Tim Gasper [00:00:01] Hello everyone. Welcome. It's time once again for Catalog& Cocktails. It's your honest, no BS, non- salesy conversation about enterprise data management presented by data. world with tasty beverages in hand. I'm Tim Gasper, longtime data nerd, product guy, customer guy at data. world, joined by co- host Juan Sequeda. Hey Juan.
Juan Sequeda [00:00:19] Hey Tim. I'm Juan Sequeda, principal scientist at data. world and we are back. It is 2024, it is season seven, episode 160, 78. I don't know, I lost cat now. But it is super exciting to be back here and we're going to kick off this year with a fantastic conversation. I mean, obviously, AI's on everybody's mind and I think one of the people I've been following the most and just really enjoying all his content and all the events that he's been doing is Jeremiah Owyang, he's the general partner for Blitzscaling Ventures. Jeremiah, how are you doing?
Jeremiah Owyang [00:00:53] Hey everybody, so glad to be here. Greetings from the tiny Airstream Studios in the heart of Silicon Valley. It is raining today so you might hear a little bit of pitter- patter on the aluminum roof.
Juan Sequeda [00:01:04] Well, you're in the Bay Area. I am today in LA and Tim is in Austin, so we're all in different parts of the country. So well, let's just kick it off right now and tell a toast. What are we drinking and what are we toasting for today, Jeremiah?
Jeremiah Owyang [00:01:20] Me, I'm a healthy guy. I'm on a health kick, so I'm drinking green tea.
Juan Sequeda [00:01:26] What are you toasting for today? What do you want to toast for?
Jeremiah Owyang [00:01:30] Let's use technology to fix the problems and woes of the world.
Juan Sequeda [00:01:36] Amen to that. Well, I was looking for a cocktail even though it's two o'clock here, but the bar's closed so I'm just drinking a lot of water today, so cheers to that. I'll cheers to that, Jeremiah. How about you Tim?
Tim Gasper [00:01:49] I am in the Austin office of data. world right now and actually we're in the process of switching offices right now, so we've aggregated all our alcohol together into one cart and I found something that I really thought looked good. I grabbed the Bunnahabhain inaudible here, so I've got a little bit of that going. I'll cheers to what you said, Jeremiah, let's use technology to make a difference. That's a big deal. Let's make that happen in 2024.
Jeremiah Owyang [00:02:13] Clink.
Juan Sequeda [00:02:13] All right, cheers, cheers. So we've got our warmup question today, so we're starting off the year 2024, but what are you least excited about in 2024?
Jeremiah Owyang [00:02:24] Well, probably the election cycle is going to get nuts and AI will play a role where we will see misinformation created by AI and we'll see people use AI to try to clean that up and find the source of information. So it's a little bit dreadful, but at the same time we'll see new technologies unlock things we've never seen before, so get ready for that.
Tim Gasper [00:02:47] I know that's going to be huge, big changes and everything's going to be at 11.
Juan Sequeda [00:02:52] Yeah, how about you, Tim?
Tim Gasper [00:02:54] You know what, for me, I'll say that I'm excited about my Cleveland Browns actually going into the playoffs, and so what I'm least excited about is the fairytale ending. I don't know when it's going to end. I hope it doesn't, but go Browns. They're my team.
Juan Sequeda [00:03:12] Well, my answer is going to be around the whole politics stuff in the election year and like, ugh. Anyways, but let's just keep that aside and let's kick it off with our-
Tim Gasper [00:03:19] You guys took it the serious way. I was trying to have fun here.
Juan Sequeda [00:03:24] No, but we got so much to talk about. All right, Jeremiah, honest, no BS. How are you seeing AI going to impact us this year? What is December 2024 going to look like?
Jeremiah Owyang [00:03:34] So Fast Company asked me to make a prediction and I wanted to give one that I think is very practical and pragmatic, one that everyone can relate to. Your email inbox, what a disaster. By the way, it seems like most people took, or many people, even in the West US, took two weeks off for the winter holidays. I mean, I saw a lot of people doing that, which is unusual. Europeans are like, " Yes, but that's what we do every year." But for Americans that's unusual. Now we're coming back to inboxes filled to the brims. I mean, yes, usually they're pretty bad, but it's even worse. So my prediction is that AI will help us get to inbox zero because all the data that it needs is right there. You can pre- train on sent mails. How did you filter emails? What did you delete? What did you respond to and which tone? So I'm expecting to see an AI agent emerge that takes the world by storm, just really takes the world by storm. You funnel all your emails in there, it reads all your other stuff, and it does the following features. First, it reviews everything, then it filters, summarizes the emails, and then it gives suggested replies. As we train it over time, eventually it's going to suggest email responses. So by this time next year, yeah, inbox zero due to AI.
Juan Sequeda [00:04:49] Look, I think we all want this. This brings back the question is like, do we need to have all this communication? I think that sometimes we just are communicating too much and I don't want all the emails zero. I want Slack zeros and all text messages and WhatsApps and all that stuff, right?
Jeremiah Owyang [00:05:06] That's really funny actually. We'll find each other on other places.
Juan Sequeda [00:05:11] Well, yeah, it will just go off and then I'm like, well, I want this also for my summary and LinkedIn, my LinkedIn feeds. Then at some point we're like, we are consuming all this information, all this, everything is coming to us. Then we're like, I want a summary of it and I want... Then we're going to generate a bunch of more summaries and then I want the summary of the summary and this is never going to go stop to the-
Jeremiah Owyang [00:05:30] Nest of summaries.
Juan Sequeda [00:05:31] Then at the end of this, I just want to disconnect from the technology and just go talk to people and face- to- face.
Jeremiah Owyang [00:05:39] So emails come in different types. Some of them are updates that you don't need to see. Some of them are spam. Some of them are just coordination. Some are follow- up. Then some are actually important. So we do need AI to at least do the first filter. So imagine if... I'm not an executive anymore, but I used to have an EA that would help me with those types of things. I don't have that now, but now everybody can with AI. I think that's what we need and I want to just underline that point, Juan. I agree, the point for AI is that we can be human again because we're not really humans right now. Look, we're stuck behind screens and the pandemic really screwed us over in many ways. So can we use AI to do all the tasks that we don't really want to do so we can have time to do and connect and make real business and personal lives and social lives and do fitness?
Tim Gasper [00:06:28] I think that's a really important point that you make there because when you first said, " Hey, inbox zero, email is going to turn into this thing that's happening automatically," my first inclination was like, wait, that sounds bad. That sounds like something where we are supposed to be doing the work of the communicating, but then when you think of it like, no, this is your EA that now is helping you, it's your assistant. What that means is that now instead of focusing on the emailing... The emailing wasn't the important part. The important part is now you're having the meeting with the person and you're connecting with them and you're having that conversation. That's not the part you're trying to replace.
