Most founders get one founding story. Shensi Ding thinks Merge has three.
Shensi started Merge in 2020 with her Co-founder Gil, watched their first customers die off, then rebuilt themselves twice for the AI era.
I had Shensi walk me through the playbook behind their night shift 5-11pm AI transformation. Not many founders talk this honestly about tearing up their company. Thereās lots of lessons here for others trying to do the same.
We get into why integrations turned out to be so important in AI, almost hiring a foreign spy, why founders with an EA are moving too slow, going straight for enterprise customers on day one, the Embarrassment Framework, vibe coding a dinner bot that 10xād their customer events, and why 6% of Merge employees get married.
Whether youāre looking for ARR or love, this conversation has it all.
Support this Episodeās Sponsors
Numeral: The end-to-end platform for sales tax and compliance.
Flex: Sign-up for Flex Elite with code TURNER, get $1,000 here.
Amplitude: AI analytics, all you have to do is ask.
Merge: Every model. One API. Total control. Check out Mergeās Agent Handler.
Monaco: The revenue engine for startups.
Inquire about sponsoring future episodes here.
š Stream on Spotify and Apple
Timestamps to jump in:
0:00 Something broke every year since 2020
2:22 How Merge went all-in on AI on nights and weekends
5:40 New launches got faster
8:32 Building connective infrastructure for AI
10:51 The dinner where Merge started
13:11 Six months of research before a line of code
19:00 Almost hiring a foreign spy
20:48 Mergeās 6% marriage rate
22:50 Hiring enthusiastic, nice, smart people
28:02 Early stage founders donāt need an EA
31:27 Shortcuts are a mentality
33:01 The AI tool that 10x'd their customer dinners
38:33 Marketing became an engineering function
40:51 Starting with SMB and climbing the logo ladder
42:07 How the product went cross-category
44:16 Launching Agent Handler and a new pricing model
46:47 Every product should be multi-model
49:58 Why most MCP servers donāt work
53:31 Startups should go all-in on enterprise
57:00 The hardest things are most defensible
1:00:26 Say "psycho shit" to be memorable
1:05:47 The Embarrassment Framework
1:10:23 Frank Slootman and being okay with being disliked
1:12:48 Giving feedback got scarier at 100 people
1:14:01 What only the CEO can do
1:19:44 The best marketing is not doing what everyone else does
Referenced:
Find Shensi on X / Twitter and LinkedIn
š Stream on YouTube, Spotify, and Apple
Transcript
Find transcripts of all prior episodes here.
Turner Novak:
Shensi, welcome to the show.
Shensi Ding:
Thanks for having me.
Turner Novak:
Yeah, thanks for being on. I think probably the coolest, most interesting thing we could talk about right now is that Merge feels like itās had multiple founding stories. Most founders donāt get multiple founding stories. So whatās the second founding story of Merge?
Shensi Ding:
Yeah. We started in 2020, which was probably... I know everyone says whenever they started was the worst time, but every single year something bad and crazy has happened since 2020.
Turner Novak:
Really?
Shensi Ding:
Yeah, I know.
Turner Novak:
What have been the bad things?
Shensi Ding:
Like COVID. We had to fundraise on Zoom. Itās weird to think about now, but at the time people werenāt used to fundraising on Zoom. So when we would do calls it was very, very awkward, and people werenāt quite used to having that kind of interaction. Now itās whatever, but back then everyone was like, āOh, this is weird.ā I think we were one of the first fundraises to happen over Zoom.
Turner Novak:
Really? So when was this?
Shensi Ding:
Like June 2020.
Turner Novak:
Okay. So what made it so awkward? Did VCās... Was it that you donāt know who should talk, you donāt know how to build a rapport with people?
Shensi Ding:
I think it was partially that. Zoom was still new. A lot of people we hadnāt met yet in person. So our first interaction was doing that. I think people felt like there was a different human component that was really missing over Zoom. And it is true, but now I think people are so used to it that it doesnāt really impact things.
Turner Novak:
Yeah. And then 2021 was probably just way too much money sloshing around.
Shensi Ding:
Yeah, exactly. Everyone had so much money. Everyone was so competitive. It was so aggressive. You had a ton of companies just randomly buying and changing ideas all the time. And then 2022, a lot of our customers were starting to die. Not yet, but starting to. And then 2023 they just all died. All the companies that were our initial customers, except for a few. Obviously Ramp, I think theyāre okay. Same with Drift.
Turner Novak:
Yeah, they seem to be doing pretty well.
Shensi Ding:
I think theyāre doing okay, yeah. And same with a few of the other logos that we ended up closing back then, like Brex, Drata, Vanta. But yeah, it was just a lot of ups and downs, and us having to figure out, okay, where is the industry going? Whatās happening right now?
In the beginning, when ChatGPT came out, we started having a few AI companies onboard, but none of them were really production ready, and none of them were getting a lot of traction or a lot of customers, just a lot of signups. So we didnāt have enough conviction on which direction we could go in to support these companies, because a lot of them just looked like SaaS companies with a little bit of AI on top.
Turner Novak:
Hmm.
Shensi Ding:
But then in 2024 we started realizing there was a really big opportunity, and the market might be a little bit too early, but the market needed integrations. AI agents are really useless without context. At the time people were really just thinking they could do it themselves, and they just needed a few core integrations in order to have an offering that was enough. But yeah, end of 2024, beginning of 2025 was really when all of it started hitting, which was really exciting for us.
Turner Novak:
So what was that like when youāre making the decision of, we have to go all in on something that didnāt exist when you first started the company?
Shensi Ding:
Yeah, it was definitely scary. I was really confident that it was the direction the industry was moving in, but the product roadmaps werenāt reflecting it. So I was kind of like, when is this gonna happen for us? When are we gonna start getting all the hyperscalers? When are all the large AI companies going to realize that they need some help with these integrations? And it did require a lot of networking. We had to continue meeting with these companies, really sharing why we were the right partner to work with. It was a lot.
We also had to actively invest in areas that werenāt quite ready yet or mature enough, and take resourcing away from some other industries that were definitely slower growth, or were probably not going to do super well over the next few years. That was very, very painful for us, because I want all of our customers to always love us. But especially when thereās a potential market opportunity and youāre the first to do it, you have to do it. Otherwise itās really irresponsible.
Turner Novak:
What was the hardest part about making that transition?
Shensi Ding:
I think it was just finding the resourcing, because we had a lot of customers that were very dependent on our product, and taking resources away from that product just felt like you were screwing your customers over. But we wanted to also launch our own AI products at the same time, while adjusting our core product to serve a lot of AI use cases. So figuring out how to make the resourcing work, how to move this engineer here for a few weeks and do that.
And then, let alone building an entirely new product. I was like, āI donāt even know how weāll find the resourcing for it.ā So what we ended up doing, which actually became a blessing, was we had myself, my co-founder, and one engineer focus on building the new product. That was it. And my co-founder and I would only be coding at night, so it would be like 5 to 11:00 PM. After all of our meetings were over, weād be coding. And then on weekends too. But because we were building zero to one, we learned so much about AI and what was possible.
When youāre building AI into an already existing product, what you feel like is possible is just significantly lower than doing zero to one.
Turner Novak:
And itās probably helpful that you were actually doing it yourself, because itās really easy to just see an article about AI and be like, āAll right, guys, letās do AI.ā
Shensi Ding:
100%. We learned so much during that time period, and it forced our team. Because we knew so much and we got fully AI-pilled, it made us really force our entire team to invest in it. So I think if we hadnāt done that, we probably wouldnāt have the kind of AI fluency that we have in our organization now.
Turner Novak:
Was it harder to build this product and scale it the first time, or the second time that you went back and redid things?
Shensi Ding:
Oh, my gosh. So the first time, we were in stealth for nine months before a single customer onboarded. And we were doing research for like six months before that. So it took a long time. Our second product, from day one to launch, was around six months. And the third product that we launched was three months. So AI is just getting better and better, and I also think itās a skill set for a company to learn how to launch new products. So weāre getting better at that too. Itās so much easier to launch new things now, and the ROI of doing so is really high. But there is opportunity cost. Thereās so many different ideas. Like, what do I end up doing?
Turner Novak:
So I know you have a lot of customers that use it, like Netflix, Uber, OpenAI, Perplexity, JP Morgan, BCG. Iām probably missing some. So what is Merge, for someone whoās five minutes in and theyāre like, āWhat are these people talking about?ā
Shensi Ding:
So we provide connective infrastructure for AI. We have connectors for products and agents, so if you want to connect to G Suite, Microsoft Suite, Linear, all the productivity tools, instead of having to build these integrations one by one, we have synced connectors. We also have live tool calls available via MCP, and we have a smart LLM router as well.
So now a lot of companies want to have model sovereignty, they want to integrate with all the different models. Every day thereās a hot new model coming out. Maintaining and connecting to all of them in-house is a lot of pain. So we end up having this layer that makes it really easy for people to build next-generation AI products.
Turner Novak:
Is it really that hard? Because you could just argue, āHey, Claude, just go do all this stuff for me.ā Why do you need to use something for integrations?
Shensi Ding:
Oh my God, we try doing that all the time with our own product. Itās just not good enough. Thereās always the edge cases.
Turner Novak:
So whatās so hard?
Shensi Ding:
For connectors and integrations, thereās just so many edge cases for what could go wrong. Thereās also a lot of context that is just not online and not available anywhere. Itās only available if you have industry knowledge, and if you have the data points from your customers for what could happen. And again, itās just not online anywhere, so an agent could try to do it on their own, but unless you have real-world experience, itās just hard to replicate.
So one thing that is also very human about connectors is that increasingly a lot of access to sandboxes, which you need, which is basically an instance where you can test against and start building your connector, you can only get access to it from a partnership. And a lot of these partnership conversations require meeting with someone, having some kind of agreement, them knowing who you are and making sure that youāre not sketchy. An agent canāt really do that.
