Tim Tully of Menlo Ventures joins Nate to discuss Navigating the AI Hype: Where Value Will Accrue for Investors, the Future of AI Tech Stack, and the Role of Open Source, Vector Databases, and Recommendations Engines. In this episode we cover:
- AI Hype, Investments, and Company Growth
- Using Vector Databases to Improve AI Accuracy
- Tech Investments and Trends in Developer Tools and Microservices
- Enterprise Customer Demands, Engineering Teams, and Data Privacy
Guest Links:
The hosts of The Full Ratchet are Nick Moran and Nate Pierotti of New Stack Ventures, a venture capital firm committed to investing in founders outside of the Bay Area.
Want to keep up to date with The Full Ratchet? Follow us on social.
You can learn more about New Stack Ventures by visiting our LinkedIn and Twitter.
Are you a founder looking for your next investor? Visit our free tool VC-Rank and we’ll send a list of potential investors right to your inbox!
Tim Tully joins us today from Menlo Park. He’s a partner at Menlo Ventures, a multi-stage venture fund investing in cybersecurity, cloud infrastructure, and artificial intelligence. Tim has invested in companies including Pinecone, Neon, TruEra, Edge Delta, Julia Hub, and Squint. Prior to Menlo Ventures, he was the Chief Technology Officer at Splunk, a software company that specializes in data analytics and observability. Tim, welcome to the show!
0:47
pleasure to have you can you take us your backstory in your path to becoming an investor? Yeah,
0:51
so I’ve been here two and a half years before that, I was CTO at Splunk. As you said, I ran the product and engineering team, and IT and security, which is a lot of fun. It’s kind of like the fox in the henhouse. If you let an engineer into the IT organization, so I had a lot of fun with that. Did you know close to four years there helped put into the cloud, we grew revenue pretty significantly while I was there, so I’m pretty proud of it. And before that, I was at Yahoo, I ran all of the media engineering, so things of you would think of as being Yahoo, like yahoo.com, Yahoo Finance, you know, big webpages that get, you know, millions and millions of views a day. I mean, it’s huge scale, Yahoo is a huge site, and people probably don’t realize it still. But you know, Yahoo Finance is one of the most trafficked pages in the world. So is yahoo.com. And that was a lot of fun. That was 14 years. So before that, you know, a couple startups and then even before that, worked on the Java language at Sun Microsystems in the in the mid to late 90s is a company a lot of people sort of hopefully did not forget about, but I did do a solid four years there. And it was it was amazing. Awesome. And
1:51
talk to us a bit about your thesis in your investment approach at Menlo.
1:55
Yeah. So like you said, I mostly invest in AI, ml infrastructure plays, things that you would think of recovering CTO would want to invest in, you know, things you’d want to buy, if you’re a CTO for your engineering teams, developer tools, I do dabble in cybersecurity a bit. So yeah, that’s basically the summary of it.
2:13
Awesome. Well, I’d love to get your thoughts on AI, maybe a good place to start is a gauge of the hype cycle, you know, do you think AI is under or overhyped?
2:22
You know, it’s a dangerous question, man. Like the question, I was worried you’re gonna ask the most, because you can be wrong both ways. But you know, I guess you’re looking for hot ticks here, I tend to worry that there’s a little bit of over hype on it, just because of, you know, from where I sit, people are treating it, like it’s this panacea that’s gonna, you know, solve all problems of humanity. And I don’t think that’s that’s the case. I think the best AI is Augmentative and makes workflows better, and enables humans to do better, better work and become more productive. And the way it’s being treated as as this like magical elixir, that’s going to just change everything. And I don’t think we’re quite there yet. I’m excited. I continue to one invest in a lot of generative AI companies. I’m just a little bit worried that we’re getting over our skis just a little bit like I’m excited about what’s to come over the next three to five years. But you know, we’re, we’re still early, right in some people will say we’re in the seventh or eighth inning, I say we’re in the second inning. So we’ll see. But I’m excited about my portfolio. I’m excited about the some of the companies out there and I’m excited about generative AI. Is there
3:28
an analogue like as you think about technology shifts of the past or paradigm shifts? You know, which one does this feel like it most closely resembles? Probably
3:36
mobile, mobile, like you knew something was happening? Yeah, like, early on, things were starting to happen. It wasn’t super clear what was going to turn out like I, you know, I couldn’t say in 2007, that I knew I’d be ordering my dinner every night on my mobile phone, which I do do now, frankly, that’s like pretty much how you can get all my groceries, services. But I couldn’t predict that in 2007, like you knew the iPhone or 2008, you knew that iPhone was exciting. You know, Android was exciting, but you didn’t you didn’t know what’s gonna change your life that much. And I think that’s happening here. Like, you know, something’s happening. It’s not super clear what it’s going to be you want to invest, you want to be a part of it. But nobody can fully predict exactly what’s going to happen in five years. I mean, people who claim to be able to do that are silly. Well,
4:20
I think they should be probably options trading instead of investing in venture if they can.