Jeremiah Owyang [00:07:04] You're a customer executive, right, Tim?
Tim Gasper [00:07:06] Yes, customer officer.
Jeremiah Owyang [00:07:08] There you go, so the most important thing is the relationship and listening to customers. So if we could just get all the, quote, " paperwork emails" out of the way, " Hey, how are you doing?" schedule, all that stuff. As you understand, right, I think you guys get it, it's going to be their email AI agent talking to our AI email agent to get all that stuff out of the way. So yeah, that's the hope and promise.
Tim Gasper [00:07:30] Interesting.
Juan Sequeda [00:07:30] It's a T- shirt. We're always talking about T- shirts. The point of AI is to be human again. I love that.
Tim Gasper [00:07:37] Yeah, no, I love that. One thing that I wonder about this idea, before we move on to some other of your thoughts and predictions, is when things start to become more automated, I feel like that often spawns new industries and new problems. So now it's like, okay, automated agents talking to automated agents, how do you stand out? So now is there similar to search engine optimization, there's email bot optimization. Are new industries going to start to spawn because of this?
Jeremiah Owyang [00:08:04] It's starting. So my next event in February... I host an event series in Silicon Valley, San Francisco, called Llama Lounge. No physical llamas are there and it's not tied to the Llama model from Meta. It's just a clever alliteration. Ten different AI startups get to demo on the floor. So February event, I'm sure you can guess what the theme is, it's AI agents. There's going to be 10 AI agents demoing on the floor. So the SEO industry, search engine optimization, and search engine marketing industry and the advertising industry is freaking out. I mean not to mention New York Times and media companies and politicians and regulars. But they're freaking out because they see that users, consumers, are going to websites like ChatGPT or Perplexity, which is growing really quickly, to go find and get information on what to buy. These tools are making recommendations to consumers and to buyers on what they should actually purchase. It's not clear where their data sources are always coming from. It's often a black box. The algorithms not clear. Not that Google search is open source algorithm either, it's not. It's also a black box. So we are seeing people trying to figure out how to unwind that. So yeah, that is going to be a whole nother industry.
Juan Sequeda [00:09:21] So we were talking about this before, is the term agent, AI agent is something that we're starting to hear now in the last couple of months. I mean AI, generative AI, has come around in the last, what is now 14 months since ChatGPT. Then everything has been AI, generative AI, rag, rag all over the place. Then the last couple of months, I mean that's my perception now, is that we're now hearing more about agents. Agents is a very well- defined thing in artificial intelligence. I mean the academic, I call this the good old- fashioned AI, has been talking about working with agents and systems of planning and so forth. What's your definition of AI agent? Now you're seeing all these companies talking about agents, you're hosting this thing, what is an AI agent?
Jeremiah Owyang [00:10:05] All right, so you are a data scientist or you are a computer science professional. I am not. So I come from the business side. So we should actually discuss and debate openly for the betterment here of what is an AI agent. I see my LinkedIn friends like Ethan are here checking this out. So welcome everybody who's joining in from LinkedIn and checking this out. So let's talk about it. So the way I'm seeing the AI agent market evolve, there's around 300 AI agent startups right now. About 100 of them are in San Francisco, which is truly the cradle of AI. They're mostly here and most of them are in SHACK15 and in the Ferry Building. So I'm getting to meet the community and talk to them. An AI agent, I don't think there's a great definition, but essentially it operates autonomously with little human oversight. Operates autonomously. Autonomous is the key word. Operates autonomously with little human oversight. Right now, the best known example is AgentGPT, which is no relation to OpenAI. You can demo out that tool and it can do five cycles for you at no cost. For example, I did a demo on a speech today. I said, " Book a trip to Bahamas." It would say, " Here's what you need to know. Here's the passport. Here's the airlines available, here's the hotels." It started to go through these things and simulated booking those things just off one sentence prompt. Now in the future, these AI agents, you won't need those prompts as much because it should look in the context in my calendar, in my email and my social and know what's going on and it's operating. We'll have all these agents, multiple agents, around us doing these things, whether it's for personal or for work. For example, Tim, you will have agents, you'll have account development rep agents that will be reaching out to customers and prospects and understanding what they're doing and pinging them. I already see many AI agent sales startups emerging. That's already happening. So that's an example of that impacting you. I already gave the email agent that would be operating autonomously on my behalf. So Juan, I would love for you to respond, retort, debate on that definition.
Juan Sequeda [00:12:17] I don't think there's much of a debate because I am in agreement with you. I think I would extend it more from a technical perspective. But I think first of all, the autonomous is clear. You want this, it's really a computer program that can by itself figure out the steps of what it could do to go solve a problem. It really needs to be able to go perceive from the environment and that's how it knows it needs to go do one thing or the other thing and it should do that. So I think the perception of what's happening in your environment is very, very important and that's what can make it seem more intelligent than others. I think that's how we're going to start seeing more of these types of how my agent is better than your agent or how we start to... They can be more and more intelligent. I think effectively the term is actually you want to have an intelligent agent because you can go create any... I mean, you can hardcode a bunch of if then rules and that can be technically an agent, but it only knows that particular world that you defined, but you want one that can be really adapt and changed and so forth. I think that's where I'm really excited about what we're going to be seeing with all the AI applications coming out. My caution out there and what I'm telling to everybody is there's so much work that has happened in this space from a computer science perspective for decades and decades, let's just learn from our history because we need to build on the shoulders of giants and not reinvent the wheel so we can advance faster. I want to have that email agent as fast as possible. Because if we look at it, we've been having spam filters and so forth already, that in a way is some sort of an agent. It's kind of intelligent but not anymore.
Jeremiah Owyang [00:14:04] But it had to be trained, right, off our rules, our business rules, right, which is slightly different?
Juan Sequeda [00:14:08] Yeah.
Jeremiah Owyang [00:14:09] I think the difference here... I love that word that you brought up perception, perceives, that's right, so it can sense. But I think there's one other thing that can enable this to happen faster is that there's an network effect from peer to peer, right? So like Tim, if you go use an agent to go book a trip to The Bahamas, that data, that click path through United or Virgin Airlines, whatever, and then to Hilton and Marriott, whatever, that click path should be recorded. So when Juan needs to book his trip, that click path's already noted. We should be able to share those things peer to peer. I mean that's what you guys do data, right?
