
Lessons from 1,000+ YC startups: Resilience, tar pit ideas, pivoting, more | Dalton Caldwell (YC)
What this covers
This is a conversation between Lenny Rachitsky and Dalton Caldwell, a group partner at Y Combinator (YC)) who has run thousands of office hours with founders. It begins with his own history. He sold his music startup imeem to Myspace, then built Pick Please, a mobile photo-sharing app with filters that Marc Andreessen backed and that Instagram eventually outcompeted. The rest draws on what he has seen across YC batches, framed around the Don't Die mindset and a defense of simple, repeated advice over novel strategy.
The middle of the episode is practical. Caldwell lays out what makes a Good Pivot: it moves toward the founders' expertise, not away from it. Brex's first idea was a VR headset built by programmers with no optics or hardware background, and Retool began as a UK Venmo competitor. He contrasts these with Airbnb, whom he advised to keep iterating because they still had untried growth ideas. He describes the Tarpit Idea, explains why investors say no, and covers customer conversations. His advice there is to spend roughly 20-30% of founder time with customers in person and to validate or pre-sell before writing code. He also warns against over-delegating to senior hires with impressive résumés.
Some positions are contested. He argues that TAM (Total Addressable Market)) matters little at pre-seed, citing Uber, Airbnb, and Razorpay's bet on Indian credit card adoption in 2015. He also says growth hacking, analytics, and A/B testing are unhelpful for startups with no customers yet. Later sections turn to the small, close-knit Silicon Valley of the early 2000s, including Sam Altman's earlier Loop project, and to Caldwell's story of a16z's conflict over Instagram. The episode closes with YC's Request for Startups, a failure segment, and a lightning round. The failure segment holds that a shutdown handled with integrity does not define a founder.
Dalton Caldwell argues that startup success fundamentally hinges on founder conviction and persistence through inevitable near-death experiences, not on having a perfect initial idea, large addressable market, or sophisticated growth tactics—and that the most consistently fatal mistake is founders losing hope and giving up when capital and moves remain.
- Nearly every successful YC company faced moments where founders rationally should have quit, but irrational persistence led to success (Airbnb, Brex, Retool examples)
- Founders fail most often because they lose conviction and stop trying, not because they run out of money; many startups die with capital still in the bank
- Good pivots work because they move toward founder expertise and prior knowledge, not away from it; conviction builds through customer feedback and product-market signals, not upfront certainty
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Even the best athletes and performers rely on being coached and reminded of fundamentals rather than discovering novel strategies, because psychological mindset on the basics is what separates elite performers from others.
“if you actually listen to what the greatest smartest most successful athletes are talking about like if you listen to what Tiger Woods is saying to his caddy it all sounds like pretty mundane stuff they're not it's not like what Tiger Tiger Woods is talking about with his caddy as some you know imp possible to decipher jargon it's like yeah you really need to keep your head down on this one”
Growth hacking, analytics, A/B testing, and split testing are actively unhelpful for early-stage (seed/pre-seed) startups with no customers, and founders should ignore this advice until they have product-market fit; this advice is calibrated for series A/B companies and big tech products.
“I think growth and growth hacking and doing all this analytics AB testing stuff um is a total waste of time for very early startups”
Razorpay, the largest payment processor in India that YC funded, was founded when India's credit card TAM was tiny (very few Indians used credit cards in 2015), but the founders correctly bet that credit card adoption would grow 100x, and this bet paid off as the market materialized.
“I funded Razer pay which is I think the largest payment processor in India and the Tam of that was Tiny because no one was using credit cards in 2015 in India so you had to believe that the size of the credit card industry in India would like 100x well guess what happened”
The Brex founders (in Winter 2017 batch) started working on a VR headset idea, appeared to be the worst company in the batch, felt ashamed of their idea, nearly gave up, then pivoted to fintech and became a decacorn.