Turner Novak:
So in a sense you are that partnerships layer for your customers. So itās, all right, we want to use this thing, we donāt have to go talk to a million people and spend all this time. We just use Merge, and it spins up really quick, and we can get going right away.
Shensi Ding:
Yes. And especially when thereās API changes, thereās new things that update, or one customer has a crazy instance thatās a little bit different from everyone else. Weāre able to account for all those edge cases too.
Turner Novak:
The interesting thing is that you guys actually came across this originally because you were building these integrations yourself, back when it was you and Gil when you first started. So I think you were getting dinner one night. What happened at dinner?
Shensi Ding:
Yeah, so we were getting dinner and just talking about work. At the time I was working with a cybersecurity company that needed a lot of ticketing integrations, and we actually didnāt have any. We were losing a lot of deals to our competitors purely based off of the number of integrations that we did not have and they did have.
So we ended up trying to avoid building these integrations as much as possible, having CSV upload, or sorry, download and upload, launching our own API, hiring contractors to build out these integrations. And it was just not successful. We ended up having to hire a full team of engineers purely to focus on these integrations.
But I have an engineering background, and I was like, āThis canāt be that hard. Come on. Let mama go after it. I will do this much, much better than you guys can.ā But my co-founder, Gil, was head of engineering at a startup in the recruiting space, so very different from cybersecurity. And he was like, āNo, this is much, much harder than you expect, because thereās so much data that we have to sync. We canāt even get started without getting partnership access, or having an established partnership and sandbox account from these different ATS providers. And thereās just a lot of edge cases that can go wrong that are not documented in the API docs.ā
So what ends up happening is customer support and customer success donāt know what the answers are. They always have to get engineering involved. So we build these integrations after a lot of effort and a lot of hard work and a lot of partnership time, but our engineers have to keep babysitting them, and we canāt really move on.
Turner Novak:
So then that comes into what we talked about a couple minutes ago, where you almost have this... Youāre almost like a partnerships team in a sense.
Shensi Ding:
Yeah, we are.
Turner Novak:
So itās almost like youāre a forward-deployed partnerships team maybe.
Shensi Ding:
Yeah. Itās actually one component of the product that was always a nice benefit, where, especially if you wanted to immediately get started, you end up having to go figure out, okay, how do I get a client ID, client secret? Where in the UI do I find this? Who do I reach out to? Originally with our unified API, it was a helpful part. But now, especially because AI is so good, itās increasingly a more and more important part.
And a lot of these SaaS platforms and sources of truth are increasingly becoming a little bit more protective of who has access to their data and how. So weāve been able to help facilitate a lot of that too.
Turner Novak:
And how long did it take to get all this going? Because you just said before it took nine months the first time you built it. Did you start coding on day one and it just takes forever to build, or was it a research project? How long did all this take?
Shensi Ding:
So we had six months of research before we started writing any code.
Turner Novak:
What was so hard about the research? Why did it take you six months?
Shensi Ding:
Because back then it was in San Francisco. This is like the olden days, but back then people didnāt meet on Zoom. So we had to meet with people in San Francisco for coffee. Each meeting had to be in person. I also needed to find the people to meet with, so I had to do cold outbound and beg people to meet with me. A lot of these were people weād never met with before, so we had to ask them questions, just to validate that it was a product theyād be willing to spend money on in the future.
But also, in retrospect, Iām like, wow, it was very generous of those people to spend this time with us, because we had nothing. Itās not even like we had a product that we could demo at the time.
Turner Novak:
Yeah.
Shensi Ding:
But yeah, it just took a lot of time, to make sure we werenāt going to be quitting jobs that we really liked for something that was just going to flop.
Turner Novak:
How did you convince people to talk to you? Because to your point, you were nobody, some random person, you probably didnāt even have a startup yet. You had this idea. So what was the general framework you used for reaching out to people and convincing them to talk to you? What kind of questions did you ask?
Shensi Ding:
Honestly, I probably shouldāve done a better job. I just did a shit ton of outbound. A ton of pipeline generation, being like, āPlease meet with me. Please meet with me. Please meet with me.ā So it wasnāt strategic, and very bad compared to what we end up doing now.
But yeah, it was just like, āOh, we have this idea. Your backgroundās really interesting. Weād love to get your feedback and advice on what weāre building. Weāll meet you anywhere.ā
Turner Novak:
Thatās probably a secret though, that you have to go to them. Like, āIāll go to your office. The meeting room beside your desk, we will meet there.ā The minimum possible effort from you to do this, weāll do that for you.
Shensi Ding:
100%. Thereās no way. These people are already doing us a favor, and then we have to make it harder for them? No way. I think thatās something people donāt really understand. This person is actively doing you a favor, so you have to make it easy for them.
Turner Novak:
Yeah.
Shensi Ding:
But yeah, if the meeting went well, or we really had chemistry with them, then weād ask for an introduction to someone else after. So that way we were able to expand our network.
Turner Novak:
Yeah. Itās always a good cheat code. Every time, I always ask, when people come on the podcast, āHey, this was great, hereās all the links, itās published. By the way, anyone else who youād recommend to come on?ā I feel like thatās a good way too. The guy Iām recording with right after this, Michael Tannenbaum.
Shensi Ding:
Oh, heās sick. I love him.
Turner Novak:
Yeah, heās great. Iād heard a couple people mention him over the years. Like, oh, this guyās the best non-founder CEO operator. And I had Sam Blond at Monaco on, and heās like, āOh, you gotta get this guy on.ā I was like, āHuh, Iāve heard of him multiple times.ā Heād be a good guest.
Shensi Ding:
You should ask him about... So we had this event at the Warriors game. Michael went to Columbia, and then Geoff Charles from Ramp also went to Columbia. And I am so stupid. We had an event, and I try to not have competitors at the same event that we organize, because Iāve been there before, and itās fucking weird. It just makes it so you canāt be as relaxed as when thereās a competitor there.
So we hosted this Warriors versus Knicks box suite, and I invited both of them. And I was like, āFuck, I canāt believe I messed this up.ā So I gave both of them a heads-up, and Michaelās like, āI donāt care,ā because Michael was about to go to Vigor anyway. And Geoff was so excited. Geoff was like, āOh, I really want to meet him.ā
Turner Novak:
This is when they were Ramp and Brex?
Shensi Ding:
Yeah. And Michael was so fucking scary in this meeting with Geoff. He was like, āWhere are you from? Okay, Iāve heard about you.ā
Turner Novak:
Iāve heard of Ramp before. And itās so weird.
Shensi Ding:
Because Geoff is also scary as fuck, right? But Michaelās the GOAT. He created the billboard, basically. He revived the billboard in San Francisco.
Turner Novak:
He revived the billboard.
Shensi Ding:
Sorry, not created. So itās his fault. He revived it.
Turner Novak:
It is his fault.
Shensi Ding:
Itās his fault, yeah. But heās just brilliant.
Turner Novak:
Anything I should ask him?
Shensi Ding:
You should ask him about that interaction.
Turner Novak:
The interaction, what happened when you went to the Warriors and Knicks game...
Shensi Ding:
Hosted by Merge.
Turner Novak:
Hosted by Merge, and you met the Ramp guys.
Shensi Ding:
He has a secret friend at Ramp, he told us.
Turner Novak:
A secret friend?
Shensi Ding:
At Ramp. But he wonāt tell us who it is.
Turner Novak:
Oh, interesting.
Shensi Ding:
Yeah, yeah, yeah.
Turner Novak:
Itās almost like a spy. You guys had a spy one time.
Shensi Ding:
Oh, yeah, we fucking did.
Turner Novak:
What happened with the spy?
Shensi Ding:
Okay, so Iām not 100% sure he was a spy, but I think he was. So there was this guy, he was interviewing at Merge. Sick engineer, strongest all around. Everyone was like, āThis guy is so good at coding. We have to hire him.ā
Turner Novak:
Yeah.
Shensi Ding:
I was like, āOh, my gosh, Iām so excited.ā So I look him up on LinkedIn. I was like, oh my God, great resume. He worked at Loom. And my friend Vinay is the founder of Loom. So I was like, okay, if heās good, I want to just back channel. Maybe Vinay can say something nice about us, and what we can do to try to recruit him. So I texted Vinay, I was like, āHey, do you know this guy? I saw he was at Loom for like five years.ā And Vinay was like, āIāve never seen this guy in my life.ā
Turner Novak:
And he was the CTO. He was the head of all the engineers.
Shensi Ding:
He was like, āThis guy never worked at Loom.ā I was like, āAre you fucking kidding me?ā And he was like, āYeah. Oh, Iāve gotten this back channel before from one other person, and this guy is just straight up lying.ā And I was like, āWho is he?ā And Vinayās like, āI donāt know.ā
And so then we were like, āShould we confront him?ā So our recruiter was like, āHey, weāre friends with the Loom founders. Are you... They said they donāt know you.ā And the guy was like, āOh, thatās weird. They should know me. Iāll circle back with another reference.ā And then his LinkedIn was just deleted.
Turner Novak:
Wow.
Shensi Ding:
And we donāt know where he is.
Turner Novak:
So youāre saying he was a foreign spy, or maybe from a different competitor kind of thing?
Shensi Ding:
No, I think it was a foreign spy.
Turner Novak:
Really? Wow. Because you see those stories where theyāre from a different nation state.
Shensi Ding:
And I heard theyāre sick engineers too.
Turner Novak:
Yeah. And they do the Zoom interviews, but they have all the answers. Theyāre getting fed the answers to the questions and stuff like that. Was it that kind of situation?
Shensi Ding:
No, I think he was actually a sick engineer. This was before AI was really good for cheating in coding interviews. He was actually a sick engineer. But weāre in person, so it wouldnāt be like he could try to join us by being remote with a fake face thing.
Turner Novak:
Thatās fair. But you guys are really big on in-person, in the office, right?
Shensi Ding:
We are, yeah.