Tim Gasper [00:14:40] That makes sense.
Jeremiah Owyang [00:14:44] Can I just pick up on the comments from Ethan?
Juan Sequeda [00:14:46] Yes, please, please.
Jeremiah Owyang [00:14:47] Ethan's a smart guy. He been a digital exec in retail and media. I've known him for a long time. So he says, " I like the nuance of agency versus agent." I've been using them interchangeably. But I think that's right, the agency is like, what would you say, the description of the ability to do things. I think that's right.
Juan Sequeda [00:15:07] Yeah, that's an interesting perspective. You're thinking about the agency versus agent.
Jeremiah Owyang [00:15:12] Yeah, it has agency like an EA, right? An executive assistant or personal assistant has agency to do things within these parameters and that's what we're missing.
Tim Gasper [00:15:21] That connects to a little bit of what you said around it's more than just if this, than that. There's at least one level of additional choice that you're hoping to empower that makes it go that one step further, take the perception and then make some decisions around your actions. So that's interesting. So that gives us a little bit more clarity around AI agents and some definitions there. It was interesting to see that both you and Juan were pretty aligned there. You gave one example so far around email, which I think is sort of a broadly applicable communication technology use case that it can apply to both business and personal use. What other industries are you really looking at that are like you see, especially in 2024, you're going to see a lot of impact as AI agents and related technology are adopted?
Jeremiah Owyang [00:16:06] So at the February Llama lounge event in San Francisco that I organize... So here's eight of the 10 companies at our demo. One is called Agent Lunar, digital workers for small business. Claros is AI personal shopper agents, shopping agents. Commit, oh this one's interesting, AI talent agents. So it's for software developers, so what it does is it'll go out there and submit applications for jobs on behalf of a software dev and try to get the interview. Oh, you might like this one. Evabot, your AI research assistant in sales. So it's going to go get information about your customers and bring it back to you as chief customer officer, Floode, F- L- O-O- D- E, personalized AI EA for daily communication. So that's going to be closer to what we're talking about an email agent and for social. Then Instalily autonomous AI agents with human level productivity. Oh, that's not very descriptive. Well anyway, so that's just a sample of... There's one that's AI AgentOps and that one is run by Alex who's a leader in the space, called AgentOps, so it's agent benchmarking testing and compliance, so it's a developer tool. So there's like the ecosystem that I'm seeing that's coming around and there's like 10 others that I'm reviewing. But I think these ones are interesting for the demos.
Tim Gasper [00:17:24] That's super fascinating. One thing this triggers for me, and I'm curious about Jeremiah your thoughts on this, is I've noticed that even within data. world we've started to leverage ChatGPT apps and things like that to create tailored prompts and sort of mini applications around different use cases. So for one example is one of the things that you mentioned was around a research assistant for sales. So actually one of them that we've created internally is a research assistant for our BDR team. I'm a little curious from your perspective, there's one direction which is more like startups and companies creating services that use AI for specific use cases. Then there's more like general purpose AI for trying to maybe you tailor it, do a little bit of prompt engineering to get certain use cases. Do you see that specializations going to take off, generalizations going to take off? Or is it both for different purposes?
Jeremiah Owyang [00:18:21] Yeah, different purposes. So a diagram that I made, and anybody can go out and search there, we can put it in the show notes, is called the AI tech stack. I built it with my colleagues at Blitzscaling Ventures, Chris Yeh, former Stanford instructor who did the class with Reid Hoffman, Blitzscaling. I worked at Blitzscaling Ventures. I got feedback from Stanford, Google and AWS on it. I'll cut to the short of it. There's on one line of the stack... By the way, my background was a Forrester industry analyst and then Altimeter as co- founder as an industry analyst. So mapping markets is what I do, it's just naturally in my DNA. So foundational models like Anthropic, OpenAI, Inflection, Claude and Llama, those are horizontal. So that's wide, broad AI. Then on top of that, there's two categories of software that I see at the app layers, which is AI apps and most of there's are narrow AI. I'll get back to that in a second. Then the emerging category is AI agents, which we've already discussed. So to answer your question Tim, in that category, AI apps, they will be narrow for the time being. I do speak to the foundational model product teams and they're not going to be doing narrow. In fact, they said they're not going to be detail connecting into CRM systems and interconnecting ERP. That's not their job. That's not what they want do. So that's an opportunity for other players to do that. Another reason why we will see a majority of these AI apps be narrow with specific tasks is because they need to have enterprise contracts legally to get access to proprietary data. Then they've got to train off that data and it's going to be fine- tuned and it's got to be in a co- location center or it's got to be secure cloud or private cloud. So that's why we'll see more of those examples of narrow in that case. Last thing, this market is growing so fast that this could change. Today, January 2024, there's 11,000 AI projects. When I was tracking this market, that website called theresanaiforthat. com, in March there was 3, 000. So that's 1, 000 new AI projects per month birthing. I got to figure out who's who because I want to fund them. So I need an AI agent to help me figure out who's going to make it.
Tim Gasper [00:20:37] This is an explosion.
Jeremiah Owyang [00:20:40] Yeah, so I gave you a lot of information at a market highlight. I wanted to give you the market overview there, Tim, was that helpful?
Tim Gasper [00:20:47] Yeah, no, that's very helpful. I think you have a pretty good article that you've posted around the AI technology stack, which has a nice diagram and it shows off this layering, right?
Jeremiah Owyang [00:20:58] Thank you.
Juan Sequeda [00:20:59] Yeah, I really like this. If we go over it, so it's the data layer you have at the bottom.
Jeremiah Owyang [00:21:04] Yeah, want to talk about it?
Juan Sequeda [00:21:05] Yeah, let's go. I mean, let me describe it here. So you have the data layer, then you have the AI infrastructure, the AI models, the foundational models, and then you have both on top of the AI apps and the autonomous agents. I would love if you can like, let's walk through this because this was one of the things I saw. I mean, this is almost a year ago and I really, really love this when I saw it the first time.