“this one company in in my group this batch was called theond that was their name at the time and it was like a VR headset thing from these uh Stanford dropouts and they basically showed up to group office hours and we're just ashamed and they're like our idea is horrible you know we might want to shut our company down”
Gustav Söderström said the most common reason startups fail is that they don't talk to customers and don't find product-market fit, and Caldwell agrees with this assessment.
“I asked him what is often the most common reason a startup fails and his answer was they don't talk to customers they don't find product Market fit nothing else matters if they can't do that”
Retool (originally Cashew, a UK Venmo competitor) was also one of the worst-performing companies in the same Winter 2017 YC batch, nearly gave up, pivoted to internal tool builders, and became a successful company.
“there was another startup also in my group called cashew which was this PTP for uh the UK PTP venmo excuse me um in the UK and it was going really poorly also and not growing”
In the early 2000s Silicon Valley startup ecosystem was very small, with only 30 people at conferences and a tight network of people like Mark Zuckerberg, Reed Hoffman, Sam Altman, Elon Musk, and Sean Parker, who were introverted 'weirdo outsiders' obsessed with internet and computers.
“it reminds me of stories about the homer Computer Club I'm not saying this is cool but when I read stories about what it was like when when the Homer Computer Club existed it was a very small number of people that all knew each other that were real like weirdo Outsiders that were into this stuff okay and so that's what in the post.com boom bay area startup scene that's legitimately what it felt like”
Caldwell sold his music startup (imeem) to Myspace, was briefly recruited by new CEO Owen Vanada (former Facebook COO) to fix Myspace with a mobile photo-sharing idea, but Vanada was fired within a month along with much of the leadership team.
“my my second company um basically I sold my first company in Myspace it was the music company that I worked on and the they recruited a new CEO um who was formerly the COO of Facebook called Owen vanada”
Because Marc Andreessen was on Caldwell's Pick Please board, a16z had a conflict of interest and could not invest in Instagram, causing a16z to miss out on the Instagram deal and creating Silicon Valley gossip about Caldwell being the reason a16z missed Instagram.
“basically because uh andran Horwitz had like invested in my company was on the board even though they were investors uh in Instagram um that was like a conflict they and they didn't do the deal”
The Stripe story illustrates founder persistence: Patrick Collison (Stripe co-founder) was personally checking in on early customers like Caldwell via Google Talk on a weekly basis, showing hands-on customer care even as the company scaled.
“I was an early uh stripe customer my start up and yeah Patrick would like um we used Google talk at the time Patrick would be sending me messages like on a weekly basis just like checking in”
Market size (TAM) matters very little when evaluating pre-seed or seed stage companies, because the market will likely change dramatically and successful companies often started in markets that looked tiny (Uber, Airbnb, Razorpay in India's credit card market in 2015).
“I think it really depends on what stage you're investing at and it's absolutely critical the later stage you get right if you're going to invest at a very high valuation it is really important um the earlier you go the less it matters and some of the most phenomenally good startups if you were really uh panic about it the Tam would be like tiny like the The Tam of uber would be like nothing right like how um The Tam of Airbnb would have been nothing The Tam of I was uh I funded Razer pay which is I think the largest payment processor in India and the Tam of that was Tiny because no one was using credit cards in 2015 in India so you had to believe that the size of the credit card industry in India would like 100x”
Caldwell founded Pick Please, a mobile photo-sharing app with photo filters (similar concept to what became Instagram), raised angel funding from prominent investors including Marc Andreessen, got traction with hundreds of thousands of users, but was eventually outcompeted by Instagram, which was a Foursquare clone that pivoted.
“I ended up yeah starting the company and quickly was able to raise an angel around because people remembered uh my company from the first one and the the major investor we had was Andre Horwitz this is one of their first board seat Investments like Mark andreon was on my board”
Failure in a startup does not define you long-term; no one remembers startup shutdowns years later if you handled yourself with integrity and honesty, and fear of failure should not dominate decision-making.
