Audio & Transcript

Welcome to the Superconnector podcast. I am Matt Joseph, your host. We've got another round of startups to look through from YC's winter 2024 batch. Let's jump into it.

First up is Yenmo, instant loans against investments. All right, let's see what we got.

Yenmo offers a better way to access cash for India's 60 million investors who either resort to personal loans at over 18 percent or have to liquidate investments when in need of money. With Yenmo, they can pledge their Investments digitally and instantly get a loan at a flat rate of 10.5% interest.

Firstly, Yenmo, obviously branding against Venmo, not a bad idea. I don't know if the Indian market knows Venmo, maybe they do, but I think it's smart to use a money transfer brand for a financial services product that is, in some ways, derivative of what the big brand Venmo actually does. Let's learn about the team. Ashutosh started digging deep into the problem while he was working at one of the largest fintechs in India. He discovered over 30 percent of users who were taking a personal loan had active investments, but they just weren't aware of an option to take a loan against their investments.

He quickly called the best coder he knew, Aryan, who was running a capital management company, to join him to build something big. And we've got a nice picture of them outside the YC office.

So the first thing I want to understand is why exactly these investors need liquidity, why they need to take out personal loans. I mean, are these the kind of borrowers who you actually want to be loaning money to?

I have to imagine that they have some sort of underwriting capacity to be able to say, Yes, these are good investors. They're going to be good for the money, but it's surprising to me that there are so many of these people.

There's a little bit of smoke and mirrors in the numbers they put out here because they say 60 million investors. But not all of those 60 million investors are resorting to personal loans and not all of them are going to have liquidity problems.

So I'm curious as to what the actual market is. I mean, they say over 30 percent of personal loan borrowers have active investments. I get that, but maybe the reason that they're taking out these personal loans is because they trust the people that are taking the loans from more than they would trust a startup.

So it's hard to parse that. But I will say this, that there's a lot of very successful startup founders in the U S who use their equity to get loans from banks. And then they turn the loans into things like real estate and they start actually making money off of their startup shares. So this isn't a space where there isn't some sort of precedent, but I'd be curious to know why exactly these investors are in a liquidity crunch.

What are they buying that they need to take out loans for? And is that a good business? It's hard to say, I do like the branding and these guys seem to have a good background. So Ashutosh, Aryan, good luck

Up next Quary, transform data together.

Quary connects your data warehouse and lets your team transform raw data into valuable insights in seconds, right from your browser, right? So we've got a screenshot here showing exactly what they do.

Quary makes data engineering, collaborative, sharing the responsibility across the organization.

Anyone who writes SQL can build production grade pipelines in minutes.

Well, that sounds really valuable. this was something that I saw at Amazon quite a bit that when you have people who are not part of the data science teams, it can be very difficult to get the data that you want especially when some of these dashboards are constantly breaking and there's issues with them, so I can see the need for this.

All right. What about the team? Ben and Lewis here were originally from Luxembourg. Oh, Luxembourg. Nice. Very, very small country.

Ben founded Tumalo and led engineering there through its series a. Lewis survived a four year stint as a software engineer in finance at Amazon.

Oh my God. I love this. I didn't know that he was from Amazon, but that just brings this whole thing together. I love it. Ben and Louis, congratulations. Louis, my apologies for your having to survive.

Congratulations on being on the other side. Good luck to the both of you.

Up next, OmaCare. Get paid for taking care of your loved ones. Ooh, okay. We're building the technical infrastructure to train and pay the 53 million family caregivers in the U. S. Now this is interesting.

Backstory. When my grandmother was diagnosed with cancer, my family rallied around her to take care of her at home. The same happened when my cousin was diagnosed with a developmental disability. Well, this is such a noble problem that they're taking on, and that's a powerful backstory.

Arianna and James have frontline clinical experience. Medicine at McGill,

a Master's in Public Health at Columbia. Plus Software Engineering Experience from Amazon.well it's not totally clear who is the engineer, but I'm assuming that's James, I'm assuming that Ariana has The med degree from McGill.

So let's understand more what's going on here. Last year, family caregivers in the U. S. provided 36 billion hours of unpaid care, worth 600 billion.

We unlock a way for the 53 million caregivers to get paid up to 28 an hour for doing what they're already doing through automated Medicaid and Medicare reimbursements.