4:24
I’m happy to talk to you about that. Two. Big options traders. Yeah, seriously. I love options trading man. Yeah, all day long. I’ve recovered calls I like love to sell puts. Yeah, we can do another podcast
4:35
on this. Yeah. Give us give us your best tip at the moment. Well, I
4:40
mean, this all came from there’s a company we acquired when I was at Yahoo. It was a Hungarian analytics company. And so I spent a lot of time in Budapest. And these guys are, you know, Eastern European Olympic math, like wizards and they developed this like philosophy and approach on options trading, which I had picked up slowly over two years. And yeah, I absorbed it and you know, did well especially when the Uh, you know, credit crunch, hit in 2007 2008. And some of the banks were doing pretty poorly. There was a lot of volatility. And so you could trade that volatility on options and make, you know, some solid money. And then I got into things like trading options on the volatility index and like, yeah,
5:16
I love it. Well, yeah, our anchor investor is a market maker that represents over 10% of all options traded. So maybe another time we’ll have another Oh, wow. Yeah. All right. Back to Let’s Talk generative AI. Are you investing in companies that are building Gen AI from scratch?
5:34
Yeah, so I’ve, you know, one company that a lot of folks know called pine cone, it’s a vector database company, we did the series A in that company back in 2021. This is before you know, the generative AI boom, hit, but pine cone is is powering what’s known as Richie Wagner generation or the bag pattern that’s, you know, pretty much read, every generative AI company is doing to implement their applications, whether it’s, you know, Harvey or, you know, the Jasper’s of the role, though these guys are doing augmented generation, it’s through some vector database of some kind. And it happens to be at Pine cones when big ones in our portfolio, we’re in anthropic, we’re pretty proud of that, you know, anthropic is very similar to open AI in many regards to have a foundation of all that, you know, millions of developers rely upon when another company called typeface, which is a digital marketing company that does, you know, digital marketing automation, which is absolutely fantastic, you know, completely augments I say, augments my favorite word these days, augment the work to be done and digital marketing teams and sort of do everything from content creation all the way through campaign management. So we’re having investors in that space. Okay,
6:36
so you will do companies that are building from scratch, Jen, and Oh, for sure.
6:40
Oh, for sure. I mean, we have that’s anthropic, and we’ll continue to do that.
6:45
Perfect. Yeah, you mentioned the pine cone, you know, vector databases are becoming increasingly important in the AI and machine learning space. Can you explain why these databases are crucial for application? Like applications, like recommendation systems and natural language processing? Yeah.
7:01
So you know, the big use case is bag, right? Where you want to ask a question through a chatbot, or some kind of knowledge base that’s being exposed in the UI. So if it’s a legal tech company, or, again, a chat bots, like one of the more canonical use cases, you’re going to use GPT 3.5, Turbo GPG, for from open AI to get your generative AI. That’s the big model that are cloud from anthropic. And so these models do what’s you know, everyone knows by now called hallucinations, right, which is it’ll give you an incorrect answer to a question you have. My favorite example of this is if you ask it about Lincoln’s assassination and John Wilkes Booth, you know, you can ask various ways, but it’ll basically tell you that John Wilkes Booth and Lincoln run different continents when Lincoln was assassinated, just insane. And then like the next gen, so it’ll say like, oh, you know, Lincoln was assassinated by John Wilkes Booth that, you know, 3pm for theater? Well, you know, and once you sit there on different continents on the next one, you said, you know, precise time and location. So you can overcome these suits, or this nations. And the way that people do that is by using a vector database. And so what they’re doing is treating the vector database as a knowledge repository. And so instead of just blindly asking, you know, one of the generative AI models to answer the question, you first go to one of these vector databases and ask for relevant content related to that question that you have. And so I would have pre populated pine cone with content about Lincoln and John Wilkes Booth, pull that out before I went to say, opening Iowa anthropic, and then made that part of the prompt and fed and said, like, Hey, answer my question about Lincoln’s assassination, using the context below that I just pulled out a vector database. And so to finally answer your question, what’s happening is you pair up one of the vector databases with, you know, one of the foundational models and put them together to overwrite the hallucinations. And what that does is effectively reduce some of these foundational models to being great writers, which they are, they’re oftentimes better writers than most college graduates.
