a lot to even get like the media permissions when you're at these big companies. Appreciate you coming on being so honest. I'll get the wavy-davy story in the next. episode of All In.

So everybody was going to, I'm going to do a cliffhanger here for episode 131 to get the Wavy Davy stories. Continued success. Thanks for talking about. Thank you. Just internally, my message to the team at Google is, yeah, get out there and talk. Like you were when it was a, you know, a pre-public company, like a three or four year company. I'm telling all these stories about, you know, talking to Larry and Sergey, they were out.

They were at events. They were talking to folks. They were showing the product. They were taking feedback. You know, Google's got to get back to that. They got to get back out there, get on the streets, talk to everybody, share the experiments, and get the startup community building with you. So that's my hope for Google is the comms team really sees how well this went and just get out there and talk more and engage with the community. We're like, it's an amazing moment.

It's also a fun moment. Yeah. And these conversations are fun, man. It's been fun, riffing with you. It's been fun, Jim. And, uh, I appreciate it. I'm holding you up to that. I'm holding you up to that return invite, man.

Oh, no, three months. I think two or three months, whenever you've got. The really, I got my pulse on the startup scene. Every startup is like, how do I get a plugin on Bard? How do I integrate it onto my website? That's the golden opportunity. And I think that's the second you have some thoughts on that. And, you know, if there's somebody I could talk to on the back end, you know, who does plug-ins to just put a couple of my companies into it, to kind of start giving you feedback.

Like I mentioned two marketplaces, Gigster and Meowatel. I know specifically I want those places. I mean, when you're interfacing with a marketplace, you're like, tell me. me, I want a sci-fi location in L.A. Can you show me some pictures, put them in a table? They're going to either build that technology themselves from scratch, or they're going to use a platform like yours. And then you're going to want people in BART to be able to interface with it.

So, man, the possibilities are mind-blowing. And we only succeed when others succeed. It's- Love it. All right, everybody. It's going to happen. Thank you. We'll see you next time on this weekend starts. Bye-bye.

All right. What a great interview with Google. Thanks, Jack, coming on the program. And we'll have you back in a couple of months to talk more. about Google's progress. Obviously, a lot of entrepreneurs are really excited to participate with Google and their tools and services. It's great that we now have multiple options, right? You have obviously Bing and Google going at it now with their products, Quora's in the game. Hugging Face is going to come on the program in a couple of months over the summer probably and great partnerships

there with AWS. So we're going to have many different options for founders to leverage these different models, open source, or in partnership with the major players. Or I thought I would just wrap the show here with my favorite TV show over the last couple years, secession. If you haven't seen the show and you don't want to talk about the spoilers, you can tune out now or you can save this episode. But I thought I would just take a second to talk about this incredible show because it is about business.

It's about family. It's about motivation and so many different things. And it's so well done. The writers and the actors, directors, everybody involved in that program deserves a lot of kudos. It just shows when you have a collaborative effort and everybody is firing on all cylinders from the writers to the actors and everybody involved in directing, recording, cinematography, locations. Just a tour to force and over in only four seasons, which is a very unique decision, leaving you wanting much more, I think, is how everybody feels about it. But I thought they stuck the landing. I really enjoyed the finale and it was long.

It was like a 90-minute finale, which was nice. But some takeaways from this, because you're left to wonder, what is this about? At the end of the day, what is this show about? And I think, you know, the show in some ways is about business. And in other ways, it's about family. I believe at its core, it's about the Murdox and Rupert Murdoch, full stop. I know they say they based some other characters or Logan Roy is based on some other real world people. But it's really about the Murdox.