Turner Novak:
Is that a reason, because of the spy? Like, well, we canāt do remote anymore.
Shensi Ding:
Well, itās a nice positive part, that itās a little bit less likely. But yeah, itās just really fun, and the team is a lot closer. It helps a lot with communication and being creative. We also have a lot of couples that have formed from Merge.
Turner Novak:
Really? How many?
Shensi Ding:
I think like seven or eight.
Turner Novak:
Holy shit. How many employees?
Shensi Ding:
Like 140.
Turner Novak:
So if you join Merge, thereās like a 6% chance youāll get married?
Shensi Ding:
I guess so.
Turner Novak:
Are these marriages or just dating?
Shensi Ding:
Dating, but very serious.
Turner Novak:
Okay.
Shensi Ding:
Also, a lot of times itās really high performers dating each other, because you donāt want to date a fucking loser. You want to date somebody thatās really good.
Turner Novak:
Yeah, thatās true. I feel like thatās one of the hard parts of modern dating, where youāve got the dating apps. Well, Iām married, so Iāve never really used the dating apps. But you get matched with someone and it all seems good, and then your interests and your ambition, everything is different.
Shensi Ding:
Mm-hmm.
Turner Novak:
Thatād be tough. But I guess you show up and youāre both at Merge. Youāre obviously great filtered.
Shensi Ding:
Yeah, filtered.
Turner Novak:
Yeah, great filtered. Interesting.
Shensi Ding:
And you know if theyāre good at their job.
Turner Novak:
Thatās true.
Shensi Ding:
And youāre also friends with the same people. Iām not encouraging workplace, you know? But it just happens, and it makes me happy that we had a real impact on them. Obviously in their career, but itās nice.
Turner Novak:
I mean, thatās potentially a really good pitch to work at Merge.
Shensi Ding:
You donāt have to say anything else. And my husband said the same thing, and Iām just like, āI donāt know.ā I feel like itās suable. Or itās scary. Actually one of my team members, okay, this is really bad, but she was just like, āYou need to hire some hot people for me, because Iām single, and right now the probability of someone finding their future husband at this company is pretty high. So you need to hire some hot people for me.ā
Turner Novak:
Just start hiring only hot people.
Shensi Ding:
Yeah, only hot people for real. Okay, Iāll try. Canāt be a filter, though.
Turner Novak:
Okay. So aside from being really hot, how do I get a job at Merge? What do you usually look for when youāre hiring people?
Shensi Ding:
People have to be really enthusiastic. They have to care a lot. They also have to be really smart, or want to do a good job. But cohesively, part of our culture is people care a lot, people are really nice, and people are enthusiastic. I think the worst part of going to a job, or talking to someone, is if theyāre just like, āUgh,ā and itās super draining. After youāre talking to them, youāre just like, āWow, Iām so tired,ā versus meeting with someone and youāre really energized and you feel excited.
Turner Novak:
So what do you usually look for when youāre interviewing? Are there certain patterns of people who act that way?
Shensi Ding:
Yeah, I think people who complain the entire interview, and where things happen to them versus them being in control of their life. Itās very palpable. Because also, if they complain about every single job theyāve ever had, theyāre going to complain about you. Youāre not going to be an exception. Theyāre just not going to be happy. So thatās something that I do look for a lot too.
I think itās totally fine to have a job that you didnāt like, but itās how you talk about it thatās really important. Sometimes Iāve heard people be like, āOh, all those people were just so bad, I hate all of them.ā Iām just like, āWhoa,ā you know?
Turner Novak:
Yeah. I mean, my very first job, I worked in a factory.
Shensi Ding:
Whoa.
Turner Novak:
It started at 6:00 AM. It was like the shittiest job ever, honestly. But it did teach me, you gotta wake up at 5:00 AM and go to work and stand all day.
Shensi Ding:
Now youāre just like, āThis is so easy compared to that.ā
Turner Novak:
Yeah. Now I just tweet all day and record podcasts all day.
Shensi Ding:
Men used to work in a factory, now they just move their fingers.
Turner Novak:
Now they just podcast.
Shensi Ding:
Yeah.
Turner Novak:
Thatās the downfall of society. Thatās why. I need to go back and work in the factories again.
Shensi Ding:
Yeah, yeah.
Turner Novak:
I mean, it actually did make me realize, wow, I need to take college seriously. I need to get a real job, because it was basically the summer freshman year after college. A friendās mom worked at this factory, and I needed a job for the summer, because I needed to make money. And I was like, āEh, canāt be too bad. Iāve never had a job before, but I know Iām gonna try hard and I do well. Sure, give me a shot.ā And then I was like, āWow, this sucks.ā It sucks working in a factory.
Shensi Ding:
Well, what was your hourly wage?
Turner Novak:
I think you got nine bucks an hour.
Shensi Ding:
Oh, thatās pretty good for a factory.
Turner Novak:
Yeah. And then you got overtime if you worked on the weekend. So I just always did overtime when they offered it. I canāt remember exactly. I think I was getting like $900 a week, or $1,000 a week.
Shensi Ding:
Thatās pretty good for a college freshman.
Turner Novak:
Yeah, it was pretty good.
Shensi Ding:
I didnāt work at a factory. I worked at a froyo store where I was getting verbally abused by this Korean lady.
Turner Novak:
A coworker or a customer?
Shensi Ding:
A coworker. But I was in high school. And it was like $8 an hour. Every time, I would love making money, I was like, āOh, my God, this is so cool.ā But standing for that long was something that really came into my memory forever. I was like, āWow, I love sitting.ā So now when I have a desk job, Iām just like, āThis is fucking awesome.ā I love sitting. I love not burning any calories, and just moving my fingers. Iām so lucky, and I like it.
Turner Novak:
Yeah.
Shensi Ding:
Yeah.
Turner Novak:
And then I kind of went on this arc where there was a job portal at school where you could apply for internships. And I was like, āOh, my God, I can make like 10 bucks an hour, 12 bucks an hour.ā So I just kept doing more internships. Every semester Iād get a new one where I made more money, to the point where I think my last semester I was making like 22 bucks an hour. Which, I was a 21, 22-year-old kid. And this was in West Michigan, in Grand Rapids. So thereās not a Google where you can make hundreds of dollars an hour as an intern. But it was pretty awesome. I was making more than my parents were making.
Shensi Ding:
Thatās cool. Thatās really exciting.
Turner Novak:
It was kind of crazy though, because itās like the most amount of money youāve ever had. I was getting paid more money every week than my mom had ever had in her bank account at any point in life. Itās crazy.
Well, on this vein of sitting in the office, air conditioning, a lot of founders hire an EA, or multiple. I actually know someone, they have four EAās. How do you feel about having an EA? You donāt have an EA.
Shensi Ding:
I donāt, yeah. Okay, so I used to be a chief of staff, and I used to manage an EA, so I have a lot of experience understanding what the workflows of an EA and a chief of staff are.
Turner Novak:
Yeah.
Shensi Ding:
I think if you have a big company and youāre very successful and very busy and traveling a lot, fuck yeah, get that EA. Get that chief of staff. If your company is 10 or 20 people, and this is different for VCās, by the way. VCās have a very busy schedule where theyāre traveling a lot and have a lot of meetings. But a lot of founders, youāre coding. You should be meeting with customers. Thereās Calendly and thereās Google Calendar. You donāt really need an EA.
Itās kind of a vanity thing. It also slows down the time for booking a meeting with a customer when youāre so small and you really need to just get that on your calendar immediately. And it just means youāre not really fast at your computer. You should be able to learn how to respond to emails fast. You should learn how to respond to your team, and book things on your calendar, and book travel. Itās a good skill set to know how to be fast at your computer.
Itās definitely a hot take, but if youāre really fast at your computer, you can go for a long time without an EA. Good EAās are a huge game changer, totally agree. But a lot of these people that Iām seeing with EAās and chiefs of staff and some kind of founders associate at the same time, Iām like, āWhat are you doing then?ā Are you really that important and that busy, where you need someone to do all these things, or are you just really fucking slow at your computer?
Turner Novak:
Yeah. Does there need to be a formality though? Letās say Iām a customer, I email you, and you say you need to be really fast. What if you just, lowercase, not a full sentence, just, āYes.ā Whatās the spectrum of, you need to make sure youāre being fast and quick to people, but then the formalities of making sure youāre presenting things? Is that part of the EA? It makes you seem more legit maybe?
Shensi Ding:
I donāt think so. Because also, Iāve interacted with many different EAās. If the EA is really bad, it makes you look bad.
Turner Novak:
Yeah, like they donāt respond for a day.
Shensi Ding:
Yeah, or if the EA is a total bitch, or if you send your availability and they just totally ignore it and send another thing, it creates friction. And especially if youāre in a situation where you really want to meet with this customer, or you really want to meet with this important partner, and you have not hired well, or your EA is not representing you well, that can really hurt the relationship.
So itās just better to have that in your control. Itās another mental load that you end up having to hold onto and think about continuously.
Turner Novak:
I think a lot about it too. I try to actively not schedule a lot of meetings and not schedule a lot of things. I feel like when you outsource it to someone else, thereās less sacredness around it. Versus for me, if Iām gonna spend time talking to you, and set up a meeting, I really want to talk to you. I donāt do 10 hours back-to-back meetings all day. I guess maybe Iām in a position where I donāt have to, but I donāt really want to do that.
Shensi Ding:
But you know what you can do with an EA though? You can have the EA blow someone off and make it seem like itās not you.
Turner Novak:
That is true. So I guess itās a shield.
Shensi Ding:
As a shield.
Turner Novak:
Interesting. Thatās why a lot of VCās hire associates. Because you can just be like, āOh man, Iām traveling, Iām just slammed. Hereās Shensi on my team, find a time to chat.ā
Shensi Ding:
Yeah. Oh, another thing about needing an EA. If you donāt know shortcuts on your laptop, like you donāt know how to go to the beginning of a sentence, or switch tabs without touching your mouse, you donāt deserve AI or an EA. At minimum you should know how to go between Slack channels with shortcuts.