8:57
How do you dress latency? And you know, the time it takes to reference the vector database as well?
9:03
Well, the vector database is an index, right? So you know, depending on who’s listening in your audience, they’ll appreciate the nuance difference between an index and a database. And, you know, oftentimes, indexes are within databases, and we get into the semantics of that that’s neither here nor there. But the response time from the vector database, or index, if you will, is is built to be fast. So you know, there is some overhead, there’s some latency, but it’s, you know, on the order of hundreds of milliseconds, not seconds. Right. So it’s like de minimis and negligible, I would say, Tim,
9:33
how do you see the role of open source AI projects in the future of AI development?
9:38
I love it. Right. Personally, a little bit reminds me of the browser wars, if people in your audience are old enough to remember, you know, ie versus Netscape or Mozilla, you know, they push the innovation curve, right? They drive the innovation, right? And if you don’t have this kind of competition, who’s going to push the innovation, right and so, you know, I have a lot of experience playing with llama and all the You know, various parameter size models of llama llama to and the quantize versions and, you know, they get a lot of recognition, I will personally say, you know, my experience there, okay, I don’t think they’re nearly as great as some of the larger foundational models, even after doing some fine tuning on them, you know, they’re better, but I wouldn’t say they’re as great. And so, you know, they’re really great, they’re great for educational purposes, I think some companies will find good use for them. But personally, I prefer the larger foundation of models from the likes of open AI and anthropic and, you know, there’s probably a reason for that these are large companies, they’re built to build the best models on the planet, right? Yeah, whereas with Facebook, Mama, it’s like, well, you have awesome people there, they have you on Macoun, they have amazing people there, and they’re gonna have great models, it’s just, that’s not they’re raised on death row on the on the planet, I wouldn’t say as a company. And so unless, unless it’s a company, steering the ship in that direction, it’s hard to have the best model in the world,
10:56
you know, as you consider kind of the three layer model to AI, which layer are you most interested in? So
11:02
I think of it mentally, as you know, picks and shovels at the very, very bottom right, which is like foundational models. And at the other end of it, right, it’s like a dumbbell in my mind, right at one end of the dumbbell, you have the picks and shovels. On the other hand, you have like applications that touch the user. In the middle, you have middleware, which is sort of something like laying chain, things that sort of connect and run the execution flow, and then in the DAG, to do the work. I like the hands of the dumbbell, like I love the picks and shovels, the foundational stuff, I love the stuff that touches the user. And the reason why is because they’re harder to change, right? They tend to have more lock in, and they tend to produce more value, especially for investors, they’re harder to remove, right? Like, it’s gonna be hard to take your vector database out and swap it for something else in the same way that it’s hard to take out, like Oracle from an architecture, right, that’s how you should think of a vector database. And then on the same side, on the application side, right, if if, if people fall in love with see an analytics tool or a marketing tool, right, it becomes their their tool to do their job. And hopefully, it’s something that’s it’s hard to rip out of their hands. And so I love tools like that, where the user loves it so much, that’s hard to rip out of their hands. Those are ones that tended to have a lot of value, and become more creative. So that’s, that’s where I spend most of my time in the middle. I don’t I tend to kind of ignore a lot of it. To be honest, I
12:15
could one argue that developer tools sit in the middle. You could
12:18
I mean, definitely, that’s the main chain. There’s no doubt about that. I, you know, I don’t know how familiar you are with something like that, or llama chain. But you could argue that certainly, for sure. And you know, I mentioned at the beginning of the show, I love developer tools, but maybe not so much in the middle.