“I guess what I learned is that you just can't let it get to you too much and you got to keep going and that if you keep going no one really remembers those as much and it doesn't really Define you”
The signal that it's time to start building your product (beyond validation conversations) is when you've found people genuinely excited and willing to be early customers or users.
“and basically if you if you find people that are really excited and you you do line up customers that is a great green light that it is time to do a startup right that can get you down the path so yeah I think that's my final advice”
Founders should avoid over-delegating to senior hires with fancy résumés early in the company's life; the best product leaders stay deeply involved in product decisions, talk to customers regularly, and maintain hands-on care about product quality regardless of company stage.
“the advice that he's referencing here is just how important it is to not overd delegate and for the founders to stay close to things um as well as watch out for the Trap of hiring super senior people with fancy resumés really early at a start up”
Founders should spend roughly 20-30% of their time talking to customers in person; spending time networking with investors or doing other activities is lower priority than direct customer interaction.
“look at your calendar and there should be you know 20 or 30% of your time that the calendar says something like customer meeting customer call like meeting with food meeting with this person”
Brex's first iteration was a VR headset startup founded by Stanford dropouts, which failed because the founders were skilled programmers but had no expertise in optics or hardware manufacturing, violating the principle of pivoting toward expertise rather than away from it.
“they thought it was really high-tech they want to do a really high-tech startup and so they're like we're going to build a new VR headset and you know they were good programmers but they just didn't know anything about Optics or the things you might want to be an expert in to build a headset”
Early versions of Uber had effectively zero addressable market because few people used ride-sharing in the early 2010s, yet it became a massive market; the TAM wasn't there at the time, it had to be created through product and scale.
“The Tam of uber would be like nothing right like how um The Tam of Airbnb would have been nothing”
Instagram's founders were strategically smart in copying hipstamatic filters (a paid app) and combining them with a free social sharing model, recognizing that users wanted what they had to pay for and could now get free.
“what I think they did that was smart is if you looked on the paid app store charts the number one app was hip stematic and hypsta cost money and what do we know about what people want they want things that are free that cost money right and so they basically built a pretty legitimate knockoff of the hip static filters combining it with this social graph”
Airbnb tried many unconventional growth tactics including selling cereal and going to conventions, and they had a large pipeline of untried ideas, which is why Caldwell advised them to keep iterating on the core business rather than pivot.
“in the Airbnb story right they they tried uh all sorts of stuff including cereal and conventions like they had a bunch of zany ideas on growth and they didn't run out of them”
Sam Altman worked on Loop, a feature-phone application for finding nearby friends on Boost Mobile, and this was years before he became president of YC and later reinvented himself around AI.
“when I met Sam he had dropped out a Stanford to work on Loop which is hilariously a way you find people around you to hang out with h interesting theme here huh he was cool just he was this really young guy and um he just kind of did that it wasn't huge and then he got into other stuff and ended up working at YC ended up getting involved in hardtech and is now kind of like reinvented himself as um the the big mind behind AI”
Before writing code, founders should do customer validation and pre-sales by talking to potential customers; this is the best starting point for deciding whether to build a startup.
“basically if you if you find people that are really excited and you you do line up customers that is a great green light that it is time to do a startup right”
Facebook got 100% penetration on the Harvard campus first before expanding to other universities, rather than launching broadly; this is a better zero-to-one strategy than broad distribution.
“I think if you look at the Facebook story them getting 100% penetration on the Harvard campus first instead of launching overall again good strategy would recommend U that strategy”
Post Hog (an analytics company) benefited from the open-source approach: they launched an open-source version, got organic excitement from Hacker News and users implementing the product, and had their calendar filled with people wanting to discuss the product, which was very helpful for early feedback.