Jeremiah Owyang [00:21:21] Thank you, Juan, that's kind. So yes, I love to chart markets. I did this for the sharing economy, Web 2, and so AI it's just natural. At the very bottom, and I know you'll appreciate this, the most important layer is the data layer. That's critical. I mean, without that you have nothing, right? It's like the oil. It's like the gold. It's everything. You need the data, otherwise you have nothing to train on. Then if you don't have the data or access to proprietary or exclusive data, then you have a GPT wrapper. I'm not talking about hip hop. I'm talking about an AI startup that has no defensibility because the next iteration of a foundational model will run you over. So getting access to exclusive data or creating synthetic data that becomes a product. Synthetic data is any image that's created by AI. Any output from AI that's called synthetic data. That in its own right is turning into a product. For example, Getty Images last week said they're going to allow people to purchase generative AI art. So they develop a new business model for synthetic data. All right, so that's layer one, data layer. Layer two on top of that is AI infrastructure. So this could be machine learning ops, this could be security, it could be web hosting, could be... The big one is chips. Right now there's a seven- month wait on NVIDIA chips and the price could be$ 200,000 for some chips. It's just the extreme bottleneck right now. The next level on top of that is the AI models, which we just discussed. Think of them as operating systems. So like OpenAI, Anthropic, Inflection and Mistral out of France that just got funded last month. Then on top of that is AI apps and then AI agents as we discussed, so that's the AI tech stack.
Juan Sequeda [00:23:10] If I get this correctly, is the argument that the AI apps, it's very narrow and this is something that you're... It's basically going to be a competition between AI apps and AI agents and the AI agents are going to win here.
Jeremiah Owyang [00:23:24] Cooperative, definitely not competition because the narrow AI apps will have, in many cases, access to exclusive data. But they use the OpenAI, ChatGPT APIs in order to create that friendly chat interface. So it's combining all those things together. So it's not in competition, it's mostly cooperative.
Juan Sequeda [00:23:45] All right, so I wanted to take... Talking about this whole conversation about agents and stuff. I am going to read something here. I just want to get your perception, then I'll tell you where it came from. It's from an article and says, " The entertainment system was belting out the Beatles. We can work it out. When the phone rang, when Pete answered, his phone turned the sound down by sending a message to all the other local devices that have the volume control. His sister Lucy was on the line from the doctor's office. She says, " Mom needs to see a specialist and then has to have a series of physical therapy sessions biweekly or something. I'm going to have my agent set up an appointment." Pete immediately agreed to share the chauffeuring. At the doctor's office, Lucy instructed her agent through her app or browser. The agent promptly retrieved information from the prescribed treatment from the doctor's agent, looked up several providers and checked for one of the in plans for mom's insurance within 20 mile radius of her home with the rating of excellent or very good." It'll continue. This is that agent vision, right, you have all these little agents going around understanding the perception around things. What I'm reading is actually the start of the semantic web article from the Scientific American article from 2001. What I find fascinating is that this is my academic background. I come from the semantic web space where we've been thinking about this stuff for 25 years or more. This is why I'm super excited about how the conversation of agents coming in because this is what we're heading towards right now. I think all these foundational models, these large language models, are helping to make this possible. But again, that's why my emphasis of we've been working at this for decades and decades. Anyways, I just wanted to throw that out there.
Jeremiah Owyang [00:25:30] I love that. I would advance that scenario, that use case, if a mature AI agent was there then it would've proactively booked those medical appointments and calls and already had that information surface. So that level of proactive- ness is what... I mean, by the way, this is key and I think you'll understand, grip this, AI agents are like the in- between step towards AGI. So it's like we're training all these narrow AI agents to do specific tasks which will then train an AGI to replicate all human capabilities. So you can see where this lineage is going.
Juan Sequeda [00:26:08] Yeah. So we've talked a lot about agents and you've already mentioned a lot of different tools and things, companies that you're tracking. I'd love for you to... You see so much stuff, I mean, 11, 000 products. What are the types of companies, startups or things that they're doing that are really going to change, really disrupt things? The one I want to bring up here, and I've seen you talk about it, is things like Perplexity and things like search is going to really change. So please tell us more of what you're seeing and how do you think it's going to change this year.
Jeremiah Owyang [00:26:43] So I do not have a financial relationship with Perplexity. I'll disclose when I do have one with a company. So Perplexity. ai, everybody go open a browser tab right now, everybody, if you're listening. Obviously pullover like if you're riding a bike or skateboarding or skiing. But when you get a chance, go to Perplexity. ai and do a search query that you would normally do in Google search. Try it out guys, Juan, Tim, do it. Okay, so I'm going to do one, what are the best SUVs for three kids? Okay, let's see what it says for me and I'll verbally articulate what I'm seeing. So at the top it shows the query, what are the best SUVs for three kids? Then at the top it shows four boxes with the sources, US News, Safe in the Seat, US News and Reddit and MotorTrends. On the right there's a couple images and then it gives me the answer. It lists the companies, I'm sorry, the cars, the vehicles, Honda Pilot, Chevy Suburban, Ford Expedition, and nine more. Then it gives other props at the bottom suggesting pros and cons of these SUVs. Now guys, this is scaring Google to bits. There's no ads on this. All the answers for all these sites are not on some... I don't have to go to a search engine results page, try to figure out which website to go to, click it, get stuck in more ads, and then get re- targeted by Chevy and Honda. All the answers are right there on the page and summarized using generative AI. This is extremely disruptive. Okay, what did you see? Tell us.
Juan Sequeda [00:28:23] So I did something similar. I was thinking about this. My question was I have two babies, should I get a minivan or an SUV. It gives me the whole minivan, SUVs and it gave me the same prompt. But what's interesting is that I put that exact same in Google. What Google gives me back is actually the first link is a Reddit link, linked to Reddit, which is the same thing that Perplexity showed in there. But if I go through the results of Google, it is the, what I call the old school, 10 blue links. It doesn't tell me anything more. I don't even get that knowledge panel. I lack that contextual information. So it just brought me back to how I was seeing search. So this is something I'm really, really interested about how by the end of the year, this is going to be really scary for Google.
Jeremiah Owyang [00:29:13] It is. Tim, what did you see? Then I'll make some comments.
Tim Gasper [00:29:15] Yeah, sure. So I also just riffed on what you said. So I asked about what's the best car to buy if you have four kids and it's luxury, high luxury. So it gave me a bunch of recommendations on different things.
Jeremiah Owyang [00:29:29] Any ads show up?
Tim Gasper [00:29:30] No, no ads for me, just pictures, some related links and things like that. So no, it's pretty much pure sort of answer and context.
Jeremiah Owyang [00:29:39] So this is really interesting to see. So the relationship between the user and the website Perplexity is a positive one where it's serving the user. When you use Google search, it's actually serving advertisers because it puts advertisers in between you and the answers because that's how Google makes the majority of their money. Listen, I'm not going to poo on anybody's pinata here, but I get that's how they make money, but the old ways are going to get disrupted by AI. So that's one exciting example. Last week Perplexity raised, I think, what, 50 million at a valuation of 500 M, including their lead investor is Jeff Bezos. Yeah, so if you were Sergi at Google, and he went back to work because AI is disrupting Google, you got to be really worried.