12:33
How do you recommend companies stay up to date with the latest advancements and best practices and fine tuning and cross engineering AI capabilities? Well,
12:42
the best ones are reading Hacker News every morning. Like you better be reading Hacker News or you know, other publications, because things are just moving too quickly. You have to be on Twitter, just just following it. That’s where a lot of the crowd is posting new models. And they’re sharing research, and they’re sharing things that they uncover. You got to be networking and talking to other companies and just talking to your network and making friends with other companies and learning how they’re approaching problems. And then hopefully, you have a great VC like myself who’s imparting knowledge because I do, you know, certainly spend a lot of time doing that. But, you know, the world is moving quickly, man, it’s really hard to stay up with what’s happening. And there’s new models coming out all the time, and they’re getting better. And you know, you just, you have to love it. Right? You have to love it. That’s what I would
13:25
say, do you do still tinker and build yourself to stay relevant? Or is it are you observing, trying to observe as much as you can?
13:33
Oh, dude, I was. I was coding before we got on this podcast, options trading, voting, investing, and I might get up as like green for activity every day. Yeah, I mean, I’m I build stuff for for Menlo Ventures. You know, I do things for my portfolio companies. Like I’ve written command line interfaces for to my companies. I’ve written API’s for one of them. There’s one that I’m writing another API for right now for a company I’m looking to invest in. And I do a lot of recreational stuff at home. Like there’s a video game I’m building in the background and unity. So it’s like, you have to keep the knife sharp. And otherwise, it’s hard to be like a technology focused investor, right? A lot of folks are not, you know, recovering CTOs. A lot of them are, you know, finance professionals. And you know, I’ve gotten one trick man, it’s like, I know tech. And so I better I better stay on top of it.
14:20
What type of game are you building? Is it an FPS or? So?
14:24
You know, cyberpunk 2077. Yeah, you must have that game. So it was super, super delayed, right? And I got tired of it being delayed and I had purchased it and I got tired and tired and I really wanted to play it. So I just wrote my own. So I wrote a cyberpunk 2077 Like game which is set in like post apocalyptic Tokyo, okay, where you know, everything’s dark and everything is neon lights all over the place. And it’s just a go into Unity store. I buy a bunch of objects, just lay him down in the Unity Editor, lay some C sharp over it. And you have a really fun sort of immersive first person shooter? Yeah, it’s just it’s like a big open world thing, cyberpunk style with fewer bugs.
15:05
So are you in gaming at all? Or is this more just me?
15:09
I have a partner, Amy Wu, who does, but not me personally, I don’t have expertise in it. I can code it. But you know, I’m not an expert in in marketplaces and the economics of gaming at all. So, no.
15:21
So you mentioned that you spent a significant amount of time thinking about developer tooling and productivity, can you share some insights into the latest trends and innovations in the space that, you know, have caught your attention?
15:32
Yeah, sure. You know, obviously, you know, the obvious answer I’m gonna give you is GitHub co pilot, you know, I’d be remiss to not mention it, and I’m sure you’ve got an earful. But I’ll give you my hot take. It’s pretty good. You know, it is really interesting, if you’re learning a new programming language, like I picked up rust, I would say about six or seven months ago. And Rust has a really high learning curve. And GitHub copilot was there. And it shows you how to do you know, idiomatic aspects of a language much faster than you know, for me to pull out the book and look at, you know, how I should do a for loop properly, right? It’s great at that the auto completion stuff is really awesome. And then you start to think like, Oh, I’m gonna be out of a job in a year because this thing is getting so good. And then you try to do something else. And then you go like, like, What the eff did it. This is wrong, like it gets it like 90%. Right. And but the problem with software is, it’s not like horseshoes, and grenades, where you can be like, sort of close. It’s either it works or it doesn’t. And they’ll give you recommendations on things that are just wrong. Or not idiomatic in some cases. So no, I love it. It’s argumentative, and makes me probably 10 to 15%. Better, right? It’s awesome. It’s just going to keep getting better. Like I said earlier, these tools are going to get more powerful. So that’s interesting. I think the other thing that’s really caught my eye is the so called workflow orchestration systems. Right, the ability to sort of like, externalize workflows across microservices. Right. So every architecture in the world now is just a composition of microservices, right, hundreds of 1000s of these things. And the question is like, how do you chain them together? How do you make them depend on one another? How do you do loops inside of them. And so there’s a few companies popping up, which basically, let you externalize that the state management that comes with it retry logic, the transactions. And that’s a really powerful piece of technology that’s going to make developers lives a lot easier as they continue to make this transact transition over to microservices. So that’s definitely caught my eye. And then just maybe the last one is this idea. And I have one company called neon that’s done this, which is like a full separation of storage and compute across all database technologies. Every company that used to be a database, so it looked like a database has gone fully separation of storage compute in the cloud. And companies that doing that are going to build humongous businesses, right, snowflake was the first one that’s done it, you know, I’m an investor in neon, which did it for Postgres, and it’s continuing over into the likes of Splunk. Now, so it’s, you know, there’s a lot of companies popping up, they’re becoming, like Splunk, but have fully separated storage, including the way Splunk has, but made it a, you know, a different tool. So really exciting times, how
18:02
does your investment framework change? If it’s maybe developer tools within the open source space versus not?