And if you look into the history of News Corporation and how they grew from Australia to the UK, the U.S. It really parallels this. And, you know, he worked any edge, Rupert Murdoch, and obviously about the Wall Street Journal, but if you go back in history into the 2010s with the phone scandal in the UK where his

corporations hacked the phones of police and celebrities and murder victims, and then you look about, look at Fox News and the relationship between Fox News and Trump and putting him in. So, so many scandals, so many questionable

decisions. And I really think that the writers had something to say about him specifically, Fox News and what led up to what I think a lot of people on the left feel is an existential crisis for America. And I think that's where this show is really the origin of it. People will talk about the kids and Shakespeare and all the stuff. I get the sense that they were obsessed with the story of Rupert Murdoch and the story of Donald Trump. And that's why the show, crescendos with Jared Mankin, basically a person that Roman frames as, you know, a great person for them to ride into the ratings, which is, I think, how we wound up seeing Fox News look at Trump. It was great for ratings.

And that revelation that the media industry cares, or some portion of the media industry, specifically cable news, news on television, they really care about ratings and who is best for ratings and how to keep people tuned in. And the writers are obsessed on that, I think, on this show. and the line that wins Logan over on Mencken, if you remember, is Roman, who is channeling his father, said dad, he's box office, right? You know, that is, I think, at the core of this is they really, really cared about winning. And that goes back to, if you go back to Logan Roy and you look at him as a character, at the end of the show, at his funeral, you find out how he was developed. What a masterstroke in letting us wait until the end to realize Logan Roy,

you know, in escaping the Nazis, had to be in the hull of a boat, looking at his brother at the age of five and told, don't speak or else the U-Boats are going to get you, and you're going to be killed by the Nazis in three days, you know, silent in the hull of a ship, trying to escape, and then perhaps bringing polio back and killing his sister and his aunt and uncle, I think we're taking care of him as foster parents essentially telling him, hey, listen, you know, you probably did kill your sister. And you realize a lot of the great entrepreneurs, maniacal people, extreme people, they come from this kind of background. And the opposite of that background is

Logan's kids, who the mother was incredibly cold and distant. They were raised by nannies. They were raised in privilege in stark contrast to their father who, you know, escape the Nazis. And so you have a broken man who creates an empire. And then you have a bunch of ignored children who are unable to manage that empire, but who are diluted into thinking that they're the same stock as their father. This insane person who would do whatever it took to win, and had a drive in him that at the end, Kendall says he hopes he has. He hopes he has that fire. And at the end, he does have that fire. But of course, that fire is too much. And he's not a serious person, as his father alluded to. Another great business lesson, Logan Roy was serious about business.

And he surrounded himself with, you know, even though he was abusive towards them, seemed to be abusive towards everybody since he was so abused as a child. And fans of trauma cycles can, I guess, go deep on that. with the show. Again, the writers did an amazing job in architecting all this. But from our perspective here on this week in startups, thinking about how he ran his business. He had Jerry, legal, Carl, I think it was a CFO, Frank was the C-O-O-I-I-thing, operations guy, Carolyn in PR, and of course Tom, Wabsganz of Triple Play fame. You can look that up. You see those people working tirelessly trying to keep this empire really chugging along. And that's in stark contrast to

these three kids who are numskills. And they, the kids, think that they are incredible business geniuses when in fact, none of them have ever done meaningful work, zero meaningful work. And until their father dies, they're not in charge of anything, essentially. And all of those serious people look at the kids as dope. They're kind to them, but they look at them like, ah, these kids are not ready for prime time. They're not serious people, as their father told them. And as Roman sums up, in the end, it's all bullshit. They're bullshit. And that really is the truth. These were not, never serious people. I don't believe this show is about kids. I think it's more about Murdoch and Trump. But I'll take from it, what I want to take from it, which is business and the business

lessons here. Logan Roy, he had an incredible team around him, but he didn't have a moral compass. He was addicted to winning. He didn't think about the repercussions of what he built. It was never enough. And he was never happy. His relationships were all toxic from top to bottom. And he had, He had brief moments of joy that were singing some limericks, you know, just very short, short moments of joy in his life. And most of his life was toxic up until the end. And he died essentially alone. The people around him were cracking jokes about his death within, before the plane had even

landed. And it was just business as usual for him. He was just a figurehead. He was this force of nature. But he didn't have friends. He didn't build meaningful relationships with his family. And so he was an animal. He was a beast.