Tim Gasper [00:30:28] Yeah, big companies are having to make big changes. I know that when I'm using Google now I get my generative AI beta clip in there, but obviously it's side by side with all my 10 usual blue links.
Jeremiah Owyang [00:30:41] So most of the AI leaders that I know no longer use Google search, they use Perplexity. So if that's an indicator of early adopters and how Perplexity will be integrated into other AI apps instead of Google, this is pretty serious. This is going to be a pretty big sea change. The only downside for me is I live near Google, so it may cause my real estate property value to go down.
Tim Gasper [00:31:06] Less property tax maybe, I don't know.
Jeremiah Owyang [00:31:07] Yeah, that's the benefit less tax, which I was bitching about earlier.
Tim Gasper [00:31:12] One thing I wonder, as we make these kinds of major shifts, which are very certain to happen over that, we don't know exactly how they're going to happen, but it's certain that big changes, shifts in power, are going to happen in companies and in services over the next few years here. One of the things that I find interesting with something like Perplexity is it's getting you very fast to the answer. There's a lot of the, I'll call it, the middleman that the internet is today, right? There's a lot of content farms and creating news sites and blogs and things like that. You're trying to play the SEO and the SEM game and things like that. There's a lot there. In a world where you're just going straight to the answer, the dynamics of that changes a lot, right? If I'm a company who wants to advertise, am I paying money to try to get my facts elevated in the chatbot and things like that? Do you have any thoughts on how the dynamics of the internet marketplace evolve?
Jeremiah Owyang [00:32:15] That will happen at some point. We will see sponsored sentences. Hey Juan, here's your search result on ChatGPT or Perplexity or Claude or whatever, the best SUVs are dah, dah, dah, by the way, Chevy has a recommendation for the Tahoe and you can get free X for this. I mean we can't imagine that the trillion dollar Google revenue line, which is advertisers, the advertisers are going to give money. They will pay to get their ads in front of consumers wherever it is. They don't care if it's Google or Facebook. The CMO doesn't care. Yes, search engine marketing professionals care because they're about to get disrupted, but CMO doesn't care. I mean, that's one example. Have you guys tried Pi?
Tim Gasper [00:33:10] No.
Juan Sequeda [00:33:10] No, that one I haven't. I've been using Perplexity. Perplexity I've been starting to use a lot.
Jeremiah Owyang [00:33:14] Can we do a live demo? We can have a conversation.
Juan Sequeda [00:33:17] Yes.
Jeremiah Owyang [00:33:17] All right, so Pi stands for personal intelligence and it's created by the foundational model, the company called Inflection AI. Now I have a relationship with Inflection AI because their co- founder is Reid Hoffman, who is an advisor to our firm, Blitzscaling Ventures, the book Blitzscaling is written by Reid Hoffman and Reid is an LP in our firm. So I have ties to Reid Hoffman and this team, so I'm favorable towards this. Now PI is a free app. You can download it. Here it is. You can see it if it's on the live stream. I'm going to turn it on. I'm going to have a conversation. It's going to be a voice conversation. You guys ready? Okay, so the fourth person's entering the chat, here we go. Hey Pi, I'm with my friend Juan and Tim, can you explain in one sentence what you are?
Pi [00:34:12] Sure. I'm an AI powered language model that provides personalized assistance, information and conversation through natural language interactions.
Jeremiah Owyang [00:34:23] Hey Pi, which NFL team is doing the best right now?
Pi [00:34:38] Right now the Buffalo Bills are doing exceptionally well with 11 wins and six losses this season. They've won their last five games and are sitting at the top of the standings in their division. It's looking like they could be a force to be reckoned with in the playoffs.
Jeremiah Owyang [00:34:53] Hey, I want to switch gears. I'm feeling sad. I just found out that my dog was hit by a car and he may die.
Pi [00:35:07] Oh no, Jeremiah, I'm so sorry to hear that your dog was hit by a car. That's incredibly heartbreaking news. I can understand why you'd be feeling sad and worried. It's a difficult situation to be in. Do you want to talk a little bit more about what happened?
Jeremiah Owyang [00:35:20] Yeah, maybe later. My last question is, if I have three kids, what's the best SUV I should buy?
Pi [00:35:35] If you have three kids, you'll want a vehicle that's both spacious and reliable. Here are a few options to consider. Kia Telluride, this mid- size SUV has a roomy interior, lots of standard features and a smooth ride. Volkswagen Atlas-
Jeremiah Owyang [00:35:51] All right, you get a good sense of it. All right, what do you guys think? What'd you observe?
Juan Sequeda [00:35:58] My immediate reaction... We started this conversation saying the point of AI is to be humans again. I didn't have that impression of... I'm not being that much of a human because I'm talking to a machine. That was my immediate reaction is like, wow, we want to be humans again, but then we're going to end up talking to the computer and we're not going to be that human. That was my immediate reaction. Now that aside, this was fricking cool, awesome to go see this. Especially because it gives you a little bit of that human nature too, that seemed like if you were talking to a human on the other side, but then in reality you aren't. So I have a little bit of an interesting conflict there, but that's me. I wonder what the next generation is feeling about this. So I think it's also a generational issue too.
Tim Gasper [00:36:47] Right, yeah, it's reminds me of the older generation sees the Instagram generation and is like, I don't really understand. The Instagram generation's looking at the Tik Tok- ers and they're like, I don't really understand. So there's a degree of that that I think might be happening. But I have just a lot of curiosity about it. For example, when it told you Buffalo Bills, so the Buffalo Bills don't have the best record. So I'm curious now, well why did it choose the Bills? Is there something about you or the context of the situation that led it to say that? So I don't know.
Jeremiah Owyang [00:37:18] I have no idea.
Tim Gasper [00:37:19] Curiosity and excitement
Jeremiah Owyang [00:37:20] Idea. I don't really talk about football to it, so I don't think it had that context. So is that old data? That's wrong then, huh?
Tim Gasper [00:37:26] No, they have a lot of... What's interesting is they have a lot of momentum, even though the Baltimore Ravens have the best record. So I don't know, it's just interesting. It's curious.
Jeremiah Owyang [00:37:35] What did you notice about the interaction?
Tim Gasper [00:37:40] The fact that it seemed like it was persistent, right? So it's like there's a person in the room with you, listening to you, having a conversation with you versus something that I would compare this to... Juan, I don't know if this was similar for you. I have Alexa all over my house and I try to have conversations with Alexa, but it's not always the most seamless thing. It seemed like you were having a much more seamless conversation.