18:12
I mean, open source is hard. You know, when I see something in open source, typically what I do is I run away from what’s a library, right? Like, libraries are hard to monetize, like, I can’t, developers don’t want to pay for a library, right. And at the end of the day, I have to make investments and things that are going to return the fund. And a library is not going to be it right, the real model for open source is looking for something that’s a platform, right? So the the best example of that my mind is, is confluent, right? Like, that’s Kafka under the hood. Nobody wants to sit up and run Kafka, right? You’re not going to do it’s painful, it kind of sucks. I rather just pay somebody to do it. And so I’m going to pay the compliment, guys, whatever it costs to get that thing up and running from in the cloud. And and I’ll just use it as a cloud service. So that’d be a great example of that, right? Yeah, that’s basically it run away from libraries, embrace platforms, look for things with GitHub stars, look for things that I personally enjoy using, which is something that I have a litmus test against. It’s like, I kick the tires on it, I got to use it got to see if it’s something that I’d want to use as a developer or buy. And it’s yeah, it’s really that simple.
19:17
You know, thinking or transitioning a bit to customers, as you consider new technologies that can drive enterprise value. How have enterprise customers demands changed in recent years? And what do you think the implication is for startups that are targeting the the enterprise?
19:33
Yeah, it’s it’s really heavy around being multicloud. Whether they want it or not, right before it was like, will you even be in the cloud? That’s where the big question was for me when I joined Splunk back in 2017. Right. The question was, like, what a CISO want a security system to be in the cloud where they want the logs from the IT department to be in the cloud. Like we really struggled with that personally. Now, you fast forward six years, and it’s like, yeah, you know, for sure. are right, and it’s like they’re actually going full, you know, headfirst into the cloud. That’s changed pretty dramatically this sort of embracing of the cloud and now wanting to be multi cloud. It’s not even a question now that most companies have to be multi cloud. Yep. And then customers are demanding it for you know, rightfully so. With that comes questions around data sovereignty, right? Like, where does the data reside? Which which continent is which? Which availability zone? does it reside in? And then there’s, you know, even that, it’s like, which VPC? Is it? My VPC? Is it your VPC? Is it what’s called on prem cloud, right, versus, you know, public cloud? Like all of these things are fairly novel questions, I would say, relative to say, three to four years ago. And then like, the last one is more around usability, I’m seeing a lot more demand for UI excellence, right? It’s no longer the case that like things can just work. They have to be fantastic, right? Because everything is SAS now. And so people are demanding this remarkable, sort of immersive experience on the UI UX side. So that’s, that’s very different. I felt like two to three years ago, engineers just wanted like a headless tool, something with API’s. Now they demand like beauty. And I think I think data dog data dog do that. Yeah. All right. And I say that all the time. Just copy with data dog did because they have a great, have a great back end. It’s great. But the UI and UX and usability is just like it’s world class. Yeah, they set the bar. You
21:26
know, how do you think about this is something I’ve dealt with recently with our portfolio? But how do you think about customers that have dev teams that are understaffed, number one? So integration and timelines are more difficult? And then the second part of the question would be maybe internal dev teams that lack some capability. And so, you know, it becomes a difficult bottleneck in, you know, converting that booked ARR into recognize revenue. So
21:56
you’re asking, like, how can you accelerate teams like that?
21:59
What do you do in those situations? To accelerate everything?
