a piece of news I saw yesterday that had me scratching my head, which was the trillion dollar pay package for Elon. >> Elon Musk could officially become the first trillionaire. Just about a trillion dollars in stock. >> It’s a striking number, but the the benchmarks that he has to meet are also equally striking. >> He’s not just the leader of the company, he’s the marketing voice. >> If we really do expect to find ourselves in an abundant society soon, we should expect to have a lot of trillionaires in our society. money will start to have far less value than ever before. >> If we’re really on the the verge of abundance, then what comes after that? What’s going to remain scarce even as energy and intelligence? The the cost of both of those goes to zero. >> But stuff is changing so quickly. Media is changing quickly. Elon is paving a new path for what it means to be a a great corporate CEO. But it’s going to change again and it’s going to change again and it’s going to change again. >> A trillion dollars here, a trillion dollars there. How do you compete? What’s your mode? So, uh, Brian and Sid, welcome. Uh, pleasure to have you both.
[00:01:04] >> Now, that’s a Moonshot, ladies and gentlemen. >> Everybody, welcome to Moonshots, the news that really matters in your life. I’m here with my Moonshot mates, Dave Blondon, Alex Weezner Gross. Uh, and we’re going to have a conversation today about David versus Goliath. We’re talking about trillion dollar commitments, trillion trillion dollar pay packages. It’s insane. But first of all, one of the most important pieces of news is Alex. I’m reading the comments and what I keep on hearing is people want you me ask you a question have you talked for the entire episode. They love what you have to say. So, uh, I’m not going to do that today, but everybody, if you’re if you’re new to Moonshots, uh, Alex has been just, uh, receiving huge fan mail because of his brilliance. Uh, >> it’s an honor to be here, Peter. I think I got maybe two adoption offers. I think you did. Uh Dave, are you jealous, Dave?
[00:02:00] >> Am I jealous of Alex? I mean, I’m surrounded by so many people that are so brilliant. I don’t know. So, I’m used to it by now. >> That’s one of the most important things is finding incredibly brilliant people to have around you in life, right? The old saying, you’re the average of the five people you spend the most time with. And if you’ve got a community of individuals that are really uplifting you and challenging you that have you do the best you can, uh, that’s critically important. So, one of the things I love about Moonshots and our WTF episodes is sort of measuring toeto toe and trying to really have this conversation in a meaningful fashion. We’re missing Celestial again. Selee, we miss you, buddy. I think he’s back in India. >> He’s probably got, you know, a crate full of iPhone 17s coming, but we’ll be talking to him soon about that. So, I want to open up with uh a piece of news I saw yesterday that had me scratching my head uh which was the trillion dollar pay package for Elon for Tesla. It’s like that’s extra. Have you guys gotten
[00:03:00] a uh an offer from your boards for a trillion dollar? >> A trillion. Well, hey, if you hit the metric, the metrics have to be, you know, multi-t trillion dollar market cap, but sure, why not? It’s just a fraction of what you create. It’s uh if if Elon grows Tesla to an $8 trillion company, so it doubles the size of uh Microsoft and Nvidia, uh he’ll earn a trillion dollars. Not like he needs help becoming the first trillionaire on planet Earth, right? >> Yeah. But he’s he’s unique. You know, he’s really changed the definition of what it means to be a CEO. And he’s not just the leader of the company, he’s the marketing voice. You know, they most car companies will spend 7% of revenue or thereabouts on marketing. Tesla spends zero because Elon’s a oneman force of nature driving consumers to the product. And we’re going to meet a couple entrepreneurs, you know, in a minute here that have that same like we embrace social media. We’re creating morale and momentum like you wouldn’t believe. But, you know, that that’s a 20% pay package as opposed to the normal five. >> It’s it’s worth it.
[00:04:00] >> I I agree with you. If you’re a founding CEO and if you’re very very shy, that’s not going to bode well for the company. I know when I’m investing in companies, I’m looking for a CEO who’s a great communicator, who’s able to go out there in front of the crowd, able to convey their passion, what he or she is loving in life, uh, and for all of the insanity Elon does with chainsaws or whatever the case might be, uh, it grabs attention and people, uh, either love it or hate it. Um, and >> one of the many many reasons we love Alex so much is because he’s he’s not just brilliant, but he has very very high situational awareness. And that’s really rare around the brilliant community. But stuff is changing so quickly. Media is changing quickly. Elon is paving a new path for what it means to be a a great corporate CEO. But it’s going to change again and it’s going to change again and it’s going to change again. So, you know, if you map your behavior to the change, but most people study backward in time and they say, “Well, what did Jack Welch do?” or what did Genghaskhan do? You know, like it’s
[00:05:01] okay. But that’s not gonna >> Anyway, we’ll go on. This branch will go too long if I go too hard on this. I >> I also think Peter, it’s worth noting if if we really do expect to find ourselves in an abundant society soon. We should expect to have a lot of trillionaires in our society. >> We we will. Um and then we’ll have an expectation that money will start to have far less value than ever before. Right. I mean, do you you you and I have had this conversation, Alex, about a postc capitalist society, right? Thoughts on that? Do you still believe that’s going to be the case? >> Well, so I I think that that’s always the the question. What what does so-called latestage capitalism even look like to the extent the concept makes sense? Uh if if we’re really on the the verge of abundance, then what comes after that? And I I think the what comes after abundance is is closely tied to what remains scarce in an abundant society. In Star Trek, a common foil,
[00:06:00] energy is relatively abundant. Intelligence is relatively scarce. The ability to travel between stars relatively scarce. So the question I would ask is what’s going to remain scarce even as energy and intelligence the the cost of both of those goes to zero. >> That is a critical question. You know, my my end point here, my mental experiment is if I, you know, in Drex in Eric Drexler’s parliament, if I build a number of assemblers that are able to rearrange atoms and I drop an assembler into my hand and I say, “Hey, make me, you know, five copies of yourself and I give each of you an assembler uh and the assembler is able to use energy and matter resonant and build anything.” and I drop an assembler into the soil here and it starts pulling the atoms together to make me an electric Ferrari and it says I need a little bit of titanium, a little bit of lithium. You add it and all of a sudden you’ve got an electric Ferrari. Everything starts to become effectively zero marginal cost and that becomes a pretty cool society where anyone can do anything
[00:07:00] >> does it or I mean again not to overindex on Star Trek but in Star Trek everyone has replicators but not everyone gets to travel between the stars. So maybe that the new post scarce ability is the ability to travel outside the solar system. >> Well, we’re going to find out cuz we’re getting there really fast. On the uh news item of trillion dollars here, a trillion dollars there, there’s a dinner with Tim Cook and uh and Sam Alman and Mark Zuckerberg and Trump. Uh, and I guess during this dinner, uh, an offer was made by by, uh, was it Zuck first or was it Tim first to invest $600 billion into the US economy? >> I think it was Tim first. I can’t tell from the clips actually because they get cut and and mingled. >> But then the other one matches it and all of a sudden over dinner, Trump is getting $1.2 trillion of commitments into the US economy. Uh, he should have >> You missed the really fun punch line there. Like Tim Cook had it all scheduled, planned, and budgeted. And then Mark said, “I’ll match that.” It
[00:08:01] was It was like a YMC fundraiser. If you can do that, I can do that. I’m sure a CFO is back in Silicon Valley going, “What the hell did he just commit to?” >> Oh my god. But it is there’s a third piece that comes out in this uh related story, which is Altman announces to his employees that he expects OpenAI to be the most, you know, capital intensive company in history. Uh, and what was the number, Alex? Uh, 119 uh was $119 billion of additional investment between now and 2029. >> It was something like that. I mean, do you remember it was a whole what a year and a half ago or so that this this number of five or six or or 7 trillion of of capex into AI chips was being floated and a lot of people laughed at that. And yet and yet we’re finding ourselves a year and a half later in a world where it is entirely plausible that the the true amount of capital expenditure in fabs and AI chips and
[00:09:00] data centers and new energy sources completely exceeds that. I think yeah >> file file that away because you’re you’re dead right and and this is the effect we see all the time. Something insanely mind-blowing is predicted 6 months in the future. everybody’s like impossible then it actually happens and then they’re like oh yeah well it’s just part of life and this trend is is going to you know it’s happening over and over and over again but the numbers you just quoted yeah everybody was like oh Sam’s just blowing smoke there’s no way that’s that’s you know that’s trillion being thrown around but that’s not a real word that’s just sort of a a euphemism but and then here we are just a few months later you’re going to hear some benchmarks actually like SweetBench later in this podcast where things have just been crushed that uh the timelines will blow your mind. Well, we’ll get to it when we get to it. >> Well, then then I guess the point I’m making here is there’s a huge amount of capital flowing here. I mean, we never I mean, go back to when all of us were starting our careers in the ‘9s and in