Turner Novak:
Oh yeah, thatās fair. So actually, how do you... I donāt actually know. I donāt use Slack a ton, so I never had to learn the shortcuts.
Shensi Ding:
Itās just like using the shortcuts for Chrome and moving...
Turner Novak:
Like Control J or something?
Shensi Ding:
Yeah, exactly.
Turner Novak:
Control K. I mean, Superhuman is a VC. I know all the shortcuts.
Shensi Ding:
Yes. I think Superhuman has added so much value for the people who never wouldāve learned shortcuts, ever.
Turner Novak:
Oh, yeah.
Shensi Ding:
But yeah.
Turner Novak:
Well, itās like the whole... Because you did investment banking.
Shensi Ding:
I did. And not to brag, but Iām on the Training the Street leaderboard for formatting.
Turner Novak:
Are you really?
Shensi Ding:
For formatting.
Turner Novak:
Okay. So for people who donāt know, thereās like Olympic level competitions of how fast you can build an Excel model without using your mouse.
Shensi Ding:
Yeah.
Turner Novak:
Do you... So youāre basically saying, using AI is just like using shortcuts?
Shensi Ding:
I think so, yeah. Well, obviously to a different level, but I just think that shortcuts are a mentality. If youāre using shortcuts, your mentality is like, āI need to be really productive, and I need to be really fast at whatever Iām doing.ā So obviously then it extends to AI. I just expect people who are good at shortcuts are probably thinking, āOkay, how do I use AI?ā
Turner Novak:
Yeah, thatās fair. I think thatās smart. So then personally, how do you use AI inside of Merge? What kind of things have you done that maybe are super obvious, maybe theyāre not that obvious, itās unique to you guys? And personally, how do you leverage it to get more done?
Shensi Ding:
So I donāt have anything personal outside of Merge. That was a stupid question.
Turner Novak:
I mean, like, as a CEO.
Shensi Ding:
Iām joking, Iām joking.
Turner Novak:
I mean, thereās some people that are like, āOh, where should I travel?ā Organizing, or what Netflix should I watch, with AI.
Shensi Ding:
But thatās like level one. So for me, itās basically anything that I see at the company where Iām just like, why is this taking so long? Iāll just build it, or Iāll have someone else build it, or Iāll give a very detailed prompt on exactly what to do.
So one example. We have these dinners that we used to organize for our customers and partners. We would do one or two a quarter, because it was just such a pain in the ass to coordinate. It would require our whole team to go through LinkedIn, find who to invite. They would invite them, and they would have to update the sheet, and they would be like, āOh my God, the dinnerās all full.ā Everyone would freak out.
Turner Novak:
You gotta organize the dinner too. Communicate with restaurants about availability.
Shensi Ding:
Yes. And the seating chart. Thereās just so many things for the dinner that it was like, okay, we can only do one or two a quarter, because this is just too much time. And what would always happen, because we couldnāt fill the dinner, and I didnāt want to be embarrassed, was that I would end up spending all my time, through two to three weeks before, going through my LinkedIn being like, āWhoās in New York? Whoās in San Francisco? Who have we talked to recently?ā And inviting them manually.
I hated that process, because it was something that I always needed to get involved in, and we just couldnāt scale it. But the ROI of them was so high. People love these dinners, and weād have such a good time, and I love seeing our customers.
Turner Novak:
Yeah.
Shensi Ding:
So one day I was just like, you know what? Iām just gonna spend like two days. Iām gonna build a dinner bot. So now what this bot does, it doesnāt do everything. Like the taste of the restaurant, it canāt decide whatās a good restaurant. You know what I mean?
Turner Novak:
McDonaldās.
Shensi Ding:
Yeah, Chipotle.
Turner Novak:
Chipotle dinner.
Shensi Ding:
Or like a $30,000 dinner. No. Right? So it will surface, hereās the top restaurants, maybe do an outreach. TripleSeat, I think, doesnāt have an API either, so we canāt automate that part. So I do fill out a form, but whatever. Once the restaurant is picked, it does the calendar invites. It looks through Gong for every person weāve met with before. It goes through our Salesforce and Apollo. It goes through our customers and contacts, adds those, and whoās in the city, and it just sends the invites as me.
And then when they respond, it syncs the updates and responses. Before the dinner, it sends the reminders. The day of the dinner, it reminds my sales team to send notes to the people who are going, so they know what to talk about. Does the seating chart. So we went from having one or two dinners a quarter to now, I think we have three to five a month.
Turner Novak:
Cool.
Shensi Ding:
Yeah, and itās awesome. So itās been so helpful for us.
Turner Novak:
So it 10xād your dinner throughput.
Shensi Ding:
Yeah. So that was something where I just decided to build it out to help the team out, and itās had a pretty material impact. But me having that experience, building my own agents, and knowing what I can tell people to do. If someoneās like, āOh, this is too hard,ā Iām like, āItās just not. I know what you need to do here.ā Versus if I didnāt know, then Iād be like, āOkay, then we just do that,ā and end the conversation.
Turner Novak:
Yeah. I think itās a good framework where you can use AI to do anything that takes a long time. The thing I always do, I use AI a lot for the podcast, for all the prepping and all the post-production stuff. And anytime I change something, I always just say, āHey, I changed this. Just do it the right way next time. Figure out what I did different and just fix it.ā
And thatās a pretty... What do they call it now? Looping? They call it a product, if youāre using AI to do loops to build everything.
Shensi Ding:
Yeah.
Turner Novak:
Do you guys do that?
Shensi Ding:
Yeah, a lot.
Turner Novak:
And whatās your extent? Do you do any AI coding still?
Shensi Ding:
I only AI code.
Turner Novak:
You only AI code? Okay.
Shensi Ding:
Itās the point where, this is really bad, but itās the point where if Iām on an airplane and thereās no Wi-Fi, Iām like, āFuck, I donāt know what to do.ā
Turner Novak:
I did Wi-Fi on my flight here today.
Shensi Ding:
Oh, really? Whatād you do?
Turner Novak:
Listened to some music. Well, I played, I have a game on my phone called Slay. Itās kind of like Risk. Itās called Slay. Itās a game from like the early ā90s I played as a kid that I found on my phone.
Shensi Ding:
You guys sponsor the podcast if you love playing this game.
Turner Novak:
Itās a free game. And I donāt know if you can even pay for it.
Shensi Ding:
How long is the flight?
Turner Novak:
Itās just like an hour.
Shensi Ding:
Oh, from Michigan?
Turner Novak:
From Detroit, yeah. I think it was like maybe an hour and 15.
Shensi Ding:
Oh, really? I didnāt know it was that close.
Turner Novak:
It just depends. Sometimes itās closer to two, sometimes one and a little bit.
Shensi Ding:
Okay. Well, welcome.
Turner Novak:
Yeah, thanks for having me here.
Shensi Ding:
Well, wait, are you in Western Michigan?
Turner Novak:
Itās Detroit, East Michigan. I live in Ann Arbor.
Shensi Ding:
Oh, I thought you were living in Western Michigan. I grew up in West Michigan.
Turner Novak:
And then we moved to Ann Arbor right before COVID. It was really good timing. My in-laws are there, and then with kids, thatās why we live in Ann Arbor.
Shensi Ding:
Are you gonna make your kids go to University of Michigan in Ann Arbor?
Turner Novak:
Iām actually not convinced Iām gonna make my kids go to college.
Shensi Ding:
Yeah, I think itās gonna be like not, yeah.
Turner Novak:
I mean, everyone has different opinions on that, so I donāt know if they even want to go.
Shensi Ding:
Thatās very San Francisco of you.
Turner Novak:
Of being open to them not going? Well, I donāt know. My oldest daughterās more into art and dancing and singing, so she might want to go to an art or dance school. I donāt know. Sheās also really good at math, really likes reading. So she might be an English major. Maybe she gets a PhD. Right? So I donāt know. Iām of the view of, Iām not gonna force you to do something, because that sucks. I donāt want to force my kids to do things they donāt want to do.
One thing I want to talk to you about. When you think of marketing, as a brand, as a product, as a founder, as a company, whatās been your approach with Merge? Howās it evolved over time?
Shensi Ding:
So I think every function has gone through an AI transformation, and I think this year marketing went through its own really dramatic change. It became really clear that marketing was shifting a lot, especially the past few months, where thereās all these taste AI companies, all these design AI companies. And itās because I think coding, AI coding got so good, that the bottleneck for pretty much most companies was marketing. Because youād end up building so many features, and you wouldnāt be able to market them at the same pace, because there werenāt a lot of tools ready for that.
So weāve basically changed marketing to become an engineering function. All of our team members have to build agents for everything, like the dinner bot. Thatās one thing, right? Everything thatās really recurring and everything that kind of sucks, we should be using AI for. A lot of our ads, we generate them with AI. We have recurring posts or content that we generate based off of things that we produce in our product, and it just pulls straight from our GitHub and then generates the social post. And we obviously eyeball them, but it just releases them too.
All of the data analysis for demand gen of our different ad channels, all of that now, we use AI. We have a calendar, everything that goes into the calendar is automated too. So weāve gone through a pretty big transformation over the past two months, and itās been really exciting.
Turner Novak:
Is it working, or is it too early to know?
Shensi Ding:
I think so.
Turner Novak:
You think so?
Shensi Ding:
Itās hard to tell, because itās definitely lagging.
Turner Novak:
Interesting. So then what are you looking at as the leading indicators of an AI-native marketing machine?
Shensi Ding:
Signups, website visits, conversion, obviously early revenue, or POCās and opportunities that have been opened in Salesforce. So those have been the leading indicators. But itās still only been one or two months, so weāll see. Iāve been seeing a lot more activity, and activity hopefully ends up leading to more other things.
Turner Novak:
You mentioned customer signups. So I know in the early days of Merge you were more SMB focused, originally. So why did you start with SMB?