“post hog talk I guess that's a different go to market they launched this open source thing to start with and it was they their calendars are filled with people that were trying to implement the first open Source version of post hog or so excited about it and people on Hacker News were excited about it and like they had this like huge influx of people that were excited that post hog existed and had lots of feedback”
Whatnot founders grew their marketplace by focusing on buyers first and understanding marketplace dynamics deeply, rather than spending money on Instagram ads hoping to get users.
“in the case of whatnot they obviously it was consum app and they needed to get users um but they were very intellectually honest on the metrics for how you get a Marketplace off the ground and they didn't just go dump all their money into Instagram ads and like they effectively knew they needed to focus on buyers and on the buy side and build momentum on the buy side um they really understood marketplaces”
YC and Caldwell have had great success funding space companies, including several that are actually going to space and are not SpaceX.
“we've had a lot of success with space companies um you know uh several of the folks that are actually going to space right now that aren't SpaceX are YC companies”
Caldwell's favorite interview question for founders is straightforward and open-ended: 'Tell me what you're working on' or 'What have you learned since you started?' because the most honest and interesting answers come from simple, blank-slate prompts rather than trick questions.
“I guess Founders in this context I don't really believe in in trick questions and I think I just start with hey so tell me about what you're working on or what have you learned since you started you none of these are trick questions but I think you can get the most honest and interesting Answers by asking the most straightforward basic things”
Zip's founders Rul and Lou grinded on customer development by calling procurement decision-makers on LinkedIn for advice about their procurement product, and did hundreds of these customer discovery calls, turning them into beta testers.
“on zip I actually have a lot of their story in one of my series on how to build a B2B startup and what they did actually as you know uh is they just called DMD people on LinkedIn and asked them for advice on hey we're we're trying to understand how you enjoy your current procurement products and then they ended up being early beta testers I think they did and I think they did hundred like hundreds of these they just had oh yeah no it was a numbers game they they were just grinding at this”
Founders fail most commonly not because they run out of money, but because they lose hope, have fights with co-founders over direction, and then resign themselves to shutdown while capital remains in the bank.
“it's pretty rare I would argue that the cause of death is that they had lots of Firepower and they were feeling really positive and they just ran out of money that's actually like more rare than Founders think um it's much more common that they still have some money left I'm not saying a lot but some money and they're just like yeah I'm done I'm out of ideas I don't want to do this anymore”
A tarpit idea is one that many people come up with, seems like an unsolved problem, gets positive initial feedback, and has good arguments supporting it—but has structural barriers that make it consistently fail despite validation.
“by definition it is only a tarpit if it seems like it's not like like if it's just a regular idea that is hard that is not a tarpet the the weird aspect of what we call a tarpit idea is an idea that a lot of people come up with and that it seems like an unsolved problem and you get lots of positive feedback for right and you have a really good set of arguments that it's a really good startup idea”
YC is putting out a request for startups with 20 categories of ideas they want to fund, including ERPs, open source companies, space companies, cancer prevention, spatial computing, defense tech, manufacturing reshoring, and better enterprise integration tools.
“recently you put out a request for startups 20 categories of ideas that YC wants to fund”
Successful founders don't share a common personality type (introverts succeed, extroverts succeed, quiet people succeed, charismatic people succeed), but they do share a common trait: they deeply want success, believe in themselves, and have an internal conviction that 'I'm the one and I won't accept failure.'
“I don't think personality types matter as much I've seen very quiet people very extroverted people very you name it I've seen all sorts of personality types so for me personally I don't think that there's a right or I don't think there's a personality type that people should copy”
Founders should identify their 'comp' companies (companies similar to what they're building at early stage), study what those companies did in their 0→1 phase, and replicate that playbook rather than copying their current advanced tactics.
“don't pay attention to what Facebook does today if you're a brand new startup founder pay attention to Facebook when they were getting their first thousand users what were those tactics”
The underlying theme of all successful YC startup stories is that rationally the founder should have given up at some point, yet irrational persistence and refusal to accept failure drove them forward.