22:02
Yeah, I mean, I think you, you put your CTO hat on, and you sort of ask yourself, like, what are the things that drive, you know, engineering, passion, or teams, and, you know, what I think can accelerate the team, the most, you know, irrespective of the skill level, is clarity, and like, clarity on what you know, the job to be done, what they’re building, and then maybe even more importantly, clarity on purpose. Like, why are you building this? Like, who’s it for? Like, why does the customer get excited and fall in love with your product? Like, what is it that you’re doing? And basically show them like the top of the hill and get them to follow you to the to the top of the hill. And I think that oftentimes, no matter what seniority level of the engineering team, as long as you got smart people on the team, engineers tend to rise to the occasion, as long as you show them what the top of the hill looks like. And you can show them what the clarity is, and what exactly they’re building. And in my experience, the teams that don’t do that are the ones that fail, right? Because engineers love clarity, right? They love deterministic stuff, they love to see exactly what it is that they’re supposed to be doing. And that’s what clarity is, right? And that’s what clarity provides. And so, you know, it might whatever, 25 years or crap, 2020 years, I guess, of working on engineering teams, 29 years, I would say it’s clarity is the one that actually makes things happen engineering teams, irrespective of skill level, is
23:27
that the same answer when we’re talking about an engineering team that resides within the customer. So there’s dependencies for your product, to get the resources on the customer side as well. I
23:38
think it doesn’t, doesn’t matter. Like, you know, good engineers, at the end of the day have a common denominator to me. And it’s clarity, right? Whether it’s clarity, for the customer, for clarity for you, or the product manager. It doesn’t, it doesn’t really matter as long as they’re good. And actually, that’s a great way to test whether you have a good engineering team or not. Yeah, because if that if you don’t drive super, if you if you provide that clarity, and you know that it’s there, and you can agree with other people that you’ve provided it, and you’re still not getting what you want, you probably have the wrong team. Perfect.
24:04
Tim, what advice would you give to founders that are building data first companies? Well, I
24:09
think the number one thing is you have one chance to not screw up data privacy and data loss. The moment you do that, you’re done. Like you can’t, you can’t recover. You just You just can’t you can recover from outages. You can recover from downtime, but you cannot recover from data loss, or outages. You just you can’t because people run businesses on data and it becomes like crack and people get addicted to it. And the moment you’re not there and you’re not reliable, you’re over. That’s it. I think the other thing is that speed kills. You know, you have to be fast these days. If you’re not providing, you know, basically real time analytics or real time responses. You’re not probably data company, you’re just not at this point in the cloud and the resources that are available for folks like you just have to be, you know, sub millisecond response time on reasonable sized squares. I’m not talking about full table scans but reasonable size groose. And then I think the other thing is just going back to what you’re talking about a second ago, focus on the usability, right? Like, it’s easy to be commoditized in the space. Right? Because know, the capabilities in the cloud are so prevalent. It’s it’s getting easier and easier to build these types of tools. And so how do you differentiate? And how do you create the moat? Well, the moat, in my opinion comes from usability and UX. And that’s actually hard to get right. So go hire some consumer UX people rip some people out of, you know, Airbnb, or Instagram, or what have you do exotic things like we did at Splunk, which is like, we created our own font, even just to drive legibility of data. So you didn’t like, confuse an L with a one when you’re staring at like a SHA 256. Hash. Yep. Those are the big ones.
25:43
Tim, why did you choose to do venture after being an operator for so long? Oh, man, good
25:49
question. You know, I grew up in the Bay Area, I grew up in Silicon Valley, my dad worked at a semiconductor company. And then as a child of the 80s. And I grew up with Apple two at home and like tinkering it with no instruction manuals and stuff and playing the games with a tape recorder plugged into the back. And like you said, I grew up with this, right, and you grew up programming, I started coding when I was like, six, and you sort of see this like, ecosystem around you. And you know, at some point, you realize, like venture is sort of like at the root node of how a lot of this sort of happens. And so, you know, I sat in the back of my mind, like, oh, wouldn’t it be really cool to be able to be a venture capitalist one day, it was always sort of lingering somewhere. And then you fast forward a bunch of years, and I was at Splunk, in a in a big role, and opportunities start to come up around me. And then my wife is a venture capitalist, she’s been at Sequoia for seven years. And so I’ve watched her be at Sequoia forever. I’m like, Ah, I could do that, that looks really fun. And it looks like a really, you know, like a fun team to be a part of. And then I started getting interest and thought, like, you know, what the hell, I think it’s a two way door, I’ll give it a shot. And, you know, maybe there’s room in the world for a technologist oriented VC to do well, and I’ll try and give it my my best and find out what I can do. And if I suck, and I, you know, I can always go back, I hope. So, you know, I give it a shot. And I think so far, so good knock on wood. But that was basically it was like, there was always a lingering lingering interest. And then the opportunities popped up and then fortunate enough to see my wife do it well, for a number of years and kind of learn the craft through the ether of of living with her
27:20
team. We recently saw that you hired Jeff Redfern, who was a CTO at Atlassian. Tell us about the hire and what what Jeff is focused on?