Shensi Ding:
We started with SMB because we were infrastructure, and no one bigger wanted to take a chance on us. It was very scary. We could have died any minute. They wouldāve been screwed. And if we had a security incident or any downtime, there were just so many risks with being an infrastructure company. So we really had to climb the logo ladder over time, and move to bigger and bigger companies as our product matured.
And then around three years ago, we made a very concerted effort to move enterprise. So that meant a complete change of who our team was, how our team was segmented, how we approached our marketing. We didnāt used to have decks for sales meetings. And then we ended up having decks, and we had a CRO who was really good at enterprise sales. That helped transform our company too.
So it was a pretty big transformation, and it took a few years. But last year really was the year when everything came in. We had all those logos that you mentioned coming to us, and it was just really exciting to see all the prep and the work that weād put in for the past few years finally coming in.
Turner Novak:
So then what was the evolution of the product? Because when you first started, it was literally accounting and HR integrations with accounting and HR SaaS tools, and this was back in 2020. So how did that evolve over time, and how did you decide what to do?
Shensi Ding:
So the goal was always to be cross-category. We just picked one or two categories in the beginning, as a wedge into the market and to introduce ourselves. But we picked them based off of what would overlap, what had big TAM opportunities, and especially as the market continued changing.
So fintech was getting really hot when we launched our accounting API. Then we started noticing a lot of sales AI and productivity tools starting to form, so we invested in ticketing and CRM integrations. But we always just kind of followed where the market was going. When we saw AI search was going to be the next frontier of what really needed integrations, we started focusing a lot on productivity and sources of truth, like knowledge base, chat, ticketing, file storage.
So itās always been TAM driven, customers. I think thatās the nice thing about B2B. Customers tell you what they want, and you just really have to listen to them.
Turner Novak:
Versus consumer, you have to be kind of philosophical, like, āI think this is what the consumers want,ā the tastes, the discourse.
Shensi Ding:
Itās hard.
Turner Novak:
Versus B2B, itās just like, āYeah, thereās a billion in pipeline that wants this. Build the thing.ā
Shensi Ding:
Yeah. Iāve always really looked up to good consumer founders, because I feel like itās much harder.
Turner Novak:
And then the evolution of the AI product. What was the very first AI feature that you initially launched for Merge?
Shensi Ding:
So there was a product that was for AI companies, and then we had our first AI products that were just very, very AI. But our unified API, we started noticing that for AI search, we needed to have file storage integrations be really good, like Google Drive, Dropbox, Box. Those platforms have so much data that people want to be able to search across. And also having knowledge base, like Confluence, Notion, chat, like Slack, Teams. A lot of companies need those integrations, especially as you want to make it very easy for people to do semantic search in your product, and no one was really able to do that. So we felt like it was a unique wedge for us, and we started going all in on that early last year.
And then at the same time we were like, āOkay, we need to be very aggressive about this move into becoming an AI company.ā So we were gonna launch our first AI product, which was Merge Agent Handler. These were tools for agents. Being able to make live tool calls, instead of our synced integrations, was a slight adjustment from our first product, but a completely different product. And it required us to really build something zero to one, do completely different marketing, learn how to sell this new product, and have different pricing. It was a lot of work.
Turner Novak:
Different pricing is scary. Because if Iām a public company CEO listening to this, I think of different revenue structure, and different distribution strategy, and sales. Basically you have to completely reinvent the company when you change the pricing. So how did you approach the new pricing model?
Shensi Ding:
Yeah. Good news is it wasnāt a full dramatic shift. The first product, we kept the pricing consistent. The second product, we layered on a different pricing model, and we had to train our sales team for how to talk about it. And then when we did our third product, it was a different pricing model again. So we didnāt shift the first two.
We will get to a point where weāll have to consolidate all three, because itās too confusing for our sales team. Weāll also probably have to repackage what the products look like, so itās a little bit easier to talk about. But itās been a really big shift.
And I think another thing that was really interesting for us to observe. So Gil, my co-founder, and I were startup people. We remember what it was like when it was very early, to get something from zero to one. And a lot of our team members had joined when Merge, that first product, was pretty established. We had customers, weāre thinking about scale. So they tried to apply a lot of the things that were working for that first product to our second and third product. So we really had to tell the team, āWe gotta move out of this. We canāt think about things that will scale. You guys are trying to protect us from the downside. Thatās probably not gonna happen. Because weāll be very lucky if this happens.ā
Turner Novak:
Someone uses it in the first place.
Shensi Ding:
Yeah, exactly. So that was something we really had to iterate to the team. Donāt be so focused on that. We need to just move really fast so that we can get those first few customers. And luckily they were very understanding, and we were able to do pretty well.
Turner Novak:
And then the second one is the agent handler product. So this basically changes which model the... am I saying this right?
Shensi Ding:
Oh, thatās Gateway. Thatās the third product.
Turner Novak:
So this is Gateway. Third product, yeah. Okay, third one. Well, I think thatās an interesting one too, where you basically route the AI direction flow through the cheapest model, or the most effective, highest ROI model.
Shensi Ding:
So the beauty, itās whatever you want.
Turner Novak:
Okay. So then why is that so important? Because shouldnāt you just have one vendor that you use, and you always get the same result? Why is it so important to be able to switch between different models?
Shensi Ding:
So before, because AI was still getting so much better at every single release, whenever there was a new model release, you were kind of like, āWow, the last version was totally shit compared to this one. Why would I ever use the last one?ā The new one actually works. So we had a couple years where that was really the case, and every single time there was a new model release, everyone would just default go to the latest model. It would just be a one-size-fits-all.
But now, thereās so many new model releases. Each frontier lab is releasing models like every month and a half or so, instead of every six months. Thereās also a lot of options. Thereās a lot of open source options. Thereās a lot of cheaper options within the same lab. And the costs are really starting to become more and more prominent.
Fable coming out made it really at the forefront of how expensive this could be, and how companies really need to start focusing on optimizing their AI spend. But I think weāre seeing, like Coinbase, DoorDash, theyāre very advanced companies thinking a lot about how to optimize their internal engineering spend, and how to make sure theyāre not overspending on certain workflows.
But for a lot of smaller companies, and also companies that are not that advanced, having that kind of functionality and focus internally is just really hard. And even if itās embedded in your product, the AI optimizations, it takes a lot of resources. It can be a full team fully focused on, how do I route, when do I make sure one plus one doesnāt go to Fable? How do I make sure hello doesnāt end up costing us three cents? Those are things that a lot of people need to focus on in-house. But itās what everyoneās building over and over again. So it was a natural extension of our existing products. It is a unified API. Itās one API to route to all the different model providers.
Turner Novak:
Do you know how many different model providers or model options there are right now?
Shensi Ding:
I think we have hundreds.
Turner Novak:
Oh, inside of Merge?
Shensi Ding:
Yeah, inside of Merge.
Turner Novak:
What about total, that you could use total?
Shensi Ding:
Oh, I donāt know. Thereās also complexities, because itās not even just... People also have in-house models that theyāre probably not disclosing. So itās infinite. People are starting to do post-training.
But weāre definitely moving to... Right now, this year was the year where everyone started realizing thereās a best model per task. Next year, people are going to start training their own models, but itāll be very good for one specific thing thatās unique to that company.
Turner Novak:
Where are the areas you think people are lighting the most money on fire right now with wrong model usage?
Shensi Ding:
I think itās in product, where theyāre not smart routing their customerās prompts. Where everything they send, because the customerās like, āI want to use the latest and greatest, Iām just gonna pick Fable 5.ā And then they just end up using so much money for the stupidest stuff. So theyāre not smart routing it when they need to be.
Turner Novak:
What do you think is the biggest risk to Merge today? Because in theory, every single software could throw out an MCP, and thereās no need to use Merge. In theory, right? Technically. Is that it?
Shensi Ding:
No, itās not. So I think MCP is still pretty new. Itās just a protocol, right? So a lot of companies are releasing their own MCP servers. Thereās public MCP servers. But for most companies, their public MCP servers were a side project or a marketing thing. It wasnāt meant to be a...
Turner Novak:
Oh, they didnāt actually work?
Shensi Ding:
A lot of them donāt.
Turner Novak:
Really? I guess I didnāt realize it was that insane.
Shensi Ding:
Weāve been doing some studies, and the success rate of most public MCP servers was like 60%. We build all the tools for MCP in-house, and then we test them really extensively, so the success rate for ours is like 99.9%. We really battle test them. And also, you can have a really dumb agent that ends up making it not successful, so we account for things like that.
Turner Novak:
Whatās the dumbest agent activity?
Shensi Ding:
Youāre basically like, āHereās the tool. Hereās the information for what to say. You need to send us the date in blue.ā And it just is like, āIām not sending you the date. Iām just gonna send it in red.ā
Turner Novak:
An agent will do that, even though itās just...
Shensi Ding:
The agent will just be fucking stupid. Weāve seen that with older agents. Some agents are just really bad at tool calling.
Turner Novak:
So instead of upgrading to Fable, theyāre still on GPT-1. Theyāre just so dumb they canāt even...
Shensi Ding:
Yeah, and youād be surprised. Thereās some later stage models that are just still really dumb. But itās not a risk for us, because thereās other things that we cover in our platform, like DLPās. We have a security layer that looks for making sure PII isnāt getting sent across. You can add custom rules too, to make sure, if you have certain policies for your company, or youāre in a regulated industry, you can block that data from getting passed to a third party.
And authentication, refresh tokens. Thereās governance, making sure agents only have access to certain connectors. A lot of those things are out of the scope of the MCP protocol, that we also include in our platform.
Turner Novak:
Why isnāt that part of the MCP protocol? Was it just, when Anthropic invented this, that was not something they were thinking about?
Shensi Ding:
I mean, MCPās already such a big... Why are you shitting on MCP?
Turner Novak:
Iām not shitting on it, but...