[00:10:01] the dotcom era. The idea that’d be trillion dollar movements of capital in any particular company or any particular industry was just mindboggling. And here it is routine. Um, >> but this is in some sense like this is sort of a wonderful opportunity with trillions potentially of of capex being invested. There’s going to be an expectation I would assume by capital markets that there’s going to be enormous revenue generation that pops out of those trillions in capex. And the question you have to ask yourself is what form does that take? At some point with trillions invested, I think there’s probably a reasonable expectation that entire classes of labor of services are going to be automated and the cost of what we currently construe as labor is going to be driven down to zero. And then perhaps at some point immediately after that, you start to need transformative science, inventions, discoveries that will really justify the trillions of capex. So it’s sort of a a
[00:11:01] a blessing in disguise. I I would argue trillions of capex is going to motivate the the demand and the supply of utterly transformative discoveries and inventions soon. Otherwise, why invest trillions in this? >> Yeah. Uh the concern of course a lot of folks have is around inflation and not are these dollars really inflated dollars? Uh we’re going to find out. But it’s it’s it’s interesting Dave and in particular as a venture capitalist, you know, seeing the valuations of companies going at this level. I mean, for the average public, how do you get into any of these companies when they’re coming out at, you know, multiundred billion dollar and trillion dollar uh, you know, sort of valuations here? I mean, being able to get in early is one of, I think, the areas that you’ve been focusing on. Uh, one of the other companies that came out of Link Ventures, you were early check in here, uh, was Merkor. And I
[00:12:00] just saw Merkore has gotten a offer at a $10 billion valuation. Um, you know, you must be pretty happy about that. >> It’s a $10 billion valuation, but it’s also a half a billion dollar revenue run rate uh, after two years, which is completely unprecedented. So, >> so go back when when did you invest in Merkore? >> Uh, two years ago. first funding um you know right out of you know what what you really want to look for is undervalued underappreciated talent and not so much concepts >> uh but they had the concept right already it’s it’s rare but 18 years old you know I mean that’s not not a lot of people invest in the 18-year-old gang >> so so a couple 18 year olds come forward with this idea do you remember what the opening valuation was when you invested >> uh 30 million plus or minus 5 >> okay so 30 million to 10 billion in two years time Yeah. Yeah. It’s gota it’s got to shatter all kinds of records. But again, I I don’t want people to feel like that’s a bubble because the revenue growth also shattered all kinds of records. Yeah. >> And so from a cold start, I don’t think
[00:13:00] anything like that’s ever been done before. But you’re going to see a lot more of them now, too. They they just are setting the trend for many many other companies. I think what’s different about them is they’re inspiring a an age of people that normally would be would have been uninvestable 5 years ago, 10 years ago, and now it’s kind of a wow mainstream. We’ve made that point that the average age of a unicorn back VC backed unicorn a decade ago was sort of mid30s right in terms of the average age of the founders and today I think Dave what you found out of the investments we’re doing especially out of MIT and Harvard it’s uh age 20 to 23 and these guys were 18 when they started >> yeah 18 uh they got through one year of college uh and then they got frustrated with the pace like everybody because >> which goes >> they met in they met in high school >> which which goes to the point Alex you made last time on the last WTF episode which was listen if you really believe we’re sort of post AGI on the verge of
[00:14:00] you know ASI advant you know advanced super intelligence going to college during those years and trying to get credits versus building something uh it’s not the right trade >> it’s going to distort all sorts of societal cues and societal expectations that the the best uh I would say fiction treatment that I’ve seen of this is a a novella by Verer Vinci, Fast Times in Fairmont High, where you see this start to completely distort the way secondary education is is run in this country and you start to see high school students and middle school students suddenly spending all of their time doing startups. And I I think it’s entirely plausible we find ourselves in a near future that looks a lot like >> I agree. And one of the one of the points here I think that we need to realize or people need to realize is the MAC you know sort of peak creativity if you measure it by when a Nobel laureate does their Nobel Prize winning work or not when they get their prize but when they actually did their work is
[00:15:01] typically in the first half of your 20s. Alex, do you have the data there at all off the top of your tongue? not at my fingertips and I’ve seen those statistics too and I’ve seen how they vary from field to field purportedly math versus physics versus other fields. I I also tend to discount this notion because I I expect that in the very near future most of the innovation is actually going to come either from pure AIs or some sort of human AI hybrid. So I I I view those statistics maybe self- servingly as more of a a retrospective. This is how things used to be at best versus how they’re going to be in the future. Every week, my team and I study the top 10 technology meta trends that will transform industries over the decade ahead. I cover trends ranging from humanoid robotics, AGI, and quantum computing to transport, energy, longevity, and more. There’s no fluff, only the most important stuff that matters, that impacts our lives, our companies, and our careers. If you want me to share these meta trends with you, I write a newsletter twice a week, sending it out as a short two-minute
[00:16:00] read via email. And if you want to discover the most important meta trends 10 years before anyone else, this report’s for you. Readers include founders and CEOs from the world’s most disruptive companies and entrepreneurs building the world’s most disruptive tech. It’s not for you if you don’t want to be informed about what’s coming, why it matters, and how you can benefit from it. To subscribe for free, go to demand.com/tatrens to gain access to the trends 10 years before anyone else. All right, now back to this episode. Well, there’s another company I want to talk about here, and we have a couple of guests to join us. Uh, if you’re an entrepreneur, uh, listen to how they built this company. This is a company on the doorstep of being a unicorn itself. Uh, a company that you’re going to hear a lot about in the coming years. Uh, Dave, you want to introduce our guests and and Blitzy? >> Ah, God, I I cannot wait. So, we have, uh, Brian Elliot, Sid Pardes, the founders of of Blitzy. Uh my uh my son Jack interned with them this summer. Uh
[00:17:02] and I tell you it it drove my wife a little nuts. She started thinking, “Wow, we’re going to play a lot of tennis this summer and have a great time.” Jack got so uh wrapped into the culture of Blitzy so quickly. It’s the most high energy place I think I’ve ever seen. Morale is off the charts. Uh and so he pulled in his best friend from high school, Yash Blesetti. They pulled in a couple of other young computer science majors at Northeast or Northwestern and and a couple other places. The whole gang uh worked all summer on SWE and crushing numbers, but I tell you the morale of this group is like nothing I’ve ever seen. The mission is incredibly cool and fun. So, can’t wait to tell you all about it. So, uh so Brian uh he uh is a West Point alum. My one experience in life with West Point was Rick Dzel, who was my biggest and most important customer I ever had. actually he ran all things complicated at Walmart. You know, massive logistics, half a million people moving around and then he got poached by Jeff Bezos to work at Amazon. Uh so he
[00:18:00] he was the number two guy at Amazon right when they were in total hyperrowth and he had a West Point background really understood morale, people, logistics. Uh and then you know Brian went to Harvard Business School after that. Sid went to bits which is you know the MIT of India. It’s actually statistically much harder to get into than MIT if you can believe that Peter. I can. >> So, top MIT let MIT let me in. So, there’s got to be some some flaws there. >> Oh, so humble. Uh, so, uh, so Sid after that was spent a long time at NVIDIA and saw it go from from tiny to monstrous. So, that’s got to be inspiring. And I actually don’t know their story uh before before that, but they met at Harvard Business School. Uh, and they’re I think inspiring to a different class of people, you know, that were already in a career path. and then AI hits the world, but they’re nimble. You know, they’re not going to watch it happen. And this is way too rare. You know, people remapping their entire life to take advantage of what’s happening right now. So, I hope a lot of the listeners today get a lot out of their backstory
[00:19:01] and their transition to building this incredible company, Blitzy. >> Yeah. And I really want to frame the story here as David versus Goliath, right? We’ve heard about a trillion dollars here, a trillion dollars there. Uh, and how do you compete in that world? Right. If you’re a young entrepreneur, um you’re building a company and you’re wondering are you going to get literally decimated in the wake of Google or OpenAI or XAI just, you know, happening to release a particular feature? Um how do you compete? Uh what’s your mode? So, uh Brian and Sid, welcome. Uh pleasure to have you both. Um and uh where are you guys? Where are you guys this morning? I am in one Kendall Square here at uh the link studio offices with MIT right behind me. So we just walk over the talent from the MIT AI lab right over to here to work which uh which works well for us. >> Beautiful nearby in Cambridge. >> Fantastic. >> Um the first thing we had to overcome was convince Dave that we could be successful when we’re not 19 years old.