Interesting. so Medicaid and Medicare will pay for it. But I guess that's only if the person who needs the care is eligible for Medicare or Medicaid. And presumably there's gonna be a lot of people who aren't.

I think that's such an interesting problem, interesting business.

I have one skepticism about this, which is that Medicare and Medicaid notoriously pay out very little to providers. This is something that many doctors and health administration professionals complain about all the time, that they'll do something really expensive, And if the patient is on Medicare or Medicaid, they get basically nothing back in reimbursement.

So I'm curious to see what the actual reimbursement rates here are. Obviously, in this case, these caregivers are getting nothing. And so getting even a small amount probably helps. But at the same time, it's probably not going to be a big hunk of cash, like they're probably not going to pay out huge amounts for people who are doing at home care.

Ariana, James, congratulations on launching. Best of luck to you.

Up next,, call centers for any business. We give any business a skilled, affordable call center. Interesting.

Even million dollar call centers lose 10 to 30 percent of revenue to abandonment. Open call captures this revenue while cutting costs 50 to 80 percent by answering every call with AIs. All right. So we've got a demo here. Let's check it out.

Hello. Thank you for calling Open Call Salon. This is Elisa. How can I assist you today? Hi. I'd like to come in for highlights. What day and time are you thinking of? Can I do next Wednesday at four? On Wednesday, I see that 4 p. m. is already booked. How about 2 p. m.? I can do two.

Okay. Google had a demo where they had something very similar a few years back I think.

How may I help you? Hi, um, I'd like to reserve a table for Wednesday the 7th. For 7 people? Um, it's for 4 people. 4 people? When? Today? Tonight? Um, it's Wednesday at 6pm.

I like this. call centers are not a sexy business by any stretch of the imagination. And you can imagine who's calling in. I mean, it's an aging population that wants to use phone calls to do business with companies.

Young people would rather use websites, apps, anything digital. However, much of the business world still runs on phone calls. So I see the need for this. I think it's interesting. That demo, not bad. It's somewhat convincing at least. So let's learn more about the problem.

When customers call a business, they expect fast personalized service. Unfortunately for businesses, that's prohibitively expensive to provide. Instead, hairdressers, mechanics and receptionists have to spend their mornings listening to voicemail and then juggle phone calls while working with clients.

So they gave us a really important piece of information. They're going after small businesses and they're going after receptionists, mechanics, hairdressers.these are going to be people who are hands on with their work and probably don't have time to answer phone calls. I also don't know how much money they have to spend on a service like this, but presumably the pricing here is pretty good.

they got to crack the code on selling into those very large organizations. Because that's where the meat is going to be. If they can get Verizon to adopt this for their call centers, they've really got something, but you've got to work your way up to those big businesses. And it seems like they're starting with small businesses to do that. I think that's a smart strategy as long as they're willing to go through the tedium of trying to fix call centers, which I can tell you are no fun. So congratulations to the team on launching good luck.

Up next, Fluently use zoom calls to improve your English.

Ooh, I like that most effortless way to become fluent.

fluently is an app that helps non native professionals improve their English with instant feedback after each zoom call. Imagine having a personal coach always listening and providing tips and they have a demo here. So let's click through this and see what we got. Okay. So they've got pictures of the type of feedback that they offer. It looks interesting. I mean, I'm curious about who exactly is giving this feedback. I'm going to guess that it's AI, The problem is that getting fluent in English is hard.

English level hits a plateau. Once you reach a certain level, it becomes difficult to progress further with traditional methods. Second problem, learning requires a lot of time and effort. So as suspected, this is being done with artificial intelligence.

It doesn't look like your Zoom calls actually have people on the line, although actually think that would be kind of a smart way to plus this up. To say, okay, yo, we've got a base plan that's free. But if you want to pay 100 an hour, there's actually a person. I think it's really smart. So we don't have anything on the founders here. It looks like. But looks like it's Uri Rebrick.

this seems like it's a business that he's building for himself. It's great. Love it. Congratulations, Uri, on launching.

Up next, Eggnog. YouTube for AI generated content. We let anyone remix AI videos. Well, my videos are all up on YouTube, so I'm very curious about this one.

Okay, from the founders. We're on a mission to enable everyone to create, share, and enjoy the magic of AI generated videos with consistent characters. Check out the most popular video made with Eggnog. It got 100, 000 views on Twitter in 24 hours. Paul Graham even said it was funny.