Shensi Ding:
Iām joking. Itās just out of scope. Because all those things, the governance layer, the observability, logging, DLP, those are things that not every single company needs. And also they might not think about. And itās just a huge build.
Turner Novak:
Well, and itās probably, to your point, some of that stuff is more enterprise grade that you need. I mean, the goal of MCP is just to get people using Anthropic products, and the most average, the median project, is a developer in their bedroom making a side project on AI. So it doesnāt have to be that robust.
Shensi Ding:
Yeah, it doesnāt. But what was interesting, when MCP first came out, we actually were like, āWow, this could become a big security issue.ā So we started off trying to build a security product for it. But when we were going out to market and talking to everyone, no one cared about the security, because there hadnāt been a vulnerability yet. There hadnāt been anything bad that had happened. So no one was like, āOkay, this is nice, but it doesnāt matter yet, because thereās not a big issue.ā
And what we really want are the connectors. So we were like, āOh, okay, I guess weāre gonna build the connectors then.ā So now the connectors are really good. But you canāt have a security incident if the connectors literally donāt work.
Turner Novak:
Thatās fair.
Shensi Ding:
So now that we have everything end to end, where our connectors are really good, and we have the security layer, and then we also have the observability layer, it makes it so itās really enterprise ready.
Turner Novak:
Yeah, thatās fair. So youāve pretty much gone all in on enterprise. You still have self-serve, thereās still a lot of smaller companies that use Merge, but what was the decision to go all in on enterprise? I think you mentioned you did it in like 2023-ish, 2022-ish.
Shensi Ding:
Yeah.
Turner Novak:
So why do it? Did you have to? Was it existential, like, weāve gotta do this?
Shensi Ding:
Thereās so many reasons. I think it became pretty clear that, with our first product, it was just a lot of effort for each customer. The amount of time that we spent with each customer, and making sure they were successful, and the amount of requests that they would end up having, was just a lot. And if the customer didnāt make it, it was just a lot of work for someone who just died. So even if they paid us a lot upfront and they died, it was very painful for our business.
So moving enterprise allowed us to have longer term relationships with companies that will probably be okay in 10 years. It allowed us to invest in them, them to invest in us, and it made our company a little bit more predictable. On the downside, we ended up becoming more of a whales business. So we ended up having certain really large logos that we close. And thereās more timing risk in our forecasts and our predictions, and if something slips, it hurts us more. Or if something comes in magically, itās really exciting for us.
So thereās definitely pros and cons. But I think the pros of having enterprise customers is that youāre building together, itās a longer term relationship, and you know that if thereās something wrong, theyāre still committed to working with you. Versus if itās very PLG, and youāre more of a commodity business where everyoneās just switching because youāre the cheapest provider, itās just hard to have customer loyalty. Itās hard to have a second chance too, if something goes wrong.
Turner Novak:
Itās interesting when you see market share of models, how it just fluctuates so much. Even when you look at maybe a year ago, all of a sudden it felt like out of nowhere, everyoneās using Claude. Everyoneās using Claude Code. And then we saw, I was talking to someone in the latest YC batch, and heās like, āEveryoneās using Codex.ā Itās just crazy that it just fluctuates so much. People just change.
Shensi Ding:
People are very unloyal right now.
Turner Novak:
Yeah.
Shensi Ding:
So thatās why sometimes, if your moat is the customer relationship, itās helpful.
Turner Novak:
Thatās a lot of things when an investorās evaluating how good a business is. Itās like, do you have sticky, hard to acquire customers? I mean, every business is just valued off the cash flows. Do you have customers that will pay you a lot of money that will not churn? Thatās essentially what a business is, a collection of customers paying you money. Really, at the end of the day, weāre not romanticizing this at all. But a lot of investors will think about that. Do you have big customers that will pay you a lot, and youāve acquired these really hard to get businesses to the point where they wonāt churn? So thatās valuable.
And so you think more startups should go after enterprise? But itās hard, right?
Shensi Ding:
Itās hard. And also the revenue growth is going to be slower than what youāre seeing with Lovable. Lovable is PLG, so itās very easy to have a ton of companies just randomly sign up, and Lovable doesnāt have to do as much, and it just scales really quickly. With enterprise customers, each customer, you have to have developed a relationship. You have to build together.
Turner Novak:
Yeah, the dinners.
Shensi Ding:
Yeah, the security review, legal. Itās just bigger.
Turner Novak:
Mm.
Shensi Ding:
So it depends. But then also, Iāve seen some companies, they just get some really big logos immediately. That helps them a lot too. But yeah, if itās really hard, itās hard for someone to copy.
Turner Novak:
Yeah, thatās true. Iām definitely a proponent of, you should just do the harder things, because to your point, itās harder to copy. The whole meme of, someone will vibe code and just steal a thing. Itās kind of true. You kind of want to make sure youāre not in a position where people can just vibe code your product, and thereās a million competitors that pop up. So itās like, what can you do thatās harder? That makes it less likely and less possible.
Shensi Ding:
Yeah.
Turner Novak:
I feel like part of it too is thereās always so many other things, whether itās the customer relationships, the partnerships. I think itās actually not that likely that someone just vibe codes and copies the product.
Shensi Ding:
No, itās such a pain in the ass to have to maintain something in-house. I would never vibe code Salesforce in-house. Like, ever.
Turner Novak:
It was so interesting, that discourse was probably... Was it 2024? I think so. That was the big thing, that everyoneās just gonna vibe code, and all the softwareās gonna just go away.
Shensi Ding:
Well, also, who has the time to build and maintain this application in-house, in addition to their core product?
Turner Novak:
Yeah. Well, you just said you make the events thing, your events tool.
Shensi Ding:
Thatās not a product.
Turner Novak:
So this is not, like, youāre selling it to people.
Shensi Ding:
No. Itās just using it internally.
Turner Novak:
What if I make a marketing event field work tool that then I sell to you? Would you buy it from me?
Shensi Ding:
Yeah.
Turner Novak:
So then how do you decide...
Shensi Ding:
Depends on the price.
Turner Novak:
Okay. So was it, you built this internally because it probably didnāt exist?
Shensi Ding:
It didnāt exist.
Turner Novak:
And so then, if Iām thinking about...
Shensi Ding:
But at the time, if it existed already, Iād buy it. I donāt want to build it.
Turner Novak:
How much would you pay for something like that, though?
Shensi Ding:
Probably not a lot. Because itās not... But, like, Salesforce, we pay a lot, because itās way more, you know? So it depends on the ROI.
Turner Novak:
But then, with Salesforce, I feel like thereās people that say they vibe coded, replaced the Salesforce. Why do you still use Salesforce, and why have you not vibe coded it?
Shensi Ding:
Because itās a source of truth that connects to everything. Everything connects to it. Weāve maintained it for years. We have a lot of automations built on top. Thereās just so much that we do in Salesforce that we canāt replicate if we built it in-house.
Turner Novak:
Yeah, itās probably the more integrations that you have, and the more different things you connect with, the more defensible you are.
Shensi Ding:
Thatās why I would tell a lot of these companies that are a source of truth, the more companies that are built on top of you... Linear does a really great job of this. Theyāve created an ecosystem around their product. That makes you really, really sticky. Ashbyās also doing a really good job of this too. So many companies integrate with them.
And the more legacy companies are like, āOh, well, these newer companies donāt understand, thatās why weāre closed off, and we have secret sauce, and people are really dependent on us.ā But thatās fine for now. That degrades over time.
Turner Novak:
Youāve kind of taken the mentality of all the different partnerships, all the different integrations. Youāre kind of that layer personified in a way.
Shensi Ding:
We are. But the thing is, whatever someone wants, weāll follow. If a company is like, āWeāre open right now, but weāre gonna be closed off next year,ā weāll work with them on that. We are always going to follow what the partner wants, and make sure that we facilitate that. We really want to be that layer in the middle, and help make companies integrate with them more easily.
Turner Novak:
And so one thing you mentioned is you still do a lot of outbound, as the founder, manually. So why do that? You could vibe code some, or use some AI automation stuff.
Shensi Ding:
So itās a different type of outbound thatās not necessarily automated.
Turner Novak:
Okay. So whatās an example?
Shensi Ding:
So for recruiting, I do a lot of sourcing. For engineers, I use this recruiting tool. It is an AI recruiting tool. I use AI for that. Iāll just basically be like, āOkay, this is what Iām looking for.ā Iāll have an agent that automatically sends outreach for candidates based off of the criteria that Iāve done.
Turner Novak:
Oh, really? So you donāt even know that itās sending it.
Shensi Ding:
Yeah.
Turner Novak:
Oh, thatās interesting.
Shensi Ding:
Well, because itās hyper-specific. But this toolās really great. It allows me to also just select all and send. So I can do that too. But for dinner outreach, itās kind of like having a good party. You want to have a good party, you gotta know the vibes.
Turner Novak:
Yeah.
Shensi Ding:
The vibes. There might be, like, Shensi would not be a good guest at this party, so I donāt want to invite her.
Turner Novak:
Or I do.
Shensi Ding:
Yeah.
Turner Novak:
Like, I do.
Shensi Ding:
I know this personās fucking hilarious. Theyāre gonna be in the middle. Or this person sucks, Iām gonna put them like...
Turner Novak:
Itās like, oh, itās one of our biggest customers. I havenāt talked to him in a while. He should probably be here.
Shensi Ding:
Yeah, exactly. So thereās just certain things that are hard to replicate with AI, or even having another person. Thatās why itās hard if we hire someone to even do some of these things.
And then another thing, the outbound is more a lot of networking. I hate to say it, because I think when we were smaller, networking was not helpful. But now that weāre bigger, thereās certain networking events that have been super useful for us, and the ROI has created partnerships.
Turner Novak:
Yep.
Shensi Ding:
Networking has actually, unfortunately, been very helpful.
Turner Novak:
Networking in what sense?
Shensi Ding:
Like meeting other founders.