[00:20:00] So he totally flipped his paradigm on on these young people. And I said, you know, Dave, when I was like when I was these young these kids’ ages, I was, you know, I was out, you know, across the ocean fighting at war. And uh I think I get enough experience to hopefully have a second career here in in technology. Uh but uh Peter, I love the question that you’re asking, right, which is how on earth do you compete with these frontier AI labs, with with Google, right, with open AI? And there’s really two reactions you can have when there’s these trillion dollar investments, right? There’s the reaction where you’ve built a company that you say, “Oh no, right. They’re going to steamroll over me.” Y >> and then there’s the reaction where when every single model gets better and the combination of those models makes your product much better, you’re jumping for joy. So we just got a trillion dollars of R&D cry for Blitzy and we’re >> rising a rising tide and you can float on top of all of that. That’s perfect. But the but it’s critical to to find that product market that is able to
[00:21:02] benefit uh for the from the rise of these technologies. Um >> I think you know we this is the second time we’re seeing this happening. So I’ve you know I was at Nvidia back in 2016 and uh I heard all about the story of KUA and I saw Jensen believe in that. That was pre Gen AI the term Gen AI didn’t exist right and I’ve been working on genative AI models since the attention is all you need paper came out so Jensen was asked to stop investing in KUDA. He started this back in 2006 right and it was negative to the to the company. So but he still invested in it. He still believed in AI. He worked with researchers and built the technology to solve problems that he foresaw and that is very relatable to what we’re doing. So Blitzy if if you we’ll go into this data for sure but it’s very unique in terms of how the product is built. It is specific to the enterprise and it is based on the opportunity that we’ve seen over many years you know working at
[00:22:00] largest companies like Initia. Um, you know, I think we should begin for those who don’t know Blitzy explaining what it is. Want to get into like, you know, where were you guys when you said, “Aha, we’re going to build this, right?” I love that story. And then I’ll I’ll unleash Alex Squeezer Gross on you to ask the most intelligent, important questions. >> Well, well, let me tell you what it what it does today, and then I’ll tell you the humble beginnings of all of this, Peter. So, uh, Blitz is a an enterprisegrade autonomous software development platform. So we ingest and understand up to 100 million plus lines of code where most single LLM tools are stuck with this finite ability to understand context. We’ve developed some really unique context engineering systems to understand enterprise scale code bases. From there an enterprise will express their work from a development perspective. Whether that’s a cobalt to Java upgrade very common in these old financial service institutions that use us or steadystate development work. Blitzy will send off the most comput intensive workload in the entire
[00:23:01] AI code generation space. We’ve done a 12-h hour run. We’ve done multi-week runs for massive scale code bases. Ultimately delivering high quality prevalidated, pre-ompiled, pre-ested code, right? And our view, our thesis is we want to increase the quality of code at any cost because the other side of a pull request that comes from AI codegen is human labor which is exponentially more expensive. And so that’s really the view that we have enterprise scale high quality code. >> All right. I want you to slow I want you to slow that down for a second. Um, which is to say a lot of companies out there, a lot of traditional companies in particular in the finance world, you’re saying insurance world have large code bases. They have software that they have uh in, you know, have inherited for how how old are some of the software systems that you’re playing in? >> I mean, we’re talking about uh PL1. >> Oh my god. >> OAL. These are like these are these are
[00:24:00] old school financial service institutions that uh quite frankly for a long time have been afraid to touch the code because the cost to get something modern just wasn’t worth it. >> Okay. So you got you’ve got a company you’ve got a company out there running cobalt uh from what 20 30 years ago. >> Yeah, that’s right. >> And and their their system is operating. It’s working. It’s not doing anything significantly um useful given that it’s 20 or 30 years old. And do they have people that can still, you know, patch that code? I mean, are there engineers still around? >> Out. They make they make you and Dave look like spring chickens. >> We are spring chickens. >> Oh, yeah. I know you’re a longevity guy, Peter. You look great. Uh yeah. >> So you’ve got this problem where where you’re too you’re too scared to touch it uh because you’d have to do a wholesale replacement. >> And so they call in the Blitzy guys and
[00:25:00] you come in and you’re able to do what for this 30-year-old chunk of code? >> For starters, we give them visibility. So we’ll ingest, index and understand the state of their underlying source code of which times they they rarely have somebody that understands the entirety of you know tens of millions of lines of code. It’s actually an impractical problem to know that >> and then we allow them to execute large scale transformations. So whether that’s uh getting onto a modern technology stack or that’s adding required functionality. These businesses, these enterprises are stuck with the inability to layer artificial intelligence on top of their existing codebase because it is so old, so antiquated, and has so little visibility in what they’re doing. >> Massive value, massive value creation. >> It’s a mind-blowing experience, too. If you take 10 million lines of unintelligible, undocumented code and you run it through Blitzy and then you say, “Tell me what it does in plain English. Explain to me where there are bugs.” It’s just you’re talking to the code. It’s mindblowing. >> Like one of my favorite one of my favorite uses of Gen AI is to give it
[00:26:01] some patent that is you know unintelligible and say how does this what does this do and how could I use it in my company right the ability to take something that’s complex and make it understandable to gro it fully to use that term. >> Thank you Robert Heinline. Um >> well Sid’s got 27 patents so I had to use that to understand what the heck you did at Nvidia. >> So so Sid you had invid you at Nvidia for five years? Um, eight years total. Eight years. Uh, yeah. Six years. >> I heard I heard Jensen recently say that most of the executives there are now are now billionaires. Did you make it to that status? >> Well, I held on to my stock. I rode the wave from like double digit billion to trillion dollars. But then I got two degrees, Peter. >> And then what happened? >> I I I got two degrees from Harvard. So they took all the money. >> You won’t know way. >> Wait a minute. >> Oh my god. Let’s not get start let’s not get started on the on the expense and value of degrees from Harvard or MIT or
[00:27:00] any any of these I league schools. >> Um so this is a David versus Goliath story and I I want to understand you know you’ve got incredible success. Uh but before we get to that story uh you guys are both at Harvard Business School. When did this when did this idea germinate? What was the positive agent that said, “Okay, let’s build that.” What was that founding story like? >> Yeah, if you can rewind the clock back to the GPT3 to 3.5 era, right, where these things could code, uh, but it wasn’t what we’re experiencing today, right? At that time, Sid and I were doing a proono project for our favorite local bakery here in Boston, uh, as a part of our time at Harvard. and they mentioned they’re about to spend $300 $400,000 on a new mobile app which got our attention uh some young enterprise and entrepreneurs and so S and I went home uh and we did what is now you know two and a half years later called vibe
[00:28:00] coding uh and we built them the application overnight right but literally over the weekend >> over the over the night yeah yeah during the week and uh which which now is like no big deal but if we >> did they pay did they pay you uh you know $200,000 for that >> we we acted like it took longer. I think that was our that was our first mistake. Uh but was so clear during that time is that Sid and I were actually the bottleneck for development. And so we were you know you get a you get an error and you feed it back to the system and then you give it to a different model, right? And then through that practice you’re able to get much higher quality code. And so we said if we could just invent a system where all the commoditized development work could be removed, right? And we could have multiple models going back and forth iterally refining and getting to code that compiles and tests. That is going to be what the future looks like, right? And we learned that by doing by being hands-on and then having an idea of what the enterprise needs from SID experience and building towards that. >> Were you guys already friends or did the
[00:29:00] allnighter making friends? >> Yeah. But yeah, we’ve only gotten I mean Sid’s actually the the godfather to my to my uh youngest son, believe it or not. Okay. Investing in best friends is uh is pretty good. >> I mean, and and that’s another part of the story, Dave, that we’ve talked about is some of the most successful companies are when best friends get together uh and and just build 24/7. It is a you know, I I say this to the entrepreneurs that I that I coach, you know, being an entrepreneur and having co-founders, you’re going to spend more time with your co-founders in the trenches than you do with your husband or wife or kids. It’s an intense period of time and you better pick somebody or some buddies that you love spending time with. >> Yeah. I always tell people it’s imagine you’re on a long international flight, one of those 14-hour flights, and you’re sitting right next to somebody. Do you walk off the plane feeling good and and having fun or do you walk off the plane not waiting to get away from this person? >> Uh your startup’s going to feel just like that every day. It can be work or it can be fun. It just depends on the
[00:30:01] personalities and the match. >> So Dave, what what do you find most exciting about Blitzy? just uh from as investor and so they they came through link through link studios and uh and link investment uh and yeah tell a little bit of that story if you would. >> Well they they’re you know definitely much uh more experienced than a lot of the entrepreneurs around the studio. So they had the plan and the idea fully baked on arrival. Uh we were still first money in and still gave them space and support and have made a bunch of introductions but they already had it more than figured out. Uh so that’s you know that’s not all the companies fit that profile. Um they’re also you know very different a lot of the companies coming right out of dorm rooms will do image generation. They’ll do you know they they don’t understand the word cobalt or PL1 to save their lives. And so when I look across the range of business plans that are right in front of us with AI more of them fit into the you need to understand the domain space than you could just think of it in a dorm room space as maybe 2/3 oneirds rep numbers. And so what I’m hoping with
[00:31:01] Brian and Sid is that they inspire a ton more people to go after these, you know, these are still multi-trillion dollar markets. Uh but they’re not, you know, AI girlfriend. They’re not, you know, apartment search while you’re in the Yeah. Another photo sharing app. They’re they’re, you know, and they they get really deep. You know, you’ve got all these, you know, manufacturing, you know, semiconductor manufacturing automation, you know, that’s that’s really deep. You get insurance actuarial risk adjustment. That’s very deep. This one is actually nice in that it’s very, you know, code generation is very broad. So, it’s a huge market, but it’s also deep in that, you know, like refactoring 10 million line code bases is a is a pretty deep knowledge set. Um, the other thing that’s that’s really cool about Blitzy to me is that we have all these code generation products. So, I use Cursor, we’ve got Windinsurf, Replet, Lovable Companies are all worth billions now. Uh, but there, you know, I write a line of code or I tell it to write a line of code, it creates a button for me. I say,