Juan Sequeda [00:38:07] Yeah, I think that's it. Today the existing voice agent type of stuff is a one way, here's a command, do this thing for me, that's it. I have a question. I get an answer. I'm done. But I think what the evolution of where this is going is that it's going to be that more fluid and more a seamless... It's going to be more of that human conversation. I think that's also something, thinking about all these chatbots and stuff that I'm working on at work, stuff that I'm working in our lab, is that we want to have these chat with the data. People want to go interact and talk with all their data. I think the way we are setting things up because of how we've just done it in the past is here's a question and I want the answer. The question is very specific. Now you're assuming that humans will come in with a very specific question, which sometimes they do have that, but a lot of the times they're like, well, I'm thinking about this and I want to have that conversation go back and forth. In reality you shouldn't even expect that when you get a question back, you're immediately going to get an answer. Because if you're talking to somebody, they're like, " Wait, what did you mean by that? Give me more context." You have a back and forth and then you figure it out. " Rephrase that. Did I understand that correctly?" So I think the way how these agents are going to be evolving and if it really interacts, maps to how we're doing things between humans, is that it will not be just a one shot in and out. It's going to be a lot of go back and forth and have that kind of a very fluid conversation. I think that's what I saw and that's what I felt from this conversation is that it was fluid, right?
Jeremiah Owyang [00:39:46] Yeah and there's people like Adam Helsinger said, Pi has the best EQ over the other ones. Jacob Alfaro said it had empathy. So I asked four distinct questions in sequence for a reason. I wanted to demonstrate the capabilities. So first was, " What are you?" It's like self- awareness. I said your names, I was hoping it was going to repeat it back but it didn't. So sometimes it says hi to Tim and Juan in the room. It sometimes says that. The second one was testing real- time data, on sports data. That's because in our pre- chat you said you were talking about football, right, I think. By the way, the Bills are number two in AFC, but they're far from the Cowboys, 12 to five, but they are highly ranked. So it was technically factually incorrect, but that was not the leader. Then the third one, as I said, my dog got hurt. If you notice that the tone paused and it changed the tonality. I don't know of any other AI systems that can change the tone and voice. Then it said, " Do you want to talk about it? Do you want to talk about it?" That's a big... No other tools are doing that. Then lastly was product recommendations, which goes back to the prior discussion, to commerce. So between those four questions, I wanted to show the breadth and width of the latest technology. Now just so you know that inflection, they've raised over a billion dollars and the two co- founders are Mustafa from DeepMind, they created Go, and Reid Hoffman who's on the board of Microsoft, LinkedIn, one of the major investors at OpenAI. That's the team and their top investor is Bill Gates, Eric Schmidt and NVIDIA. So NVIDIA, in part of the deal, they gave him 22, 000 H100 chips. So it's one of the largest supercomputers in the world right now. What we just demoed is version two of their foundational model. So just imagine what version five or version 10 is going to look like. I don't think we're going to be able to distinguish it from a human.
Juan Sequeda [00:41:48] No, that's true. It's always the reminder. I have a iPhone right here and this is what version 15.
Jeremiah Owyang [00:41:54] Yeah, maturity felt like version 12, right?
Juan Sequeda [00:41:59] We have to realize that if people are not impressed, some people are not impressed, like, okay, don't worry about it, it's just going to get better and better and better at some point. It's like it's going to be so human- like.
Jeremiah Owyang [00:42:10] So PI stands for personal intelligence. So you guys I think are around my age. I think you're like... So you guys know Star Trek Next Generation?
Tim Gasper [00:42:20] Uh- huh.
Jeremiah Owyang [00:42:20] So Microsoft invested in OpenAI and Inflection. So the best reason why you think they do both is what I think is OpenAI is like the Enterprise ship computer and Inflection, PI, is Data, the android who eventually develops real feelings. So they've invested in both, so get ready. You want to have an emotional conversation with your data, Juan, with your customer's data, that could happen. I know you guys have clients that are B2C or healthcare where the data is an emotional story. It is a personal story. So I can imagine personal AI's interacting with your data.
Tim Gasper [00:43:08] That's a really good point. I'll say that as all of this AI advancement has been happening and the chat interfaces, the natural language conversations, I and I know many others go to that analogy of the computer talking to the computer. I think that's probably the more often used analogy. But we forget about Data. Data's a computer too. He's a computer with legs who initially-
Jeremiah Owyang [00:43:33] An android.
Tim Gasper [00:43:33] An android, yeah. Initially he's trying to mimic emotions. For those who watch Star Trek, know how that evolves over time. He's also an AI. Both of these paradigms are useful, but for different reasons.
Jeremiah Owyang [00:43:49] Yes, great. So yes, you guys know me. I talked to the young Gen Z AI founders. I'm like, " Yeah, it's like ship, computer and Data." They're like, " What are you talking about grandpa?" I'm not even 50 yet.
Tim Gasper [00:44:02] You've got some good TV you got to watch.
Jeremiah Owyang [00:44:04] Yeah.
Tim Gasper [00:44:05] That's good stuff. Before we start to hit some of our lightning round and some of our final questions here or final wrap up here, one sort of final question for you Jeremiah, is what do you think is the net impact of this on the job market over the next couple of years here? There's a lot of people that are fearful of AI taking their jobs away in one respect, but in another respect, creating new jobs, amplifying the ability of what a person can deliver and do, making a person more important than ever. Do you have a take on the impact this is going to have on the economy and the job market?