Turner Novak:
Oh, okay. And youāre just manually, you come across someone and you just DM them, like, āHey, want to chat?ā
Shensi Ding:
No, not quite. Itās like youāre at a conference, and you meet them, or youāre at some kind of founder event and you meet them. And you just have to say psycho shit so they remember you.
Turner Novak:
Wait, what?
Shensi Ding:
Yeah, you just have to say... Because the thing is, you just have to be memorable.
Turner Novak:
Yeah. I like that advice, that you have to say something that is memorable, and then itās easier to follow up. Because the follow-upās so important. When you read those networking 101 guides, itās always like, āMake sure you follow up after the meeting.ā Well, thereās been times Iāve gotten a follow-up from someone, and I forget meeting them.
Shensi Ding:
Yeah, no, because they suck. So you just have to be really memorable. I met the Runway ML guy on the street, when we were going from one founder event to another founder event. And I was like, āOkay, I need to just not be a bitch, and I need to just say hi to people.ā And so then I talked to him. But I talked to him, and I was like, āOh, Iām never gonna be able to sell to him. I donāt even know why I bothered.ā But I was really nice to him, and I made a few jokes. And now thereās a deal, and he remembers me, because I was the only person to talk to him. Everyone else was being so unfriendly, and I was the only person that talked to him on that street.
Turner Novak:
This was when you were walking, or at an event?
Shensi Ding:
We were literally just walking. And I just talked to him.
Turner Novak:
Interesting. I actually remember there was one event in Michigan, University of Michigan, that I went to. And one of the... She was a professor, also has a company, that I talked to in between some events. And I just saw her randomly walking through downtown, and I was like, āOh, I know that person.ā I just said hi to her. And again, I remembered her because we talked at this thing.
Shensi Ding:
Yeah, you have to be memorable. So this is like four years ago. We were at Web Summit, and I was like a baby founder then. I wasnāt super experienced. But at the time, Carta was like the hottest company in the Valley. And I remember being like, āOh, Iām so bummed,ā because Carta is using our competitor, and I really want them as a customer. It would just be a dream logo. Weāve used Carta for a long time. Iāve always looked up to Henry Ward. I really wanted Carta as a customer.
And so then I look up and I see him, and Iām like, āOh my God, I have to talk to him. Iām never gonna see him again if I donāt talk to him.ā
Turner Novak:
Yeah.
Shensi Ding:
And Gil, my co-founder, heās like, āWhat are you doing?ā I was like, āI need to go talk to him.ā I was shaking. I was like, āOh my God, I need to just fucking man up, and I need to go talk to him.ā So I go up to him and I introduce myself. Itās so creepy, because I read, knew everything about him by the time we...
Turner Novak:
Itās like one of the things where you introduce, and you read their biography to them. You recite every single piece of information that you know about that person.
Shensi Ding:
100%, yeah.
Turner Novak:
So how do you avoid... What do you say when you first come up to someone like that?
Shensi Ding:
I just had to be, Iām not usually, but I had to be pretty charming. I had to be like, āHi,ā and really think of things to say, and comment on the event. I donāt remember exactly what it was, but he remembered me. And then I got a text, or an email right after, that was like, āHey, Iām doing a dinner tonight. Do you want to come?ā
Turner Novak:
You got that from him?
Shensi Ding:
Yes, from him. And I was like, āFuck.ā
Turner Novak:
Wow. Thatās huge.
Shensi Ding:
It was crazy. So that night we had dinner with Henry Ward. And it was such a cool thing, because I had looked up to him for so many years. Heās just such a great founder. And I got blackout.
Turner Novak:
So you donāt even remember the dinner?
Shensi Ding:
I donāt remember most of it, but I had a great time. We had a lot of jokes. We had a good time. And then they became a customer.
Turner Novak:
Wow.
So you have this thing called the embarrassment framework?
Shensi Ding:
So I always tell my team, whatās more embarrassing, posting something on LinkedIn about Merge, or the company dying? And so thereās a clear answer.
Turner Novak:
Yeah, I mean, that would be terrible to actually have to post on LinkedIn.
Shensi Ding:
Yeah.
Turner Novak:
But so thatās the framework, is just...
Shensi Ding:
Itās the framework. And itās helpful for the team, because I think a lot of them are kind of like, āOh, itās so embarrassing.ā Iām like, āYeah, but it helps us so much.ā
Turner Novak:
And imagine if you worked at that company and it died, and itās just because you didnāt post on LinkedIn. Come on. Just make the post.
Shensi Ding:
I, yeah. And obviously, a LinkedIn post isnāt gonna save the company. Thereās so many different things. But it certainly helps. And itās free. And it definitely propagates more to the founders. It is the founderās responsibility to do the thing thatās embarrassing. You need to send that email. You need to go up to that person. You need to shoot your shot.
And thereās so many times Iāve had to send an email that is so embarrassing and so horrible, and then Iām so glad I did. So Dropbox is one of our customers. We went through a deal process with them, and they were like, āNo, weāre not gonna move forward.ā I was like, āOh, so sad.ā And then I stayed on it. I was like, āHey, howās it going? How is building in-house? Any updates?ā And it was so embarrassing. Obviously they didnāt want to fucking work with us, right? But then, I think like the sixth time I emailed, they were like, āLetās chat.ā And then now theyāre a customer.
Turner Novak:
Interesting. So how do you get past that embarrassment? Any advice for doing these uncomfortable things? I get it, the companyās gonna die, but you still have to do the thing.
Shensi Ding:
Itās, let your inner autist come out.
Turner Novak:
So you almost take down your guard of social acceptability, just being like, āHowās it going building this thing in-house?ā
Shensi Ding:
Yeah, pretty much.
Turner Novak:
Interesting. Okay. How do you channel your inner autist?
Shensi Ding:
I think you just have to give no fucks, but itās hard to get to that point. But you can be memorable over text too. I show my friends and my team members some of the messages that I send to our prospects, where Iām like, āHey, Tim, why the fuck are you ignoring me?ā
Turner Novak:
You actually will say that?
Shensi Ding:
Yeah. Tim from Apollo has gotten so many texts. All blue, but I love him.
Turner Novak:
And then that works?
Shensi Ding:
It works, because we love each other.
Turner Novak:
Okay.
Shensi Ding:
Or maybe I love him. I donāt know. But when we hang out, itās a good time. He remembers me, you know?
Turner Novak:
Yeah.
Shensi Ding:
Also one of my friends, sheās a really hot girl, but sheās definitely autistic. She went to Stanford, but she looks normal. So it comes off like sheās a bitch, because she looks normal, right? So she tells me that she knows sheās a little autistic, and she just leans into it when she needs to for business purposes. And that actually really inspired me, because Iām definitely not there. Iām not like that.
Turner Novak:
I think everyone is somewhere on the spectrum though. From zero to 100, most people are not zero.
Shensi Ding:
Yeah, but Iām not her level. But if I need to, Iāll just try to do it.
Turner Novak:
Itās kind of like... You know how the first time you drink alcohol, you drink beer, and you get drunk, and youāre like, āOh, this is what itās like being drunkā? Itās almost like it brings that inhibition level down, and you can tap into it later. I wonder if itās the same thing, where you get an example of it from someone else, and you learn, āOkay, this is how I do it.ā I donāt know.
Shensi Ding:
Yeah. Sheās really good. Sheās very direct. So yeah, will this get me canceled? Will people be like the anti-autism... I donāt know.
Turner Novak:
I mean, if you didnāt get canceled before, in the first hour and a half, this will definitely do it.
Shensi Ding:
No.
Turner Novak:
No, I donāt know. I feel like itās almost un-cancellable, talking about being autistic. Itās acceptable, right?
Shensi Ding:
Yeah. I think if it stays in the tech sphere. Once it gets out, then itās kind of like...
Turner Novak:
Yeah, thatās fair.
Shensi Ding:
Also, you never know when things will... At least right now itās uncanceled, but in like 20 years someoneās gonna be like, āShe said the A word.ā
Turner Novak:
Yeah. Thatās true. Itās so funny when theyāll dig up old videos of Trump, and heās just like, āRepublicans are idiots. If I were to run for president, Iād be a Republican, because theyāre just really dumb.ā And he probably wouldnāt say that today. But he said it in the ā80s. So I donāt know. Anyways, I feel like being canceled is kind of overrated.
Do you have a favorite CEO, founder, or business, maybe throughout history, at any time, that you get a lot of inspiration from, or that youāve always looked up to?
Shensi Ding:
Yeah. I love Frank Slootman.
Turner Novak:
Frank Slootman. Okay.
Shensi Ding:
I know thatās kind of a basic bitch response, but...
Turner Novak:
So why?
Shensi Ding:
Heās alpha as fuck.
Turner Novak:
Okay. So whatās been the biggest thing youāve learned from Frank Slootman?
Shensi Ding:
Well, Iāve never met him, so itās a little harder.
Turner Novak:
Itās that parasocial relationship.
Shensi Ding:
It is, yeah. I mean, I think heās really good at saying the thing thatās hard to say. Heās very, very direct. And heās not afraid to push people, which I think a lot of people forget. And heās okay with being disliked.
Turner Novak:
Is that a hard part about being a founder, being disliked? You say something and somebodyās like, āI quit,ā or āI donāt want to work for...ā
Shensi Ding:
Yeah, 100%. All the time.
Turner Novak:
Really?
Shensi Ding:
Like, you say something thatās true and correct, and for the good of them. Because, maybe this is also because, I donāt know, maybe because Iām a woman, but I really care about people. And so sometimes Iāll share feedback because I want them to really do well. And theyāll take it personally, in a way where they think Iām trying to hurt them, or itās just malicious. But itās not. Itās for the purpose of them and their personal growth. But yeah, it makes me get disliked.
Turner Novak:
How have you learned to give harder feedback as a founder? Because Iām assuming you have to give it a lot.