[00:32:00] I don’t like that button, make this other widget. And, you know, you’re doing it in real time. >> But you can’t build something really big, you know, and when you put cursor in full agent mode, it’s right in no man’s land. It sits there and grinds for like five or seven minutes, which is too long to wait, but too short to build something substantial. >> So, they’re getting stuck in no man’s land. Blitzy just said, “No, we’re going all the way to the other end where it’s going to run all night long or all week long like Brian was just saying and come back with something really big.” And that’s just fundamentally a much different engineering problem than what love lovable repleter winds are doing. It’s just it’s a different kind of company. I don’t know of any other company that’s there. >> Amazing. Well, we’re here today to announce uh a particular piece of news as well. Uh some some groundbreaking news. Uh is it Brian or Sid? Which one do you want to talk about what you guys have accomplished? >> Sid please. You are the inventor of the technology here. >> Sid. So >> tell us. >> Um you know so every time a new model
[00:33:01] comes out they benchmark on this leaderboard which is called Swebench verified. The leaderboard itself was built by OpenAI. It is a subset of Sweetbench. Uh it contains 500 problems that were wetted by the researchers at OpenAI and they confirm that these are solvable problems and these are worth testing models on and it’s been ubiquitous. So every time a new one comes out, you always see results. The current top of the leaderboard on the Sweet Bench website as of filming is 75.2%. So we hired a bunch of extremely talented interns. So So Dave, going back to Jack, if we could hire him today, we would. Uh that’s how good some of these interns are and every single intern who worked for us, uh they were amazing. We were, you know, we really credit this to their effort and to Nirj who led the uh you know, efforts on our end. But we ran Blitzy on Sweepbench. And as it turns out, these are 12 repositories, but they have 500 branches and that equates to 400 million lines of code if you ingest them on Blitzy, right? U we’ve ingested
[00:34:02] anywhere from 1 to two billion lines of code overall on Blitzy depending on how you count it. Uh cuz you count updates and the whole raw thing as well. So it’s 2 billion lines of code if you count all of the updates, right? Including Sweepbench. Um so we ingested all of that. We ran Blitzion on solving the problems and our final result u accounting for everything that we’ve tested and verified using SBCLI was 86.8%. That is uh that is a significant jump over the current leaderboard on the website and the last time this was done was when Devon had a 13% jump from 1% to roughly 14ome percentage. So we’ve come a really long way with the system and the primary reason that we were able to achieve this you know echoing some of the points that we made earlier is we’re very different from the way existing tools are structured right so one thing is you can reproduce these results in production using pity right we’ve not
[00:35:00] added any custom scaffolding just for sweet bench we’ve not tampered with any of the features to achieve this we’ve seen reports from you know some of the other labs that claim that even though for example let’s say the latest frontier model claims 80% on stream if you actually run it and reproduce it, you get 60%. Right? And we wanted to not have that problem. We want we we we care deeply about reproducibility and the practical real world applicability, right? Which SweetBench verified has been vetted to be good at. Um so you can reproduce these results and um it’s it’s live as of today. >> Amazing. Hey, Alex, help us understand how big and important this uh this particular hallmark is for the for the company and for the world. Uh give us some background here. >> Sure. Well, well, first to to Brian and Sid, congratulations on your announcement. Uh I I think there I would expect there’s going to be an enormous amount of interest from the community once they they hear these results in in trying Blitzy uh and in reproducing
[00:36:02] those results. So congratulations in advance on the onslaught of interest that I expect you will receive. I I think to to answer Peter’s question, I I think software engineering is arguably the first major vertical of human labor that is very high economic value, very high productivity that is perhaps succumbing to automation. So any sort of step function improvement in software engineering is arguably super transformative to the global economy. And and maybe just pivoting on on that thought, one of the the first things that uh that I was wondering when uh when I heard that you would be announcing these results and maybe jumping back 8 months, we all remember when Deepseek uh aka Highfly launched R1 January of this year and there was sort of an aha moment all around the world. They they didn’t just announce a new reasoning model. They announced and maybe this got a lot less attention.
[00:37:00] They announced a bunch of new open- source libraries at the systems level, like a new file system. So, one of the first things that that I was wondering when when I learned that you’d be making this announcement is the the whole world is is sitting on this sort of palumst of legacy libraries and operating system code, billions and billions of of lines of code, Linux, Python, GNU, all of these libraries. Is there is there something that that you and Blitzy and uh and this new remarkable capability that you’re announcing can can do to to speak to what can we do to improve the performance of this entire tech stack that the whole world runs on at this point? >> That’s a that’s a fantastic qued question. You know, we’ve been running some of these experiments. We’ve been taking some of these open source libraries that for example was written in mat lab for one of our customers that they were using and we converted it to Python. Mat Lab was written 20 years ago. It was specific to Windows and we
[00:38:00] made it OS agnostic. We’ve we’ve run these we’ve run these PCs all the time where you know we go from OS agnostic to OS specific to agnostic from traditional to modern but we’ve also been running other kinds of PCs where for example we picked an Nvidia repo. We identified an issue that was marked as open um and we just put blind at it and we solved we created a pull request that solved the issue. Right? So if you think about that problem, you can identify bugs, issues, feature requests in any of the modern frameworks and systems and like like we the system is not limited by how much code how big the repo is, right? So you can send blitzy add it, it’ll come back with a solution. You don’t like it, you can iterate over it, you can create five projects, get five different pull requests, see how that works and deploy it uh all you know within a matter of days. I think that’s a fundamental shift that really is going to change the way people work with open source and also closed source technology. >> Amazing. What what is the largest repository of code that you’ve actually tackled?
[00:39:01] >> Brian, you want to say? >> Yeah, go ahead. So, we’ve we’ve we frequently see 20 million lines, but I think the absolute largest we’ve seen is about 60 million lines that we onboarded successfully. >> Crazy. And how long, just just for comparison, for fun, if you had to guess, how long would it take in terms of human labor hours to to do that? >> It’s insane. >> It’s insane. Brian, I think we frequently scope these, right, as part of the PC process. And I think you you probably speak to that. >> So what what we do So there’s there’s a question of like how long would it take to grock 60 million lines of code? And yeah, the the reality is it’s just too big for a human to understand. So you might pay, you know, Accenture $100 million in three years and they might come back with some some diagrams over the 60 million lines of code and by the way million >> by the time they did that what they came back with would be out of date. >> Exactly. Which is why you can see that you know essentially the industry has been stuck. Uh this is why your your airlines are always uh misouted and they can’t get their software updated is this kind of fundamental problem. So, um, we every time we run Blitzy, we actually,
[00:40:01] uh, estimate for the the clients in production, all our enterprise clients, how many hours were automated. And the CIOS love this because it’s like the KPI that they go give the board on how many hours they’ve automated away by their intelligent vendor selection. Um, but grocking is sort of an impossible problem, but the the real value is in the code generation, right? Being able to accurately affect and develop code and accelerate that life cycle for the development team on that large underscoring corpus. Uh but importantly like uh and Alex, I know you want all all developers to sort of go away and and and we’re going to live in a society of abundance, but I think it’s kind of going to go in the opposite direction where you know there’s almost an an infinite demand for for code, right, and for software development. And so Blitzy doesn’t do everything. Uh Blitzy does about 80% of the quantum of work on average for these large scale problems. Uh but it knows exactly what it doesn’t do, which is really the power. And it kind of hands off that batch of work to the human developers to finish things out. So we get a really clean full uh pull request plus human labor at the end uh to accelerate the development but not sort of remove the need for uh you know
[00:41:01] the developers altogether. So just if I may just to pull in the thread, Brian, I mean I I I would argue we’re about to enter sort of an age not necessarily of of just abundance but of great projects when it’s possible to send lot basically let lots of automation loose on the world and fix all the problems, solve everything as it were. In the case of of Blitzy, this is letting AI agents loose on an enormous sprawling legacy codebase and just fix everything. I think it’s a good name for the episode. Solve everything >> or or or a book or or or you know, fill in the blank. But it are are you familiar? Maybe you’re tracking this project. I I love this idea. The great refactor. This is a project. Yeah. Yeah. I love this. >> What is that, Alex? What’s the great refactor? Refractor. >> So So the great refactor. I I love this. This is like a classic solve everything concept. This is we’ve built our whole civilization on a bunch of software libraries that could be better
[00:42:01] maintained that are filled in many cases with legacy memory vulnerabilities. There are statistics out there that most of the the insecurity of present- day software is due to the way the software is written that exposes them to certain type of of cyber security vulnerability, memory vulnerabilities. And if we could only rewrite all of these libraries that sort of if you know the the meme that goes around of the the entire stack being built on uh of civilization being built on just like one block it hangs by a thread. If we could rewrite all of these libraries and dependencies and software supply chain upstreams that our whole civilization depends on in Rust or some other memory secure language. Suddenly that that would fix almost all that would solve everything in terms of so many vulnerabilities. I think I had like 200 customers send me that project. Like they just like immediately saw that it became hot news the day and they all sent it to me like, “Oh, are you guys going to do this?” And I said, “Well, are you gonna are you going to pay for it?” Like the thing you think it’s worth it.