Not a great life, actually, if you look back on it. Pretty meaningless life. In fact, you know, you can run up the scorecard to billions of dollars. I know people have doesn't mean that you're going to get happy. And you can have all the business victories you want. If you don't have meaningful relationships with your friends, your family, your coworkers, your kids, it's all for not. It should be a lesson to anybody who's in business. And then if you're going to look at it as a, you've got to really enjoy the ride, right?

And I don't get the sense that there was any joy in this ride for him. It was just maximizing profits, maximizing power, trying to buy the next Pearson or New York Times, Wall Street Journal, whatever the metaphor is, trying to do the next deal, trying to win the next battle. It was never about the actual joy of building some enterprise or some product. It was just pure, unadulterated, chasing of power and money. That's it. It was vapid. It was meaningless at the end of the day.

He died, dropping his phone into a toilet. it alone. Family didn't love him. They had a perverse relationship. And his family didn't even know his own story, which is why in that incredible scene, you gotta give the kid who played Roman his Emmy now. Man, what a performance.

Kieran Kalkin, give that kid the Emmy, as great as the kid who plays Kendall is. Kieran Kocken just, that scene where he gets crushed me when he says, is my dad in there? Can we get him out? That basically summed it up. He didn't know his father.

until his brother just absolutely told those couple of handful of stories about him. Great series. Can't wait to watch it again. Can't remember being so excited for each episode of a series. I think it's going to be one of these amazing series that comments on late stage capitalism or late stage media, perhaps better. And we've talked about the media business here on this podcast and all in. And I've written about it since I grew up for the last 30 years in it.

This is about late stage media. The empires that were built are crumbling now. They're broken. They're chasing things that they shouldn't be chasing like ratings. And, you know, the kids who are going to take them over are qualified to do so. There were a bunch of clowns. These kids were a bunch of clowns. I guess there's some lessons in parenting here, too.

He could have had meaningful relationships to these kids. He could have put them to work. Seems to have put Tom to work and mentored him as a son in some ways. Kind of paradoxical that he didn't do it with his own kids. Couldn't put them to work. Couldn't build meaningful relationships with them with all of those opportunities. He insisted on putting them in top positions instead of having them work their way up. Just a failed parent across the board.

So, great show. I'm going to watch it again next year, maybe over the summer or something. I think it's going to be just one of these great shows like The Sopranos, the Shield, The Liar, that you could really rewatch them over the decades. Congratulations to the team. I hope they do a movie. I think it's perfectly set up for the movie. Just my thoughts on the ending.

I think Roman is free. I don't think he's got closure. I don't think he wants to run the business actually. He just wanted his dad to love him. His dad's gone, he got the money, and now he can move on with his life. I think that's the significance of his smirk at the end, looking at his drink.

And for Kendall, I think he's broken and he's going to try to wrestle that company back or start another one if he doesn't kill himself or go on a bender on drugs and overdose. And then for Shiv, man, what a paradoxical, insane statement that was about a woman who now finds her herself. In the same situation, her mother was in and she's, her husband is running this thing

and she has some amount of power, but she's kind of trapped in all of this. And she's just a woman on the sidelines. They really smashed to those kids. But I think Roman actually comes out the cleanest of them all. I think he realizes he never wanted it. And Kendall is crushed for life and Shiv is crushed. One out of three will probably move on to do great things in their life and maybe have some peace. And for the oldest son, Connor, he seems to have, uh, have, uh, have reconciled a decent amount of this. All right, anyway, those are my thoughts on it. Happy to hear yours.

You know how to reach me on Twitter, whatever. And we'll see you all next time. Bye-bye.