Shensi Ding:
Yeah, I just am direct. You have to just say it. When youāre not saying it, it ends up hurting the company and the team a lot more over time.
Turner Novak:
Yeah. I feel like the most honest, nowadays, is from my wife, telling me something, like, āYou need to change this behavior,ā or āYou do this thing.ā Iām pretty avoidant, I have an avoidant personality. And sheās always like, āYou need to stop avoiding this.ā Iām like, āYouāre right. I do.ā So I think itās really... Itās hard. Because you donāt want to hear it. You hear this negative thing, this thing that you messed up, or you need to be doing something better.
Shensi Ding:
Yeah. And if the person knows, and they clearly donāt want to talk about it, then Iāll just be like, āHey, I think we both know. Letās not talk about it, but it canāt happen again.ā
Turner Novak:
Is that a good way to approach it? Like, we both know?
Shensi Ding:
If they know, then yeah, itās not really worth it. Because theyāre gonna be harder on themselves than I could ever be.
Turner Novak:
Yeah. And thatās the people you want. You want people who realize, āI messed up. I really want to be better at this. I want to improve at it.ā
Shensi Ding:
Exactly.
Turner Novak:
Yeah.
Shensi Ding:
So yeah, itās hard. And especially because the companyās bigger. I think the thing that was hard for me to adjust for was that the feedback got scarier.
Turner Novak:
Really?
Shensi Ding:
The people that I worked with much earlier, when the company was smaller, we were just so close, and there was so much trust from two-way feedback, that it wasnāt really... I didnāt even have to think about giving feedback, and they didnāt have to think about giving me feedback, because it was just how we worked together. It was like, āOh, this part wasnāt super good,ā or āHey, this was a little delayed, you canāt do that next time,ā or āCan you do this better?ā It would just be two-way.
But then, almost like once we hit 75 or 100 people, all of a sudden it was scary. People would just be really offended by things that I said. They didnāt know me as well, so they didnāt really understand. I met Howard Lerman from Yext, and it was around the time where I was going through a really hard founder transition, where I just wasnāt used to how removed people felt from me. And he was like, āYeah, itās really hard at this point, because they donāt understand you. In the beginning, they really want to understand you.ā
Turner Novak:
So when did that transition happen?
Shensi Ding:
It was like 75 to 100 people.
Turner Novak:
Okay. And the thing you go through is that youāre just not spending as much time with every single person, so no one knows you as well. So what do you do? What did you change after that?
Shensi Ding:
I realized I needed to go through my managers a lot more, and be a lot more reliant on my executive team. And I had to spread out the responsibility of sharing feedback and ownership with my executive team more. That helped a lot. So we have a really strong manager bench. Our managers are very... I spend a lot of time with them. But I donāt give feedback to ICās anymore. I give feedback to their managers, and their managers already know usually, and will be giving feedback to their team members. Same with our exec team. Our exec team will usually end up talking directly with the managers and then the ICās too.
Turner Novak:
So then what is your job as a CEO, when itās my first day as an engineer, Iām meeting you? Iām probably pretty nervous, honestly.
Shensi Ding:
Iām like, oh, man, Iām meeting the CEO.
Turner Novak:
Of this hot company.
Shensi Ding:
Oh, yeah, something really bad happened. So thereās a guy who started, and heās so nervous, and he was introducing himself to me. And then on his second day, he actually called me another Asian girlās name.
Turner Novak:
Wow.
Shensi Ding:
I know. For me, Iām like, āI donāt care.ā But I know he will never forget that.
Turner Novak:
Yeah.
Shensi Ding:
I know he will never forget it. Heāll always think about it. Iām like, āThatās so fucked.ā I feel so bad for him. But heās so nice.
Turner Novak:
Yeah.
Shensi Ding:
So, yeah.
Turner Novak:
So then is your job as the CEO mostly making people feel excited about what youāre doing at the company? Or is it motivating people, or showing the vision? Itās less about, hey, specific tactical things. Itās more big picture. Is that more how you think about it?
Shensi Ding:
Yeah, so itās things that I can uniquely do. So itās recruiting, showing the direction of where weāre going, making sure morale, making sure the team is excited, and knows what weāre doing. And a lot of those things can be aligned and get close together too. And then thereās certain strategic big deals, and sales things, that Gil and I need to be involved in. And then product, obviously.
Turner Novak:
Whatās your level of involvement in sales? It sounds like youāre still texting and emailing customers all the time, but whatās the level of involvement? Have you gone through any cycles of more or less?
Shensi Ding:
Oh my gosh. Yeah. Itās always in two to three-month stints, where something will be my 100% focus. So earlier this year it was 100% recruiting, with some product. And then at some point it was 90% coding. We just really needed to get a product out, and I wanted to know what was going on, so I was coding so I could help contribute. And then right now itās mostly a lot of marketing.
Turner Novak:
So then how do you know when to lean in and out of different things?
Shensi Ding:
Honestly, I donāt... This is so bad, but itās just gut.
Turner Novak:
So itās just your gut is like, āWe need to get the word out, I need to do marketing,ā or āWeāre shipping this thing, I need to work on the product this month.ā
Shensi Ding:
Yeah. I go to a spot where itās extremely high ROI, only I can do it, and if I donāt help out with this, then itās gonna really impact the company negatively.
Turner Novak:
What was the hardest thing to give up doing, on a daily basis? Because Iām assuming, when it was you and Gil in the room during COVID, you did everything. Whatās been the hardest part to give up?
Shensi Ding:
I love coding. Yeah, I really love coding.
Turner Novak:
And so you obviously donāt do that all the time anymore.
Shensi Ding:
I donāt get to do it. Iāll do it for fun, or to help the team out. Also, because I work a lot, and Iām kind of a loser, so I have a lot of time to focus on Merge. So with the time that I have...
Turner Novak:
Your free time is coding.
Shensi Ding:
Yeah, for fun. Like, when Iām watching Real Housewives, Iāll code.
Turner Novak:
What is it, Girl Housewives?
Shensi Ding:
Real Housewives.
Turner Novak:
Oh, Real Housewives.
Shensi Ding:
Yeah.
Turner Novak:
Whatās so good about Real Housewives?
Shensi Ding:
Oh, my God. You can go in and out and multitask in a way that you canāt do with anything else. Itās not just Real Housewives, itās all Bravo shows, like Summer House, Vanderpump, Southern Charm, all the shows. I donāt really need to know whatās going on, because it doesnāt matter.
Turner Novak:
Because just some fight is happening and youāre like, āOh, this is entertaining.ā
Shensi Ding:
Exactly. So I can do it while coding.
Turner Novak:
Okay.
Shensi Ding:
And it allows me to spend... Versus if itās silence and Iām coding, itās not the same vibe. Versus I feel like Iām enjoying it. Iām watching TV while Iām doing something very productive.
Turner Novak:
Yeah. Have you ever watched Traders? Have you seen that one?
Shensi Ding:
Yeah, I fucking love Traders.
Turner Novak:
I love that show.
Shensi Ding:
I love it. Itās so good. We should do a VC Traders.
Turner Novak:
I thought about it. We should just film it like what Founders Fund did.
Shensi Ding:
Yeah.
Turner Novak:
Iāve thought about it. So the issue I thought about is, itās a lot of work.
Shensi Ding:
Itās a lot of work.
Turner Novak:
Like, this is helpful, because I learn a lot through this, and itās just kind of fun to hang out. You get to meet new people. I make money from it. It helps the portfolio. You have founders that you either want to meet, or want to help. Thereās just this whole thing thatās pretty helpful. Versus if I spent a month doing a Traders show, it would be really fun. And Iām sure people would talk about it. I donāt know.
One other one Iāve thought about is making parody music.
Shensi Ding:
Oh, parody music. Thatās pretty good.
Turner Novak:
So, what was this guy? You know, by Drake, itās called Godās Plan.
Shensi Ding:
Mm-hmm.
Turner Novak:
Some guy made a song thatās like Claudeās Plan or something like that?
Shensi Ding:
Yeah, I saw that.
Turner Novak:
So good. That was really good. One that I thought about is, whoās the guy, Chief Keef, heās got a song like I Hate Being Sober. He could probably make a song like I Hate Being Bootstrapped, or something like that. And all the words are talking about AI, like token maxing. You could probably just take popular songs and change the words, the same pattern, same flow, and itās just about something startup, tech, VC related. And people would identify with it and get it, and itās just unique and different.
Because I think one of the most important things with marketing generally is, you just donāt do the same thing as everyone else. You figure out what is a new thing to do or try. So we kind of go through cycles with watch videos, everyone starting a podcast, etc.
Shensi Ding:
Also ABG marketing.
Turner Novak:
ABG marketing. Have you guys gone through any ABG marketing waves at Merge? Followed any?
Shensi Ding:
Unfortunately, no.
Turner Novak:
Unfortunately?
Shensi Ding:
Unfortunately, no.
Turner Novak:
Does it work? Like, should you?
Shensi Ding:
It probably does.
Turner Novak:
Hmm.
Shensi Ding:
All those posts have gotten hundreds of thousands of views. I donāt have a single post that has hundreds of thousands of views.
Turner Novak:
Thatās not true. You have a couple launch videos.
Shensi Ding:
Okay, yeah. Thatās true. So maybe I accidentally was my own ABG.
Turner Novak:
Yeah.
Shensi Ding:
Yeah, I donāt know. I mean, eyeballs works.
Turner Novak:
Yeah. But youāre also literally finding love at Merge. The employees are literally getting married. So that is a good... I feel like thatās success.
Shensi Ding:
Yeah, weāre creating ARR and love.
Turner Novak:
Well, this has been a lot of fun. Thanks for coming on the show. Thanks for doing it.
Shensi Ding:
Thanks for having me!
Find transcripts of all other episodes here.


The night-shift rebuild story and the way they forced AI fluency into the whole company is interesting. I liked the honesty about tearing things up and the Embarrassment Framework make it feel unusually useful.