I live in the Dog Patch near downtown San Francisco, the epicenter of innovation. My name is Ben Bitdittle.

I'm 24 years old. I believe in taking care of myself and a balanced diet and rigorous exercise routine.

in the morning, if my metrics are down, I'll check the analytics dashboard code to see if there's a bug.

There never is. After checking the code, I check Twitter. I scroll through a carefully selected list of top tier founders, then cutting edge journalism, then my retirement account. Okay didn't think it was that funny, but I think it's an interesting concept. So let's get into the problem.

Inconsistent characters kill so much of the magic of AI videos. It's hard to get invested in the protagonists of a story if they look different in every scene.

AI video creators are the customer here. And everyone loves Eggnog. I do like that name. You guys can tell I like these whimsical names. Duckie we covered in the last video, now Eggnog. like when these guys are playful. It makes the world of startups more fun. Okay, so what does Eggnog do? It gives creators the control to design their characters exactly how they imagine and lets them use those characters in cinematic scenes. Creators can also remix characters made by other eggnog users, A lot of the AI video I've seen, even from OpenAI Sora, there's like a lot of issues with it. And I don't know that character consistency is the biggest issue today. The biggest issue is that the clips you can generate are really short. It's very expensive, very time consuming to make them. Those all seem like bigger problems in AI video than character consistency.

You kind of need all those other things to get to character consistency. But if they somehow have a model that is far ahead of where some of the top guys in the space are right now, which is 60 seconds max or less, and OpenAI hasn't even fully released that yet. But overall, I do like what they're building.

Up next, RadMate AI the AI co pilot for radiologists. We help radiologists dictate error free reports with 30 percent fewer words.

Well, I got something to say about this. My brother is a radiologist and I've talked to him at length about the impact of AI on multiple occasions. And so I actually have some opinions about this.

TLDR we're building the AI co pilot for radiologists. While reading images, we allow radiologists to dictate what they see and RadMate AI will generate and proofread reports for them.

So one quick note on this name, Rad Mate, there is a massive radiology company called RadNet. So it's not uncommon for radiology businesses to have "rad" in the name.

Rad is not like the rad that you would hear commonly, like "cool" . Maybe if you were, born in like the 60s or something, you're like, that's rad. But in this case, rad is radiology.

The problem, radiologists spend most of their day reading patient images and dictating reports based on the images. As patient access to imaging continues to rise, radiologists face escalating pressure to produce detailed reports quickly.

want to give a little bit of context about what radiologists actually do when you have an issue that you need to have scanned, whether that's through an x ray, MRI or otherwise, sometimes, not all the time, but sometimes a radiologist will review it and the radiologist is essentially a highly paid physician who knows how to spot things. For example, a radiologist might look through an MRI scan and say, I think that patient has cancer. Or they might look at an x ray and say, I think this patient has a hairline fracture in their tibia, so they're going to give context about what actually happened to a patient so that the doctors who are then treating the patient know what the expert thinks on it, right? Just based on the images. What are the images showing? This is important because some things are very, very hard to see. So even a doctor who has a lot of experience, this happens a lot with surgeons, who are like, oh, I can look at the image. I don't need a radiologist.

But in practice, radiologists are so highly trained in examining these images that it's actually very important and very helpful to have them in the mix. So little primer on radiologists.

Anyway, back to Radmate. Let's talk about the solution. We've created a platform with an AI copilot that decreases stress on radiologists and increases accuracy by Minimizing the words radiologists have to say. reducing the number of clicks required for report generation, and proofreading reports to reduce errors.

I actually really like the way they're approaching this.

For so many of these AI startups, they're basically trying to replace A worker of a certain type. They're like, Hey, our technology does everything that this low level software QA engineer does, like, you no longer need to pay that guy. Like, get rid of them. I like that these guys are proposing this in a complimentary way.

I mean, this is something that my brother harped on. He's like, yeah, everyone has been saying that aI is going to replace radiologists for a long time, and he's like, it hasn't happened. There's a lot of business reasons that it's not happening and it won't happen. But I like the idea of instead of saying, well, the machine can read these images better than a radiologist can to instead say, well, here are the problems that radiologists have.

And our solution is just going to help them with their problems. It's not trying to replace the core thing that they do. Which is reading images. It's just trying to replace some of the drudgery of the work that causes radiologists not to be happy about it.