27:28
Yeah, so Jeff is very much focused on enterprise software, like I am, he’s also a recovering operator, he was CTO of Atlassian, as you said, and I think there’s an interesting topic here around operator oriented VCs these days, which seems to be a bit of a trend, but he’s looking at similar things as I am developer tooling, enterprise infrastructure, SAS software, you know, just like I said about myself things you would think of the ex CEO of Atlassian. And, you know, head of product at LinkedIn would would focus on so now you got, you know, a CTO and a CTO in the building chasing down similar stuff.
28:04
Give us the best debate that you’ve had with Jeff, about the direction of maybe a particular technology or, you
28:13
know, man, I know, you want me to come up with some like, like heartache on something that we don’t agree on. But I think, you know, we have very similar takes, right? Like, we love the same stuff, give
28:26
me something that you really agree on that maybe could be contrarian in the market.
28:32
I think it’s probably the UI UX piece, right? Like, I don’t think a lot of people really think of it the way that we do. And that may just be because we’re such product people. Like yeah, I’m a CTO and engineer. But I’m also very much a product guy. And I have a lot of empathy for the user. I care a lot about that experience a ton. And so I just don’t hear a lot of people going around on podcasts like this one talking about UX UI so much and like pounding the table on it. Like it matters, man, like matters a lot, right? Like, would you sell a house without a fresh coat of paint? It’s the same exact thing, and they just don’t hear it. And companies win on UI UX, like data, dog wins on UI UX, like at Splunk. I know we built the more scalable back end, it was faster was more I know it for a fact we measured it. And they still won. Right? They won because experience is better. The onboarding was faster, was better was richer is like faster time to value. Like all of that matters a lot right now, as things continue to get more commoditized one
29:29
of my first jobs out of school, I worked in product for this high precision motion company. So it did motion for like printed circuit board manufacturing and semiconductors. And that one of the biggest unlocks I got in there. And you had to be so sophisticated as a customer to use these systems. And just by adding UX we 10x in our market says,
29:51
see there you go, man, right, it’s yeah, maybe that’s the results speak for itself.
29:56
Amazing. Tim, if we can feature anyone here on The show Who do you think we should interview? And what topic would you like to hear them speak about?
30:04
I think you should get Dario who’s the CEO of anthropic to come on and talk about? How do you take on, you know, what are they doing to continue to take on open AI? I mean, those are the two big behemoths and the foundational model space. And I think like, that guy is hard to get a hold of, but he’s just not a big, you know, out there sort of extroverted creature like Sam Altman is in a lot of ways. Maybe he’s introverted, but I see him all over the place. You know, either way, he’s all over the place. Yeah, you just don’t see Dario a lot. And, you know, there’s a lot of chatter, that they’re doing some really fantastic stuff and building really great models. And it’s like getting him on the show to talk about that, I think would be fascinating. And get a lot of listens.
30:44
Make it happen, we might have to have a special guest host named Tim Tolley.
30:49
I’d be happy to do it, man. Happy to do it. Love it, Tim, what
30:52
book article or video? Would you recommend the listeners something in recent memory that you found informative or inspiring? And
31:00
like, I don’t want to say it. But like, I’m reading the Elon Musk book by Walter Isaacson, the new one, or Yeah, the new one. Like everyone’s reading it, but it’s like for those who are not reading it, like I do recommend it. It’s pretty good. Walter, Isaacson’s great. Like, he just does such a great job. And, you know, the window that it gives you into Elon Musk and sort of like, what from his childhood is created where he is today is like, simultaneously predictable, but just like shocking, some of the stuff that happened to him as a kid, and it’s just, it’s a great book, man. It’s well written. I think it’s great for any entrepreneur founder to get a sense, because like, you can’t argue with his outcomes.
31:39
Tim, do you have any habits, tactics or techniques that are a secret weapon?