[00:43:01] >> Here’s the most critical thing, right? If if you think about this idea where you can give these projects, refactors or whatnot to AI and have it come back with the code, right? You can do that with any chatbot. You can give it to any AI, have it write code. Getting code from AI is a commodity, right? But if you add constraints to that problem where the code needs to replicate the existing functionality, it needs to compile all the unit test needs to pass it. It should not have newly added security vulnerabilities and all the other items that are you know crucial to the enterprise or the or the problem itself right that uh make it valuable. That’s when the challenges begin right now. That is not something you can achieve. So Sid, my question is I’ve got 3.2 billion lines of code uh which is my genome. Um can you recompile that for me? Can you like go and and identify more, you know, fix the fix the broken parts? I I would find
[00:44:00] >> LM can as long as LLMs can write as long as it’s in the language that LMs are trained on understanding, we can do it. The scale is the problem that we’ve solved for. And the other problem we’ve solved for is making sure the requirements match so that when we put you back together or edit you, you you actually look like you, right? It’s it’s validated that it is you. We didn’t change or break something that we shouldn’t have. >> Alex, what do you think about >> Yeah. I So maybe narrowly on the bio. I I I there are many other projects that that speak the language of the genome and the proteome that I I think Peter for rewriting your genome you’ll have the opportunity over the next few years to to use one of these biological sequence-based foundation models to to do some variant of that. I I do want for for Brian and and Sid though really pull on the economics of this. So, so I I really want to press you guys when when uh when we talk about the great refactor or some of these great projects to basically rewrite the the source code basis for much of our civilization today
[00:45:00] and and you think about the economics of that and it there’s a school of thought that that says we’re seeing generative AI hyperdelate by 10x per year or so an order of magnitude cost reduction every year. At what point in your minds do you think using Blitzy or or maybe competitive tools does it become reasonably economical to basically rewrite all of the legacy code out there that civilization depends on? >> I mean I would argue from a value perspective it’s there today because the value that this would be providing to society is just dramatic. Now this is a this is a question of like uh who’s the payer? Uh a line of code from Blitzy is 100x more cost basis than a line of code from sort of any other provider. uh which is uh maybe 100x less than it would be from a human developer, right? And so we’re talking about huge orders of magnitude difference. Uh and so would it be worth it from a society value to rewrite all the software today with like absolutely um but am I going to continue to serve these financial service institutions and insurance companies first that are uh readily paying me
[00:46:01] today? Like yes. And so I think if you if you grab the uh if you grab the capital funding for us to uh to break even on this AWG, we’ll start rewriting all the uh we’ll rewrite Linux if you need to. I want to insert I want to insert another topic in here. You know, you guys shot a really cool podcast. Uh it’s on your LinkedIn. Um where you’re just bantering between the two of you about the fact that the the definition of truth within large scale software has always been the functional code. You know, here it is. This is the final thing. It runs the PL1 code that does all the nav accounting for the mutual funds, you know, over at State Street. It’s like millions of lines of legacy PL1. But that debugged code is the core asset. And then the documentation is just something around the edges. Post Blitzy, the truth moves to the documentation because you can regenerate the code overnight anyway. And so your your actual core asset has moved from code to a document, but it’s going to move again. And this is where it was really cool to hear you guys bantering around like well what what is then the foundational truth of this piece of
[00:47:00] because because like Alex is saying the entire infrastructure of society is about to move and also expand 100 or a thousand or a million x you know because because code is so cheap to create all of a sudden we have much much more of it so you got a much bigger world but the the ground truth is some other format than just you know PL1 or cobalt or Python code it’s this human readable today spec becomes the the central asset. That’s a big shift. >> The source Yeah. the spec is still an abstraction layer, right? Um and so for the that’s easy for the human to look at, right? The real the real source of truth or understanding is actually we we create a customer specific hybrid graph vector database that understands exactly what is going on from a functionality perspective. And you could change that functionality from one language to another, but we are capturing the core essence of what is required there. And we can display that as a spec which is 200 pages. Uh but of 20 million lines of code that’s a intermediate representation. Uh and really we want to
[00:48:00] get back to the the core DB level understanding that’s the core asset for these folks and and and Blitzy makes it. It’s the it’s the property of the enterprise. >> Everybody there’s not a week that goes by when I don’t get the strangest of compliments. Someone will stop me and say, “Peter, you’ve got such nice skin.” Honestly, I never thought, especially at age 64, I’d be hearing anyone say that I have great skin. And honestly, I can’t take any credit. I use an amazing product called One Skin OS01 twice a day, every day. The company was built by four brilliant PhD women who have identified a 10 amino acid peptide that effectively reverses the age of your skin. I love it and like I say, I use it every day, twice a day. There you have it. That’s my secret. You go to onskin.co co and write Peter at checkout for a discount on the same product I use. Okay, now back to the episode. Brian, given your background, you know, in the military, I mean, probably the one institution that’s got the, you know, largest repository of of ancient
[00:49:01] code has got to be the US government, right? I mean, so can you attack all of that? I mean, unlock massive productivity. There’s a a lot of fear that the US is a falling empire. uh its inability to you know understand and and and legislate efficiently. I mean couldn’t you have like a just a single massive impact on on the US government? >> You know, Peter, I uh I live in Backbay here in Boston and and so does this uh lead investor at at Incel or so he tells me because he keeps running into me uh when when I’m walking to work and and just bumping in and seeing how things are going, which has me skeptical, but but the short answer is short answer is yes. It’s like the US government is a sort of a fantastic end customer uh to to ultimately modernize to get your flights there on time to uh to make getting ids easier, right? And so this is like absolutely part of critical infrastructure that is a target customer that we have certainly. Do we have them today? No. 12 months will be serving them like I think yes. >> Yeah. I mean that could catapult you uh
[00:50:02] into a you know deca billion dollar company easily just landing that kind of a customer because once you’ve modernized you know for one of the agencies they’re all going to want it. >> Yeah. I think we have the most top secret security clearance to patent ratios uh of any uh of any company out there. So >> fascinating. >> I’d like to maybe pull in the theme that that we’ve talked uh about on the pod in the past uh the elephant in the room which is recursive self-improvement. So how much of Blitzy is written by Blitzy? >> Uhuh. >> A lot. So it’s it’s interesting, right? Where you this is actually a question of like where’s the core value? Um and I would say every single sprint that we do from a software development perspective is driven by pity. Uh and so I would say a significant amount of the the corpus of the code is is driven by us. Now there are uh algorithms that are not about writing code, right? they’re about uh core invention and so I think most of the companies you know core IP is not
[00:51:01] going to be the software that it creates but it’s going to be some core invention that this is around you think about Google with page rank right page not actual pages but page rank is really the the core source of their original IP and SIDS invented a number of algorithms at the core of blitzy that uh allow us to for instance always compile code to never have circular dependencies so you could actually rewrite sort of a corpus of blitzy with blitzy uh pretty readily uh and pretty quickly and it would look like it. Uh, but would it do what we do? The answer is sort of sort of no. And and it gets the question like what is the source of IP for companies? And it’s it’s got to be a breakthrough or an invention. >> What we’re doing, right, for these large companies, we’re telling them for many of them, we’re telling them how to use AI, right? We’re coaching them and consulting with them on how to use Blitzy. You cannot do that unless you’ve actually used it yourself and perfected the process, right? because they’re not just making the engineering velocity changes or the you know tool usage changes they’re also making the process changes and that’s why this is critical. >> Alex I’d love you to take a second and
[00:52:00] and dive into the SWE bench metric here. uh again if you could for those who are not familiar you know a little bit of the origin you said from Princeton and and and ratified by open AI but what is it measuring and and who were the top of the leaderboards before and then I want to get into the conversation about how do you compete against the mag 7 or against you know sort of the frontier models in this regard so let’s contextualize it first understanding what what the sweep bench metric really is >> sure so for context >> and why don’t we start with the title of the white paper cuz that’s going to drop same day that this podcast does. So, so people need to find the paper. >> We should we should put a link in to the white paper in the in the show notes here as well. Alex, please. >> Sure. So, maybe let me just speak to to Sweepbench. So, Swebench uh is is a benchmark that measures the ability for AI systems to solve typical software engineering SWE for SUI tasks uh in in the specific form of responding and
[00:53:01] solving issues on GitHub uh very popular source code management system. So, SweetBench as a whole, not SweetBench verified, consists of couple thousand instances of tasks in which the the the central challenge that’s posed to an AI is to respond to uh to an issue in a codebase. And what would happen in a a normal software engineering context is >> what kind of what kind of issue, Alex? >> So, a wide range of issues. uh could be bugs that need to be fixed uh other performance issues and the the usual workflow in in a software engineering context is an issue will be identified and uh pull request will be submitted. So identifying an issue, responding to an issue, submitting a pull request that responds to an issue, satisfying unit tests, the these are all standard parts of what would be uh archetypically
[00:54:01] considered software engineering. And Sweepbench, I think, is sort of an excellent industry standard at the moment that attempts to capture the life cycle of of high value ad labor that a typical software engineer would would perform. I’ll let the the guys respond to to their white paper. >> Sure. Yeah. And so I also I’ll respond by also answering your core question, Peter, which is uh how does one compete with the mag 7 in this utterly important labor task, right? And the reality is there’s a significant amount of the mag 7 at the core of what Blitzy does. So we use Gemini’s models, we use Enthropics models, we use OpenAI’s models, but the uh uh what’s unique about this technology moment in time is when you use these models against one another, the quality moves up quite dramatically. And when you use them against one another hundreds of times, right, in hundreds of different combinations, which hundreds of different tool sets and prompts, the combination goes up