Now, there's a little bit of a wolf in sheep's clothing here, because at some point I have to imagine that these guys are going to look up and say, well, Hey, we've got the whole market for AI copilots for radiologists, and we have all this data on imaging.

Maybe we should start to just read the images ourselves. And like, it'll start out as a suggestion. And then the next thing, the hospitals are saying we don't need radiologists anymore. That's a doomsday scenario, right? could see where these guys are going.

All right. So let's talk about the founding team. Muhammad and Adam are technical co founders who met while studying computer science at Cornell university. Before starting Radmate AI to solve the problems his dad faced as a radiologist, Adam worked at Palantir on the initial launch of Palantir's AI platform.

So you guys may know by now, I don't love it when teams going after healthcare, don't have providers on the team. But this is a case where the guy's dad is a radiologist.

He's probably has been hearing about all this stuff since he was a kid. I think that's cool. Like, the reason you want to have providers on the team, I think is actually because of non healthcare founder burnout. Like if you've never been in healthcare and you start working on a healthcare startup, it is like entering a whole new world.

And for me, what made it easier was that I come from a family of doctors, parents, siblings, other relatives. So I'd been around it long enough where I wasn't phased by the things that I started to see, but it doesn't work the way that other technology businesses work. It is a long haul and it's a slog. So this guy's dad was a radiologist.

I think that's pretty cool. Like the business. Mohammed and Adam, congratulations. Moving on.

Attunement. Patient monitoring and treatment recommendation for better behavioral health. Another healthcare startup. I like it.

One in three Americans suffer from a mental health disorder. Medicaid alone spent approximately 110 billion per year, that's an eye watering number, on behavioral health services to address the crisis last year. Without better measurements to track whether the patients are improving, it is a challenge to evaluate the impact of these dollars accurately.

Our software platform provides automated patient assessments at regular intervals, incorporating chat based mental health detection, physiological biomarkers, and self reported inputs of patients currently in treatment.

Okay I like it. I think getting patients to use software is sometimes very, very difficult.

I would be curious to know how they make it fun and interesting. Because I think, when you're building for the professionals, it's more just good enough to have like a clean system with a lot of information. And they'll do it because it's their job. Patients, on the other hand, often don't do the things that their physicians or providers tell them to do, including fill out this form every week.

If they tie it to money so if they say something like, Hey, your insurance company is not going to reimburse this unless you fill out these forms, Well, then you've got something.

So maybe that's part of their plan. I think partnering with payers on this could be really, really powerful. Obviously the largest payer would be Medicaid, who they referenced here. So you know, if they can get those guys great, but partnering with any insurance company would be great. Insurance companies have such a long sales cycle for this kind of stuff. It's very difficult. So as a startup, you just need a long runway and you need to be very persistent about trying to get them to run pilots. But, I think for a business like this, getting insurance companies to get involved would actually make them much, much more likely to succeed because a lot of these patients just aren't going to listen to what their doctors tell them to do, unless there's money involved.

Like the business. I don't see anything about the founders here, but Angie, Briar, congratulations. Good luck to you.

Up next NowHouse, modern post trade processing, building a better Broadridge, starting with intelligent trade reconciliation. Interesting.

Many of you acutely recall quarantining at home in January 2021, nearly a year after the pandemic started. If you're anything like me, you passed a lot of time switching between the same three or four apps on your phone amid curfews and indoor dining bans. If any of those apps was a brokerage, you might also recall this headline.

This is from the New York times. Trading platforms are limiting trades of GameStop and other companies. TDTwo weeks after the wall street bets driven surge in the price of GME, that's the GameStop Robinhood restricted trading, drawing the ire of its users.

One of the things Robinhood had to do was reconcile or match its trades against its executing brokers and ultimately against those cleared by the NSCC.

a few words about this, the back office of executing trades. That's buying and selling stock through the stock market is so much more complicated than what we see as lay people with brokerage accounts.

So you might have A Robin Hood account or TD Ameritrade or Schwab. And in those systems, it's generally very easy for someone to go and buy a stock. It's simple, but there are a whole lot of people in between you placing your order and you actually getting the stock you want.

They've figured out how to make it relatively fast, but a lot of it involves things exactly like what now house is doing. So I think this is actually a very interesting base for a business. I like companies that are focused on, the back office of financial services. I think that can be very powerful.