Jeremiah Owyang [00:44:43] Oh, that's a loaded question. So I met with some federal people from the government, can't say which cabinet group that is, but yeah, they're obviously watching this. There's no question. I see three different things are going to happen for sure. One, people's jobs will be lost because any task that's repetitive done digitally will be automated. So some tasks and jobs will be lost. Two, some brand new roles will be created. There's 11,000 AI startups. That's all new jobs, but here's the downside, they're not going to hire that many people. It's not going to fulfill that aspect. Then the third aspect is we will definitely see every single job, 99% of jobs will be augmented and impacted by AI. The only jobs that will not be impacted by AI is if they're not connected digitally in any way. So for example, people say plumbers won't be impacted. Oh yes, they will big time. So they'll use startups like Productive AI where a plumber could use that app. It'll listen in conversations and book and schedule on phone calls. They'll use augmented reality. They will order parts using VR. Of course every homeowner can use AR and AI to solve level one plumbing issues. I'm no plumber, but I can use YouTube already. Why can't I use AR? So even plumbers, trades will be impacted by AI. Nobody's immune unless you are a hermit. So to answer your specific question on the economy, it's pretty clear to me that we will need to have to change the economic distribution system, which is such a tough topic to discuss. You're a Texan. I'm a Californian. We can have a pretty hardy battle on this topic, but it's very clear that there could be inequities if we don't address this and not everybody's going to be able to keep up with this. AI is moving at an exponential rate. So I'm expecting that, I don't think it's this president or the next one that will address this, but a form of UBI, universal basic income or heavier taxation to big corporations that use AI. I think these things might start to happen. It's really hard to tell. I'm not sure exactly which way. On the other end, we might see that every single worker just might become super productive and everybody keeps their jobs, so there's just another scenario on that. The last thing is this, will every single role, even entry level roles be replaced by AI? I don't think so because just from a pure economic standpoint, if many people lose their jobs, we'll have more workers. So we would technically have more customer care folks, more junior sales ADRs, more waiters and waitresses available, just a surplus of workers, so therefore it's not as expensive to hire them. So there will be those workers, which will result in an interesting economic twist. Then I'll land the plane here, is that we will see human premium, premium on humanity, for service and sales and customer care will actually rise to the top because people are going to be willing to pay for that instead of dealing with an AI agent. So a long answer because I think about all these permutations and scenarios and I'm not really sure which one will land on. It's probably all.
Tim Gasper [00:48:05] Right, yeah, there's a lot of second, third, fourth order effects that are going to happen from all of this, including... I think about when the pandemic happened and everybody was hopping onto Zoom, and the immediate assumption was, well now it's all going to be remote. But then what it did is it put a premium on in- person interaction, and you have the most popular travel season ever and the most popular conferences ever. So it's really hard to predict how all these things connect to each other.
Jeremiah Owyang [00:48:29] Which has its own third order effects after that, right? Yes, exactly, all that.
Tim Gasper [00:48:35] This is awesome. What an exciting way to think about the future and all the possibilities here.
Jeremiah Owyang [00:48:40] Thanks.
Juan Sequeda [00:48:41] Yeah, no, thank you so much. I think this is a great way to start wrapping up, I think.
Jeremiah Owyang [00:48:46] Cool.
Juan Sequeda [00:48:47] These are the discussions, the hard discussions, we need to start having more. It's interesting that you say that not this president, I mean, probably eight years from now or whatever, this is where it's going to be on the top of everybody's conversation.
Jeremiah Owyang [00:49:00] I think that's right, eight years.
Juan Sequeda [00:49:02] Eight years. All right, well let's go into our lightning round questions. So I'll go with the first one here. Foundational models require so much investment, Google, Microsoft, Amazon have big incumbent advantages because of the dollar requirements. Is that a problem for AI startups?
Jeremiah Owyang [00:49:18] Yeah, it is. Also, yeah, the rumor is it's 500 million to train OpenAI. So yes, it is a problem and it means they will be the dominant winners. So we're not sure if the AI startups can and sustain over that. So that is true unless they have access to exclusive data.
Tim Gasper [00:49:38] That puts a premium on the data, interesting.
Jeremiah Owyang [00:49:39] Correct.
Tim Gasper [00:49:41] Second lightning round question for you. So in 2024, so looking at the short term here, will AI agents most heavily be adopted by companies or are you betting more on personal use in terms of the biggest wave of growth here?
Jeremiah Owyang [00:50:02] I think the first adopter, as far as I can see by how they're targeting right now, will be on solopreneurs, like a one person company, TikTok stars, plumbers, fitness coaches, therapists. They need support staff. They can't find them. They can't hire them. They don't have enough money. So they'll be the first to adopt. Then from there, small business. By the way, large companies, they can afford to hire more people, so they don't need the AI agent for this use case.
Tim Gasper [00:50:36] The incremental value is different. That's super fascinating.
Juan Sequeda [00:50:39] That's an excellent, excellent insight. I didn't think about it that way. Right, next question, if you could pick one industry to be most positively impacted in 2024 by AI, which one would you choose, healthcare or something like that?
Jeremiah Owyang [00:50:55] Yeah, that's it. It is definitely healthcare. American's health is on decline. We have the highest rate of anxiety and depression ever recorded since psychologists have been recording in our society. For the first time, yet also due to Covid, the American life expectancy has dropped. The wealthiest nation in the world, that should not be happening. So obviously there's systemic issues, institutional issues, regulatory issues, but AI can help with that to decipher our data, to understand our bodies and our minds better so we can live a better, longer, healthier lifespan and health span. That is key.
Tim Gasper [00:51:37] I'm very hopeful and optimistic about that. I think there's so much potential there. Last question for you on the lightning round is you mentioned plumbers as an example of something people think might be unaffected and then you came up with so many reasons why it's going to be affected so many different ways. Is there truly an unaffected job around AI?
Jeremiah Owyang [00:52:01] They have to be digitally non- connect, but they'd have to be digitally non- connected by two orders, right, because if they have their supply chain, they get any of their resources, food or information through digital in any form or they sell anything that eventually goes to somebody who uses the internet, then they would be affected. So who would that be? It would be a remote farmer in the woods that is only building a homestead for his or her family. Can you think-
Tim Gasper [00:52:32] Even, I mean, they could benefit. It's just maybe they don't have the technology, right?
Jeremiah Owyang [00:52:36] I mean, can you think of an example? I really can't.
Tim Gasper [00:52:39] I'm really struggling to figure it out. I was hoping maybe you had one in your hat and you could pull out, but yeah.
Jeremiah Owyang [00:52:46] I mean, whether it's direct or indirect, they will be impacted by AI because the supply and demand, and even global supply chain, will be impacted by AI. So even if you don't work with tech, but you rely on a supply chain that is using digital, which is every supply chain, unless it's a farmer's market in a village. I mean, that's really it.
Juan Sequeda [00:53:08] Yeah, you live autonomously in the middle of nowhere.
Jeremiah Owyang [00:53:12] Yeah, if you're unconnected, then you will not be impacted by AI.
Tim Gasper [00:53:16] That feels marginal enough. The answer is basically no, everything will be impacted by AI.
Jeremiah Owyang [00:53:21] Yeah, and anybody who is that remote, unless it's by design... For example, Africa, they want to be online really bad and they're using satellite- based wifi to get connected. Now they're going to have Pi, they're going to have Inflection, they're going to have GPT and their societies are going to boom because they can rise to develop world within one generation using AI. It's really going to be phenomenal what we're going to see.
Tim Gasper [00:53:50] AI native generation.
Jeremiah Owyang [00:53:51] Yeah.