[00:55:00] even more exponentially, right? And so really it’s the art of orchestration through what we call extended inference time validation to move up the quality of code ensuring at every moment in time the system has the right context to operate despite the large scale underlying codebase. So we say we’re excited when Gemini or or OpenAI release a new model like our product gets just dramatically better every single time. I mean that’s a really important insight right building so that the better your your components are the stronger you are as a whole and um that’s a unique niche when did you realize that I mean that’s sort of a fundamental for you >> I think we realized it back when we were just initially figuring out this first project together but I’ll let Sid expand on this >> I think we made a bet and we said that uh there’s there’s not a doubt like we we we were building this when models had 5,000 tokens of context right u and we made a bet that there’s no doubt that context Windows are going to expand and the models are going to get better at writing code. Now, do we want to go and compete with the with the Mac Max Mac 7 and build our own model or do we want to
[00:56:01] stand on the shoulders of giants and use the technology to solve the problems that they’re actually meant to solve and that’s exactly what we did. >> Amazing. And just for context again, who had the record before you? Who had the record last week? I think there were some open source labs and there’s also been some other you know unpublished um reports that have claimed um you know around 80% but the 86.8% that we’re claiming has is unprecedented and the highest number we have seen. >> I think bite dance was the most recent tray model to to be at the top there. So you know we we can rebrand this US versus China if you want to. Every time I hear 80s something percent, 90 something% I’m thinking that these benchmarks are getting super saturated, right? Uh and so where does this go next? I mean, what are you going to measure when you’re at 100%. >> Yeah, we talk about this in the white paper. Like this is um we we certainly need new benchmarks, but the the reason
[00:57:00] we didn’t do sweet bench verified for a long time is it’s just not representative of the scale of problems that most people use Blitzy for. So the typical pull request size is like 100 lines of code and the largest repository is a million lines uh within this and so we really need a set of benchmarks that’s on you know Linux and VS code which is 20 million and and 4 million lines respectively uh with with hold out sets against those trying to do larger scale work uh to to ultimately show how far we can push the bounds on on autonomy. >> Yeah. And in terms of Peter’s saturation question, uh, the paper does a really really good job of describing the landscape of benchmarks and Swebench in particular and the need for a new benchmark. But one of the points it makes is that when when you score 86% on this benchmark, that’s effectively very close to 100%. Cuz the remaining subset of questions are just are just flawed. Um, they’re not harder, they’re just not they’re not structured well. Mhm. >> And so you’ve basically capped out this
[00:58:01] benchmark now. >> So folks want to look up the paper. What’s the name of the paper? >> Do you have it? >> Yeah, we Alex is urging us to retitle it. Uh so I’ll give you the subtitle which is which is uh domain specific context engineering paired with extended inference time validation breaks the barriers of LLMdriven software development. So that’s that’s really what we’re talking about at >> that rolls off the tongue and onto the floor. It is a technical favor. >> Yes, >> you know, Sid, Sid, you can search Blitzy in Sid’s name. He has a searchable name. Brian Elliot, there are thousands of Brian Elliots, it turns out, >> but Sid Pardeshi plus Blitzy will get you to the paper. >> Okay, perfect. Alex, where do you want to take this next? What what have you found uh important and fascinating about what Blitzy is doing? What’s the implications in the long term, buddy? >> It’s so interesting. So many different directions to go in. I maybe just to go back to this idea of great projects because I I think Blitzy has the
[00:59:00] potential to be sort of an embodiment of an era when we just again turn all the AI agents loose on all of the problems in a discipline. So what I heard I I think Brian you say a few moments ago was something like 100x price difference per line of code or or per file between Blitzy with its you know again congratulations uh state-of-the-art performance announcement and other competing tools. When I hear you say 100x price difference I immediately uh internally say oh well that’s just 2 years worth of of cost hyperdelation. So you’re sort of two years more expensive than the competition is is the way that I heard that. So if you project forward two years, three years, four years, do you think we will find ourselves in a world where AI really has, you know, the the great refactor has been completed and we’ve rewritten all of our foundational systems with Blitzy or maybe copycats of Blitzy? Do do you think we find ourselves in in a near
[01:00:01] future like that? >> Let’s say Pontificate first here. I think um you know I I think we will we will see the models get significantly better at doing this and the cost go down right the the key thing that I would like to underscore is you know a lot of the approach at some of the labs and then what what what’s happening right now is the labs aren’t really making money on the inference that they’re running right for all the models u but what we’re doing is because we’re using the labs and we’re able to charge a premium right for the work that does and also provide the validations with it. We’re not losing money on the uh code that we’re writing, right? So as this equation improves over a period of time, the difference that Blitzy is able to create is going to also grow. So I I I think somewhere in that double negative, I I heard the the answer is that that yes, as as hyperdelation kicks in, call it an order of magnitude cost
[01:01:01] reduction per year, maybe more. Uh not only does Blitzy become very profitable, but also becomes very feasible to start to tackle these solve everything level grand challenges in in software engineering. Um so so I I’ve been attempting to kick the tires on on Blitzy myself. My my first project with Blitzy was I wanted to rewrite Python, the the very popular programming language. And and I I I gather you, Brian, and Sid, you you’ve had access to your your own product longer than I have, which has only been uh 2 days. Have you tried to take some large scale project? I think Brian, you mentioned Linux a few minutes ago. H have you tried to take some large project and either say gosh I I want to ask Blitzy to improve performance by 10% on some relatively mature codebase or add some crazy transformative new feature. Have you tried that? Yeah, we did this project
[01:02:01] where we’ve done a number of these Alex but the the one of the most fun thing we did we onboarded VS code um and we said hey add a chat experience to VS code at the time there was uh you know cursor I still is right one of the biggest uh tools out there uh we tried and we did we built one of a subset of the features of cursy and we tried using that internally, right? Uh so anytime we consider SAS products at this point of time, we’re trying to first see if we can replicate that internally using Blitzy and if we’re you know a few months out or a few years out, we’ll say hey let’s just use the SAS products as a starting point and uh consider rebuilding it later on. I think a lot of enterprises will do this. H have you thought I mean so so sort of uh free marketing advice uh before the public. Have you thought about taking all of these open source projects that are in many cases starved of core
[01:03:01] development team members and and hungry for human capital or human capital equivalents. Taking these projects and just aggressively setting loose the AI agents to submit very friendly, very polished poll requests to to these projects to to launch improvements. We’ve actually done that for ML flow, I believe, right, Brian? >> Yeah. Yeah, we’ve done this for especially for for one that enterprises specifically rely on. It’s uh it’s quite possibly the best BDR, which is uh which is just sending pull requests to open source libraries. Um and we uh there’s an open source one right on the homepage of our website that I think you’ll find fascinating, Alex, which is uh it goes all the way back. We were talking at the beginning of the episode. So AWS invented this uh or created this repository specifically to be incredibly messy mainframe code, right? So all different styles to represent, you know, different decades of people working on it. Uh and to ultimately say like how how would one use code generation to be able to move this from target cobalt to to target Java. Um
[01:04:00] >> Oh, cool. >> We ran that through Blitzy because you know mainframe is such a big problem in all these large even government organizations and we moved it from Cobalt to Java completely autonomously with the ability to compile right out of the box. Right. Now there’s rem sort of like remaining development to work on some runtime stuff, but this is a this is a sort of like a multi-yearlong project to move mainframe from a target messy cobalt into Java. And uh and the results of that have have been probably one of the best uh business development tools that that we’ve ever created. >> How fast did you do it? >> Uh how how long did that take? A couple days. That was a Yeah, that if you count everything from start to end, it was a week’s worth of inference >> essent. Yeah. >> Amazing. So, so if we project that going maybe if I may back to Peter’s question about what the human equivalent of this is, do do you have any metrics that that you can point to for either cost savings relative to to humans for a given unit of code, a line of code or per file or how much faster than humans this this
[01:05:01] is. In general, >> we we typically see from a speed perspective a 5x velocity difference when this is brought into the enterprise. And the biggest challenge is really like operational deployment, right? And so you’re used to sort of starting your work uh the same day that you pick pulled up your IDE. And so what we’re having these organizations do is sort of start the sprint for the next development work the week prior. And so instead of developers starting with tickets and tasks, they’re starting with code that’s mostly written and a project guide with all of the human tasks to begin. So we really focus on this 5X. Anytime we engage an enterprise to work with us, we say pick a real world real world project that you have upcoming next quarter. Let us prove a 5x difference and we’ll, you know, we should do the remaining development work so we have an endto-end solution. And if we do that, then you adopt across the org, right? And uh it’s incredibly successful because people don’t realize the cost of coordination and development and and all of the sort of uh requirements to actually get a piece of full software out that when you can offload a significant chunk of that to agents, the the velocity gains are
[01:06:00] honestly unbelievable. Brian, >> that’s a really cool insight because a lot of the younger teams are going into enterprises and then they’ve never been in enterprises before and they’re they’re saying, “Look, the raw raw code generation is a thousandx 10,000x.” >> You’re like, “Yeah, but what’s it going to do for me in my enterprise?” And there are very few that are credible in saying well we actually have done it and we we know the final answer and right now it’s 5x apparently but but they don’t know like the the overhead of the you know the the organization and the the documentation and the just getting you know all of the things that that happen before you can even start code generation. And so it’s it’s really nice actually to have at least one vendor that understands how to get the real thing. You know we we actually need this to work in the end. It can’t it can’t just be a hypothetical thousandx. It’s 3 years from now. We’ve got digital super intelligence has landed. Um, it’s come out of, you know, I’m going to put my bets on Google, but we’ll see. Uh, what does Blitzy look like? >> Yeah, I think Blitzy is the core system of record and system of action for
[01:07:02] software development in the enterprise environment, right? And so the source of truth moves from documentation uh uh and and code which is sort of like this hybrid today to organizations relying on Blitzy’s hybrid graph vector database that understands a core functionality and organizations are going to be able to move incredibly quickly from a software perspective and the source of value is going to be sort of core IP uh that is not easy to replicate. Mhm. >> To add more onto that, you know, Peter, there’s always going to be some tasks where it’s better to have a human in the loop and do them sequentially, right? You’re solving a problem that has never been solved before and you need uh quick feedback from AI, right? That’s always going to be there where you use the co-pilots, but there’s always going to be this other category of tasks where you can automate them away, right? Build the code, run it, deploy to production, and execute maintenance. Blitzy now, you know, gives you the code. The code is the final output, but we’re going to go into auto autonomously maintaining,