I mean, one of the crown jewels of the FinTech market is plaid. Stripe is another one YC company. So you get these companies which are doing things which the consumer doesn't necessarily see, but that's actually very important for your ability to make payments. So think this is interesting. I actually really like it.

The vision is to become the de facto system of record for firms and their regulators on the three trillion shares that change hands in the U. S. markets every year. don't think they need to get that many of them to have a meaningful business. So founder's background. I was previously the growth lead for SoFi brokerage, where I saw this problem firsthand. Prior to that, I designed the core ledger API and reconciliation tools at an e commerce startup acquired by Shopify and study computer science at Stanford.

I really, really liked this guy. Deval, you're the man. Congratulations on this. Godspeed. Good luck.

Up next, GiveFront. Banking for non profits and organizations.

We empower organizations to manage banking, spending, reporting, and compliance through a single, unified platform.

Our individual journeys began in the nonprofit sector, and we met while managing the finances for an organization at Harvard.

Our platform is more than a bank.

It's a comprehensive financial operating system designed to help nonprofits streamline operations, maximize resources, and ensure financial transparency.

This r eminds me a bit of what ramp and brex do so brex is a yc company ramp was started by Couple of my batch mates from the summer 15 batch at YC.

So YC is obviously having a massive impact on financial services overall. This is similar in the sense that what ramp and Brex do is something they call spend management. they're essentially trying to get. People to spend less in an organization overall, or really to get them, benefits for spending in a certain way and to help them organize and manage how much they spend on all kinds of stuff.

So those companies have done fabulously well. These guys are saying something similar. They're basically saying, yeah, we want to help you manage your banking. But there's a whole lot more you need as a non profit that we can also do for you.

The big distinction of a non profit is that they get tax breaks. Like, they can run an organization that has huge budget without having to pay the same types of taxes that a regular company would have to pay. But in order to get that, there's a very high compliance burden. So what these guys are doing, they're just making all of that stuff easier.

I think that's cool. I think they can find a lot of great partners. I think it's a noble mission. Really like the business.

We got a photo of the founders here hanging outside of the Mountain View office for YC. Matt, it's a Harvard alum and formerly founded a FinTech startup.

He previously ran and served on the board of two national nonprofits.

One small note about this. I love second time founders. Once you've had a chance to go through this and you know what it's like, you're much more likely to succeed.

you get a lot of scar tissue. So you do this again. Over time you start to really feel that monkey on your back of just the weight of carrying a company. It's a lot, but when somebody has done it and they're onto their second one, these guys look like young guys. I really liked that they are former founders and they're doing something that is noble. GiveFront, congratulations on launching.

Up next. Creo build internal tools with AI. Okay.

All right. What's the problem? Internal tooling is a massive space with a lot of no code builders taking the cake. We feel strongly this is somewhat of an anti pattern as it's mostly engineers who are interfacing with these tools.

With LLMs, the speed at which you can scaffold an internal tool in code is orders of magnitude faster than trying to fiddle with a no code builder.

I've talked to a number of CTOs about this who agree strongly with the idea that startups should build their own internal tools instead of spending a bunch of money on SaaS.

Now, I think when you're just getting started, it makes sense to outsource as much as you can. But if you can build a tool that you use internally that costs you almost nothing versus paying 20 per user per month, it really can help your organization and save you a lot of money in the long run.

And a lot of times those tools that you would build for yourself, at least as, a couple of my CTO friends have told me, are going to be better and more customized to what you do than the SaaS product that you would hire to do the same thing. So I actually think this space is going to be really big.

I think it's smart.

Let's see what they have to say about what they built. all you do is write the React you're used to, or ask our AI to do it for you, in a single folder called Tools in our starter project. Once you deploy, we give you a seamless interface to interact with your tools.

You can also add your organization members, configure permissions, and manage your deployments. And we have a demo.

Here's a warehouse shipping tool where we keep track of our shipments. I want to make some changes to this. Let's start by adding a card below the table to display data when a shipment is selected. Our card got added below The table with row information. Now let's move the card to the right of the table to make it easier to view. Okay. I like it. I like the business. This is cool.

Rohan Saif. Congratulations. Cool concept. Best of luck.

And that is it for our startups for today. We covered another 11 great startups. Thanks for listening. Thanks for watching. And we'll see you again soon.

Watch next