31:44
Well, they wouldn’t be secret if I told you. So. No, I mean, I think I mean, as it pertains to venture, you know, I think it’s just, I use the stuff, probably in ways that most people can’t write, like, I have a datacenter. In my basement, I’ve got 342 u x of hardware that I deploy things against, you know, I can deploy clusters in my house and kick the tires, I use it programmatically deployed against use cases I have in my house, and just put myself in the, in the shoes of the buyer, I think in ways that most folks can’t, and probably won’t be able to, unless they’ve sort of been in the seat, like I had for a couple of decades. But you know, on the other hand, there might be some others, you know, CPLC to types, you go on the BBCs, like, my partner is off, and who will be able to use these tactics. But I think that’s sort of it. And then I think the other thing is just, you know, having ran large product and engineering teams over the past, like, I think you have to be decent with people. And I think having developed that muscle across, you know, 1000s, or 10s, of 1000s of people over the past couple of decades, perhaps helps me connect the two founders in ways that are different, right? Not only can I talk shop in ways that are that are different, but it’s like it takes one to know one, right? It’s like I understand their pain, I understand their struggle. It’s like, you know, I’ve been in your shoes, man, it’s like, I want to go on this seven, eight year journey with you and help you build this thing out in the winter. Right. And I think founders can feel that I think so. Yeah, I mean, those aren’t so secret. It’s not like I’ve got this like secret widget in my back pocket. I think it’s just more experience. Love
33:26
it. You know, I’m curious if you’ve ever invested in sort of like a certain type of model where, and I was just talking with a unicorn founder this morning about it. He’s currently building a an RPA company. You mentioned something earlier that made me think of robotic process automation. But this model where you find a technical founder, or maybe a product person within a large organization that has built something for that organization to solve a problem. But really, it should be built outside of the organization and could serve many, you know, examples could be like zoom and WebEx, have you ever done that before? And how does your you know interaction? If you have, you know, how does your mode of working with the founder different in that instantiation when you have you no more of an operator less of maybe a true entrepreneur.
34:16
So I have one company like that it’s called squint. And squint is this mobile app company that basically overlays information on augmented reality. So if you have something like standard operating operating procedures at a big factory, you can enable workers to take a iPhone or an iPad, it’s pointed at a machine and it says like, hey, go three feet this way and push this button, go this way, three feet, pull this lever, what have you. And that was born out of an idea we had at Splunk. Right? This idea of you know, could you get access to information using augmented reality by pointing it at a machine so you know, we took that modified it took the first or second derivative of it and did something slightly different. You know, if it didn’t borrow any of the tech, you know, we wrote it All from scratch, we’re all aboveboard here. But you know, that happened. And it was, it’s a huge success, that company is doing really, really well. We’re proud people should all go check out squint, because the technology is just like mind blowing. So Korea wound up doing a seed, you know, they have a Series A now that they’ve just completed as well. So like, they’re growing rapidly. And they’re selling mobile app licenses, which is like, almost unheard of these days, like, you just don’t see people selling licenses on mobile apps. So things like ripping. On the flip side, I would say, like, I don’t think that’s normal. It’s, I don’t know how you feel. But like, I think founders who don’t have that natural sort of ingrained DNA in them to be founders that you can’t make someone’s be a founder. Right. So I’m, I’m a little bit reluctant to take an NG star engineer out and be like, Hey, dude, like, I think you’re a CEO. Like they got to have that DNA, man. And you can feel it, like, you know, you can tell when people have like that blue blue flame, you can like see it in their eyes. And they, like I said earlier, like, I want them waking up in the morning, like, running through walls, eating glass to like, make crap happen. Totally. And I think I can’t force someone to be a founder. So like, unless I feel that, you know, I feel it through a handshake or like, you can see it in their eyes. It was like, There’s no way I’m gonna like, force them to start a company out of, you know, greatness inside of another company. For
36:16
sure. For sure, you can’t make somebody an entrepreneur. Yeah. Wonderful. And then finally here, Tim, what is the best way for listeners to connect with you and follow along with Menlo?
36:26
Man I know you want me to say Twitter. I am not a Twitter VC guy. I have not. I have not like or activated them. It’s not me. I do have Tim t as my handle on Twitter as early Twitter user back in 2006. On Tim T. You can reach me there or just Tim at Menlo vc.com works. I’m pretty responsive. Feel free to reach out.
36:47
Alright, he is Tim Kelly. The firm is Menlo. Tim, thanks so much for the insights today. Appreciate the
36:51
time. All right, man. Thanks for having me.