Juan Sequeda [00:53:52] You're going to come in fresh enough to deal with emails and stuff like that.
Jeremiah Owyang [00:53:56] Yeah, they can design something better than email inboxes. They'll figure something better.
Juan Sequeda [00:54:01] Well, to close the loop here, Tim, take us away with takeaways. We got a bunch of stuff here we're taking notes on.
Tim Gasper [00:54:07] So we started off by asking you honest, no BS, how is AI going to impact us this year? Throughout this conversation you've come up with really incredible and impactful ways across industries, across use cases where AI is going to have a huge impact. The technology that's going to lead us there is innovating really quickly, including some live demos that you did today. The number one, first example that you gave of something that's going to be something that happens in 2024, one of your bets here, is that we're going to get to email inbox zero. It's not just going to be from us spending all those hours at 10:00 PM trying to get to inbox zero. It's going to be leveraging AI and AI agents and AI services to do things like review, summarize, reply automatically. This sort of AI agent becoming your EA, it's going to let us be humans again. I think that's a really great example and microcosm of how this is going to apply to so many different industries, services, technologies and to not just make things easier, but hopefully help us become more human. We can hopefully really capture the best part of what this is going to do for us and all new industries and things that are going to spawn out of it. You mentioned AI agents as a really important aspect here, and we talked a little bit about the definitions around that. You said AI agents operate autonomously with little human oversight. Juan, you agreed and said" Yes, autonomous, that's a key thing. But also the perception, the perception and the sensing making that that's what allows it to be intelligent." So I think that was a really great definition that we aligned on around that. Jeremiah, you also mentioned a bunch of different startups and services, ones that you interact with and are exposed to, a couple that we showed off today. Some of these companies are things like shopping agents, AI talent agents that will submit applications for jobs for you, helping you get a job, research assistants, email assistants, AI ops, agent ops. I mean, there's so many different types of services that are being spawned here. You mentioned over 11,000 AI projects where in March there were just 3, 000. It's a thousand new ones every month. This is really a Cambrian explosion of innovation right now, and it's super exciting. Obviously not all of these services are going to bloom and blossom, but this is the point where we're trying things, experimentation is happening and so much is going to evolve over the next couple of years here. It's truly exciting. Before I pass it to Juan for his takeaways, I asked a little bit about sort of broad AI versus narrow AI, which of those is going to win out. I think you very astutely mentioned that they're both valuable and they're both a little different. You talked about your AI technology stack and how they both fit into that. So I'll use that as a transition to pass it over to you, Juan, for your takeaways.
Juan Sequeda [00:57:01] Yeah, I think that AI technology stack, the article you have, is a very critical one. I hope everybody who's listening can go find that right now. The most important layer we've talked about is the data layer, right? It's a gold critical. Without the data, you have nothing but just a GPT wrapper, which in the next iteration, the foundation model will just take it over. So what's interesting also about the data, you have synthetic data and that's actually going to get bigger because right now that synthetic data being the output of AI, Getty Images, you can now buy, they're selling out synthetic data, so that's going to be the next thing there. There's a whole AI infrastructure, like AI ops, security hosting, the chips, and then you're going to have, on top of that, you have the AI models like Anthropic, OpenAI, and then we have the apps and the agents which are going to be working together around that. I loved how we went and talked about these demos. So we did look through Perplexity. I think that's a very, very popular one. It's interesting. I like how you say it's serving the user while Google is really serving the advertiser. Hey, the AI leaders right now, the early adopters, they're using Perplexity and not using Google anymore. I liked how you said but we're still going to probably see sponsored sentences. We went through this demo of Pi. ai. I think it's an emotional conversation. You went through four things, it's self- aware, you can talk about real- time facts, it's empathetic and you can talk about product recommendations. I think this whole notion of having a personal AI interacting with your data. We wrapped up the conversation today with jobs, how will they be lost, and effectively any digital repetitive task will be replaced, all jobs will be augmented, even plumbers. We've talked about this already. There is going to be a change in economic distribution, so we need to have these conversations so there'll be some form of UBI and that's a conversation that probably it's going to be eight years from now. It's going to be in the top big topic at the presidential level. People will pay a premium on humanity, I like that. Pay more to deal with humans instead of AI. How did we do? What did we miss?
Jeremiah Owyang [00:59:04] Well, I'm canceling my GPT subscription because you guys are definitely pros at summarizing unstructured data. Thank you.
Juan Sequeda [00:59:11] Well, as always, we say this is you. This was all you. So thank you so much because these takeaways is all from you. Thank you so much.
Jeremiah Owyang [00:59:18] Thanks for having me. Guys.
Juan Sequeda [00:59:20] Just to wrap up quickly, we always ask people, what's your advice? Who should we invite next? And what resources do you follow, people, conferences, blogs or podcasts, whatever?
Jeremiah Owyang [00:59:31] Oh my God, what advice? Do not fall behind on AI. I have young kids and with my supervision, we're using these tools because I want them to be AI natives. I need you all who are listening, I know you are if you're listening to this advanced podcast, you are leading into it for personal, family, career health, don't fall behind on this one. This one is a big one. Who else should you have on this podcast? I would suggest two of my business partners. One is my current one, Chris Yeh, who wrote Blitzscaling with Reid Hoffman on how do companies grow using data. Then Charlene Lee, who focuses on AI and leadership and who was the founder at Altimeter Group, which I was co- founder of, also Forrester. I would recommend them. What was your third one?
Juan Sequeda [01:00:15] What resources do you follow?
Jeremiah Owyang [01:00:17] Yeah, one of the best newsletters... There's two out there in the AI space. One is called The Neuron, neuron like brain neuron. Another one is called Synthedia by Bret Kinsella. If you subscribe just to those two newsletters daily you'll be up to speed on AI. I really mean that, those two, you should have enough info on real time.
Juan Sequeda [01:00:40] That's excellent. Jeremiah, thank you. Thank you. Thank you so much. This has been a fantastic conversation. Just quick reminder, next week we have Samia Raman, who's a director of enterprise data strategy at Seagen. We're going to be discussing governance, meshing AI knowledge graphs and everything. With that, thank you so much. This has been a phenomenal conversation. I'm looking forward to-
Tim Gasper [01:01:00] Cheers, Jeremiah.
Juan Sequeda [01:01:01] ...how much we're advancing this year and keep conversation going and see what's going to happen next year when we chat again.
Jeremiah Owyang [01:01:07] Thanks guys. Say hi to Bret for me.
Juan Sequeda [01:01:10] Will do. Bye.
Special guests