[01:08:00] deploying, and keep keeping the applications running. Uh, so you’re not going to need humans for specific sections of the entire enterprise. It’s all going to be driven by AI. >> Interesting. >> If you could, I think Alex was about to paint a kind of a two-year view. Okay. He asked a question about what’s the force multiplier today, but then I think we were going to next segue into Okay, but there’s 100x and another 100x coming. So, I would love to finish that thought, Alex. >> Totally. Um, so I there’s a lot of to to Dave’s point, uh, and and Brian and Sid, um, I think you were starting to to gesture in this direction as well, there’s a lot of interest in the the benchmark out of of meter that’s measuring the effective time of autonomy, the the characteristic time scale over which AI systems, including AI codegen systems, can basically operate without a human intervention, sort of like a disengagement with a driverless car. how how far can it drive without a human needing to to take the wheel as it were. So I’m I’m I’m curious like have you thought about the characteristic time scale over which
[01:09:00] Blitzy is able to do autonomous codegen or the human equivalent really of autonomous coding before which the human needs to step back into the loop and be involved. Right now I if I remember correctly the the the current state-of-the-art is something like 1 to 3 hours. there’s a a a nice very clean on a at least on a semi-log plot expectation that and I think we’ve discussed this previously if you project it out a decade or two we we get to many many years and perhaps hundreds of millions of years and in a few decades where does blitzy fall in this apparent exponential trend towards exponentially increasing times without humans needing to be in the loop >> I think this is a project >> if you if you think are all the pieces needed to achieve this right let’s let’s take a very small example let’s take the AWS example right there’s the part where you identify the requirements uh decide what you need to do get the code and there’s the part where you get it all the way to production right and if you
[01:10:00] look at those parts each of them have already been automated in isolation for example CI/CD how do you deploy the code to production you have automations for that um debugging security analysis monitoring in production tracing and viewing the logs ensuring that the system is not doing anything malicious all of these These items exist today and there are you know these glue layers that we are now seeing like for example MCP A2A right that allow agents to form this mesh and automate work the only thing that’s left to be done really in my opinion for these projects is to just connect the dots and that’s exactly what we’re working on so to answer your question you know how far out I would say we are months out actually from delivering projects completely autonomously as long as they meet a certain set of criteria and conditions. So what I just heard you say Sid, correct me if I’m wrong, is that the documentation writers, the spec writers are the new limiting factor for the speed of software development. Is that correct? >> That is correct. >> Great insight. >> And h how do you think about automating that process if at all?
[01:11:01] >> That is also automated. So if you if you go to if you go to Chad GPT, right, and you ask it to write documentation, it will following your criteria. The reason we have document, you know, writers in the first place is to have quality, right? uh we we we have concerns that models lose context over a period of time uh and they skip and omit things or they can be gamed into you know adding things that you don’t want and we’re really concerned um for quality control which is why we have humans but as you can see we’ve solved the context problem for large code bases and there’s nothing really stopping anyone for that matter from effectively adding in the right safeguards and you know layers of protection to ensure that we minimize the need for humans I think it’s a matter of us becoming comfortable with AI doing that and I definitely see that happening over the coming months >> to try to >> begging >> begging for a follow-up white paper because Alex’s question is infinitely recursive right if you said okay well then then that’s not the constraint then what what you know because there’s always going to be a constraint is turtles all the way down you got to just
[01:12:00] ask okay >> speed of light buddy speed of light >> yeah Douglas Adams right that famously pointed to that the problem is is far more far harder to pose than than the solution. And to the extent that the new limiting factor is the spec writer or or the the prompt engineer, whatever we end up calling it in in the future, I I I really would like to to press you sit on this like when do we get our automated program manager, product manager, spec designer, documentation writer if if that really is the limiting factor for the speed of software engineering in the near future all the way down. >> I I’ll tell you something, Alex. You know, you know how the Blizzy platform works. We have these thousands of agents and each of these agents has has a persona. There is a product manager agent. There is a software architect agent. There is a QA agent and there is an agent that writes the prompts for the other agents. >> So all of the challenges that you’re describing are live today in production with the platform.
[01:13:00] >> This is such an important question though because you know I know it sounds very hypothetical but you know look at this timeline on Swedbench here. This was only 18 months ago that you got like 12%. >> And and now it’s saturated. It’s only been 18 months. So, you know, like what we think of as the dis distant scientific future. You know, it’s all science fiction. It’s only a year, a year and a half in the future. >> Star Trek’s coming, buddy. >> It’s really really hard to anticipate. Yeah. >> Now, I love this question. This is another white paper waiting to happen. I want to wrap this episode with a conversation amongst all of us on a particular topic. We opened up talking about trillion dollar pay packages, trillion dollar investments, you know, uh numbers that are extraordinary and the sovereign funds, the venture funds, family offices are just supporting this with massive capital inflows. And so the question is that I put to uh to all four
[01:14:02] of you uh think about competing in the long term with the Mag 7 who’ve got this incredible access to capital. How should founders consider going about that? Right? What’s your advice to others who are are getting in here during this, you know, this period of exponential growth in the AI economy? How do you compete? How do you think about that? I think you want to be a large customer of those folks as well. Uh I mean we we are major customers of of all the uh AI Frontier Labs and so they’re quite quite excited that we’re going to continue to push the bounds of autonomy and their market paths are going to continue to grow probably dramatically uh in line with that return on investment and you know Blitz is going to ride those waves as well and so um if if you’re happy when they’re successful and uh they’re happy when you’re successful then I think you’re in a pretty good strategic position. But there’s and there’s one more thing you know to add on to that what I’d like to go back to what Dave said right Merkor for example it was able to do that because it went deep I
[01:15:00] think that’s also the case for us right we’ve seen the enterprise Ben and I from the enterprise side of the challenge and the enterprise perspective and the security roadblocks and the product roadblocks and the process gaps that stop them from taking the full advantage of the product. So if you’re an entrepreneur or a founder and you’ve seen this personally and you’ve struggled with this problem and you think you have a solution that addresses the core of that and you’ve been able to test that with the actual enterprise and demonstrate effectiveness, I think you’re holding on to something that is core, right? >> So you’re saying >> understand the problem deeply. >> Yes. >> Understand the problem because look, you have these max seven, they’re giants. They have all the money. That’s fine, right? You’re an entrepreneur. You’re nimble. you can find the right investors. We were grateful to find, you know, Dave who believed in us the moment we we pitched it and we were able to get just the right amount of capital to get started and that’s really all you need. If you have the right talent, the right amount of capital, and you have the right problem that you’re going after, that you’ve convinced about because you’ve, you know, experienced and solved
[01:16:00] it, then you’re going to be so nimble and make these moves and get a product out that is significantly better than anything that the Max 7 can put together because they’re struggling with their own challenges like bureaucracy and, you know, all of the hurdles that they have to go through to actually put out >> politics, struggling with what what to say over dinner with Donald Trump. >> Exactly. So while they’re distracted all of that, you can build a kickass product, get it to market, solve real world problems, and you’ve you’ve changed the world effectively. >> Alex, what’s your thought? How do you how do founding entrepreneurs uh compete with companies? I mean, I remember famously, you know, Amazon was out there as a platform for people to sell their products, but then when Amazon saw a product that had incredibly high, you know, margin and uh and growth, they would clone the product and compete directly. How do you keep from that happening? Two words, solve everything. >> The name of this episode, solve everything. >> The world is filled with so many
[01:17:01] problems that a startup standing on the shoulders of the trillions of dollars of capex that are being invent invested in cloud, AI, chips, fabs, energy are now poised to solve so many problems, thousands of problems. I I think Brian and Sid and again congratulations on on the benchmark announcement are well poised potentially to solve the problem that we face of decades of civilizational software croft legacy code that’s just piled up without enough human capital to invest in reinventing it and now I think we’re arguably on the verge of doing that that’s one of thousands of problems entire domains that can be solved protein folding was solved by alpha fault essentially overnight transforming a subset of structural biology. So many more opportunities. >> Before I go to you Dave, I just want to remind people, you know, I define an entrepreneur as someone who finds a juicy problem and solves a juicy
[01:18:00] problem. And the more entrepreneurs in the world, the more problems that get solved, the better the world is. It’s why we’re going to hit on this over and over again. I think the career of the future is being an entrepreneur, finding a problem, falling in love with the problem, not the solution, not the tech. Because uh if you understand the problem deeply, as the tech evolves and continues, you’re going to use the newest version to go and solve that problem. And again, some of my favorite lines, the best way to become a billionaire is help a billion people. And the world’s biggest problems are the world’s biggest business opportunities. So that’s what entrepreneurship means. Dave, you see hundreds and thousands of companies. You’ve got how many companies right now in the link studios? >> Yeah, 28 in the building and about 50 total. >> Amazing. uh what do you when you’re looking to invest in a young entrepreneurial team like uh Brian and Sid or like the founders of Merkore or again some of the incredible unicorns that that we’ve backed out of uh out of
[01:19:00] link exponential ventures? What are you looking for to make sure that that company isn’t going to get disrupted in the wake of a of a, you know, open AI or or Google slight jog to the right? >> You know, it’s funny. Uh Kevin Wheel, we asked that exact question in that podcast we did two weeks ago. Uh and he answered exactly the way I had hoped he would answer, which is in a world where the foundation model companies get to AGI and can do virtually anything, are you just going to take over the world? And you know, Kevin was really clear that maybe we can do that, maybe we can’t. We probably can’t anyway, but even if we could, we don’t want antitrust to come in here and break us up. You know, that’s the last we want a huge thriving ecosystem of partners that give us money. You know, is Blitzy one of those companies that gives us money? Yes. Therefore, they’re our best friend. Go conquer the world. Take over. Change the entire foundation of all legacy codebase. Make a trillion dollars and give us half of it. We’ll all be happy.
