The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch
00:22:50 · Harry Stebbings, Adarsh Hiremath
[00:22:50] S1: If you get further and further up the spectrum of we're able to 70, 75, 80, 85, 90? Do you not need fewer and fewer humans because the frequency is much less as you move closer and closer to perfection? [00:23:04] S0: That's a great question. And I think that begs the question of what do labor markets look like later on. I think the key thing is that the market will move towards specialty and sophistication, meaning the types of work that we see fifty years from now will be more specialized and oftentimes will require people with kind of like a higher level of sophistication in that specific thing. [00:23:27] S1: Can I ask you, when you sell to clients, what is the moment where they're like, wow, shit, we've got to use my core? [00:23:34] S0: When we're able to find exceptional people at the cost of software hundreds of times over. [00:23:39] S1: But when you're in that sales cycle with them today, when are they going, yeah, we've gotta sign up? Is it when they see the AI interviewer? Is it when you show them the price? Is it when they meet a candidate? Like, when when is that wow moment for them? [00:23:54] S0: It's usually when the first couple candidates start working with them. [00:23:57] S1: What what is that buying process? They buy one at a time? Is it on a per talent basis? Is it on a timeline basis? How does how does the deal with work?
00:23:57 · Harry Stebbings, Adarsh Hiremath
[00:23:57] S1: What what is that buying process? They buy one at a time? Is it on a per talent basis? Is it on a timeline basis? How does how does the deal with work? [00:24:05] S0: One interesting thing about Merkur, we don't have a sales team. There's not a single person who works, on sales at Merkur outside of, you know, the founders. These days, what we're seeing is mostly customer inbound. Folks have heard great things about Merkur from other people who have hired through Merkur and then reach out to us, and then we go from there. So right now, it's more of a a bandwidth thing than than any, like, tactical or coordinated sales motion, I would say. [00:24:32] S1: What percentage of hires is end to end done by software versus has human in the loop? [00:24:38] S0: On our end, the entire process is automated. This is everything from a candidate hearing about Merkor and going onto the Merkor platform via job listing, us pulling in their resume, their salary expectations, and whatnot, administering a personalized interview based on both their background and the role, allowing them to get paid for their work. That entire process is automated. [00:25:02] S1: What does a take look like on a per candidate basis?
00:25:02 · Harry Stebbings, Adarsh Hiremath
[00:25:02] S1: What does a take look like on a per candidate basis? [00:25:06] S0: It all comes back to quality. You know, I briefly talked about Uber and just going back to that example. Right? When I get into an Uber, there isn't that much of a difference between the 4.8 star driver and the 4.9 star driver, because the unit of work is not exponential. But with something like Merkur, there's a huge difference between the top point 1% and then the eightieth percentile person. So usually for customers, it's not a question of price. It's a question of quality. If we're able to find those point 1% people reliably at the cost of software, what what we take is often a second thought. [00:25:43] S1: And so I'm sorry. I'm... What what does that take then? Is it like a standardized take? Is it like on a case by case basis? What does that look like? [00:25:51] S0: It's on a case by case basis. For some customers, it can be, you know, over 30%. For some, it can be less. [00:25:56] S1: Over 30. That... That's great. Amazing. Well done. When you look at candidate completion rates, how much of that is India versus rest of world today? I know you specialize in finding amazing talent in India specifically.
00:25:56 · Harry Stebbings, Adarsh Hiremath
[00:25:56] S1: Over 30. That... That's great. Amazing. Well done. When you look at candidate completion rates, how much of that is India versus rest of world today? I know you specialize in finding amazing talent in India specifically. [00:26:09] S0: So the reason we started with India is because our parents immigrated from from India, Suri and I, so they went to these amazing schools. So we, like, started these recruiting campaigns from those schools specifically. And actually, like, one of the things that got us really, really excited about labor markets in general and the inefficiencies associated with it was just because, like, one of the best engineers I've ever worked with on our team. We found through a Facebook ad, and I manually interviewed him. Actually, he didn't pass the interview, but the reason that we ended up hiring him is he sent me a really, really long message about what exactly he got wrong in the interview and how to correct it. And I just felt like we gotta work with him. It was sort of that that prompted us to start in India. But if you fast forward to today, actually, the number one place that, you know, workers on the Merkor platform who, you know, have jobs through through us are from is actually The United States. [00:27:02] S1: Percent wise, is that 60% US style? [00:27:05] S0: It's high up there. [00:27:06] S1: Yeah. And client wise, all US too? [00:27:09] S0: Mostly US.
00:27:09 · Adarsh Hiremath, Harry Stebbings
[00:27:09] S0: Mostly US. [00:27:10] S1: Yeah. A lot of young exceptional people are being told to say that they shouldn't maybe study CS anymore because actually CS is becoming so automated. 41% of code is now written by AI. In five years' time, that'll be extortionately higher. Do Do you agree with that advice? And how do you think about whether or not young people should learn programming today? [00:27:31] S0: My take is that programming is actually more important today. It's just gonna happen at a different level of abstraction. One could argue that the leap from assembly to Python was actually maybe even a bigger leap than the leap from Python to natural language. My answer there is that the way we define programming will look very, very different. It may be a person who, you know, has like average skills by today's standards in in computer science, who's orchestrating thousands of superhuman coding agents to achieve more than we thought was even possible. But that skill set, which, you know, we can define as programming at a different level of abstraction, programming in English is gonna be super important. [00:28:11] S1: Can I ask how has the way that you program changed over the last two years?
00:28:11 · Harry Stebbings, Adarsh Hiremath
[00:28:11] S1: Can I ask how has the way that you program changed over the last two years? [00:28:15] S0: I I definitely use a lot of the the AI tools. They've gotten really, really good. A great example is Cursor. A lot of members of our team use use Cursor and love it. It makes doing things that would take a lot of time just so simple and elegant. Right? A great example is testing. With a couple of prompts, you can just generate a more thorough test suite than than, you know, anyone could have imagined for your application, for example. And that just wasn't possible before. Maybe it's like bringing the same consistency from, like, one part of the code base and refactoring it for another part of the code base. You can basically snap your fingers in cursor and it'll get done. And I think, like, the implication for software is that software is gonna get commoditized very, very quickly as these coding agents get really, really good. [00:28:58] S1: What does the world look like where software is commoditized? What does that mean? [00:29:02] S0: It means that people will be able to build applications much faster than was historically possible. It also means that the businesses that succeed in a world where software costs approach zero will be built on network effects. The companies that could even give away their entire code base and then still be alive. The marketplaces and companies like Meta, you know, and Airbnb that have built these really, really strong network effects will be the ones that thrive.