[01:20:00] that’s what they want. >> I took a 2-hour walk yesterday with a dear friend of mine here who runs a large venture fund and uh we’re talking about the notion that his bet was, you know, Google had so much more capability than they unleashed and they said, “Look, it’s an open AI. Go and do as much of this as you can because we need someone out there competing with us, otherwise we’ll get broken up for antitrust reasons.” Um, which is a fascinating idea. you need you need viable competition uh to help you price to help you remain on the edge to help you not be you know sort of broken down by the government. So >> and be a good partner and that that was when Google was growing like crazy and we had all these portfolio companies we made a made a ton of gains. Uh but be a good partner to Google while they’re growing like crazy and now it’s the foundation month. Just be a good partner. Talk to them all the time. Make sure you know where they’re going and they’ll love you. >> Amazing. I I have a a selfish question. I I don’t know if we’re running out of time here, but >> No, that’s fine. It’s Let’s We’ll close with your selfish question.
[01:21:01] >> Okay. Okay. Well, this I’m always looking for trades like this has obviously been one of our best investments ever. Uh and the sky’s is the limit from here. Um and I’m always looking for trades of success and the morale at Blitzy is like nothing I’ve ever seen, you know, which is not a no-brainer. when you’re doing video generation for a movie studio or whatever, it’s easy to keep high morale, but when you’re doing, you know, 5 million lines of code core cobalt conversion, but yet you guys have just this crazy thriving culture and and Sid mentioned, you know, we were we’re first money in. I don’t remember why we loved the deal so much. I do remember we absolutely was a no-brainer to invest in you guys. So, two things jump out at me. One one of them is Bits, which is just the hardest place in the world to get into, and in video, which is, you know, you’ve seen growth. The other one is Brian. I think you had Army Ranger Bangalore Institute of Technology. >> That’s the BA Ba Institute of Technology. >> Okay. >> Yeah. He has a cool name though. Bits is like MIT bit but it’s bits you know. >> And it was you know it was by the way MIT designed the curriculum for bits. So that that that statement that was
[01:22:00] actually true. >> Oh that’s cool. The other one though was uh you know Brian I think uh you had uh not just West Point but Ranger training which is freakishly hard uh and then uh first boots on the ground in Syria. So literally the first people touching a war zone. Um so I got to feel like there’s something in those experiences that puts you a cut above in terms of building a team, managing logistics, building morale. So any clues there that other founders can pick up on and and >> yeah, >> I think some some evidence of of being incredibly missiondriven and ambitious is is what you would see if you were an anthropologist looking at both of our backgrounds. But if you take me for instance and if you fast forward or I guess rewind to to 2017 when I was serving in the 75th Ranger Regiment, the the mandate was, “Hey, go into Syria. Uh there’s about 2,000 ISIS fighters in that hold Raqqa. Uh we’re going to send you with 100 guys. uh recruit everybody else and take back the city, right? And
[01:23:00] oh, by the way, we can’t let anybody in the United States know we’re here because we’re there covertly, right? And uh and to be able to sort of sort of go in and solve that problem, like that’s a very ambitious undertaking where we’re like conquering cities isn’t something that like most people have spent their their time doing. So when you look at sort of the the level of ambition of of the company, everybody here at the business at the business of what’s uh has that ethos, right? And the very first thing we do when we interview is we screen for ambition and the ability to invent and create. >> We have those core values. And if you if you if you talk to any single person that sits in this building right here, uh they will tell you and they’re right that what we’re doing is one of the most important things they will do in their lifetime. Because the economic expansion that the globe gets, the GDP expansion that you get from automating software development or at least huge chunks of software development, there’s almost no better incremental use of energy than driving towards that goal. >> Wow, that’s a beautiful thought. Are you guys a 996 or 997 shop?
[01:24:01] >> It’s Saturday today. We’ll be as tomorrow. Yeah, we’re we’re a 997 kind of crew. >> Oh my god. >> Just just to back Dave up on his on his 997 job. >> That was a mistake. Well, I I hope Right Dance was the last leader on Street Vengeance Verified so we can we can uh we can we can work longer than them and beat them on the leaderboard. So, >> oh my god, guys. Listen, congratulations on hitting that new benchmark. But more importantly, thank you for the work that you’re doing uh from the companies that will benefit from our government that will benefit from the world that will benefit. You know, this is you’re upgrading the DNA of uh of industries and of our planet. So grateful for you, Alex. Dave, any closing thoughts here? >> Uh, I’m I’m just super excited to to see what you can Yeah. what what you guys can bring to the future. I would very few things would excite me more on the the software engineering front than few years from now to to learn that the entire software stack that that I run
[01:25:02] that um that companies that I I work with run has been 99% rewritten by Blitzy by Blitzy’s agents to remove all the vulnerabilities improve all the performance. I I think it it’s the the sort of challenge before you guys that that sets us on the road to recursive self-improvement and abundance and also solving everything in in software. >> Solving everything. That’s my That’s my phrase for the day. Let’s solve everything. >> That’s a good catchphrase. It’s better. It’s better to to Dave’s 997 than Tang Ping. Tang Ping the opposite of 996 lying flat in response to overwork. >> Dave, let’s give my closing thought. Definitely everybody read the white paper. Uh the title may sound very complex but the paper itself is very very readable. Uh so please read it and then your takeaway will be wow okay now we need a new benchmark inside baseball blitz is already working with MIT to create the next generation benchmark. Uh
[01:26:01] so but but catch up to what they did right here by reading the paper. >> To all our subscribers thank you for following Moonshots and WTF episodes. Uh we’re grateful for your time. We hope that, you know, in spending the time with us, you’re able to understand how incredibly powerful this technology is for transforming our world, our lives, creating a future of abundance. I hope this counters all the dystopian news you get on the 6 and 7:00 news. That stuff I don’t watch. This is the stuff I focus on. I hope you do, too. I’m grateful to my moonshot partners, AWG, Dave Blondon, Seem, wherever you are transiting the Atlantic to come back here to the US. Uh, and again, Brian and Sid, congratulations on your epic wins. Excited for your future success. Every week, my team and I study the top 10 technology meta trends that will transform industries over the decade ahead. I cover trends ranging from humanoid robotics, AGI, and quantum computing to transport, energy, longevity, and more. There’s no fluff, only the most important stuff that
[01:27:00] matters, that impacts our lives, our companies, and our careers. If you want me to share these meta trends with you, I write a newsletter twice a week, sending it out as a short two-minute read via email. And if you want to discover the most important meta trends 10 years before anyone else, this report’s for you. Readers include founders and CEOs from the world’s most disruptive companies and entrepreneurs building the world’s most disruptive tech. It’s not for you. If you don’t want to be informed about what’s coming, why it matters, and how you can benefit from it. To subscribe for free, go to dmmandis.com/metatrends to gain access to the trends 10 years before anyone else. All right, now back to this episode. [Music]