“if you look at all the startup stories that we have at YC and all the companies we funded over all the years the underlying theme is that rationally the founder should have given up at some point”
Successful founders have the ability to reinvent themselves across multiple 'eras' in their career (e.g., Reed Hoffman at PayPal → LinkedIn → VC; Sam Altman at Loopt → YC → AI), showing that careers are long and individuals can shift directions while maintaining core commitment to internet/tech.
“you'll see these folks just reinvent themselves multiple eras right okay like Reed hoffen right he was a worked at PayPal right and then uh he did LinkedIn and then he was like a VC and like like he's kind of had like all these different eras where it's the same person but it's it's almost like a different figure right”
Most founder applications to YC have similar startup ideas because they consume the same information diet (same podcasts, Twitter feeds, blog posts, and social circles), so founders should deliberately diversify their information sources to find differentiated ideas.
“people all kind of have the same idea like like basically imagine this imagine your information consumption where you're listening to the same podcast wink wink um you're reading the same people on Twitter you're reading the same blog post basically you have the same information diet of all these other Founders and you're friends with all the same people does it seem surprising then that you would all end up with similar startup ideas or similar philosophies on what makes a good startup idea of course you are”
A founder deciding whether to quit should check if they are still having fun and enjoy working with their co-founders; if the answer is genuinely 'no,' and it's not just a temporary rough patch, quitting is reasonable and not shameful.
“one are you still having fun do you still enjoy doing what you're doing do you enjoy spending time with your co-founders you know like is this actually a fun thing you're doing and if the answer is is yes I would tend to lean on the keep going and then if it's more of wow this is actually profoundly affecting me a negative way and my relationships with people in my life and my team you know I don't really want to work with my co-founder anymore and things like that then I would lean on the probably don't do it anymore”
The core mantra that proves most impactful with founders is 'just don't die'—keep your startup going—because being coached on fundamentals and persistence puts you in the right psychological mindset to execute.
“one of my mantras is just just don't die just keep your startup going just keep going and I say that over and over again and and honestly that is often what people tell me is the most impactful thing I said it's not that I said some Ninja 5D chess move that they never would have thought it before it's just the constant affirmation that continuing to keep going and doing high quality reps is the game”
Most founders avoid direct customer conversations because of social anxiety and fear of looking stupid, and they need to accept and work through this discomfort until real customer engagement becomes normal.
“a lot of the reason people do that is they they're just shy and they don't want to put themselves out there because it's a little awkward to go talk to people and and you kind of have to Su yourself up to go out in the physical world”
The best introductory book for founders learning sales is 'Getting to Yes' (a popular top-seller on Amazon); reading basic sales books provides 80% of the value of expensive sales coaching.
“go on to Amazon and find the most popular sales books like getting to yes that everyone reads and just read those and that'll did you 80% of the way there”
A good pivot is one that moves closer to something the founder is already an expert in or has prior experience with, and builds on learning from the previous idea, rather than pivoting toward entirely new territory.
“usually a successful pivot gets warmer instead of Colder from what you're an expert at and somehow Builds on what you learned on the prior idea right”
Caldwell advised Zip founder Rul to identify publicly traded or private-equity-owned companies that are large, hated by customers, and have terrible software, then build a software solution to disrupt them.
“my suggestion was to start by looking at what companies are um publicly traded uh and or owned by private Equity that are large and that also are hated uh by their customers and to try to intentionally find where there's a knowable big Market with an incumbent um combined with the software is horrible”
Investors say no to startups most often simply because they have limited capital and opportunities, and are holding out for ideas that hit a higher bar—not because there is a hidden secret objection or because of minor flaws like poor Zoom setup or shirt color.
“a lot of investors just don't make that many Investments and as for what we talked about earlier life is short and so there's lots of things that an investor that in their hearts thinks is like pretty good and they're oh like I like this person and I like their pitch but I only am going to do a few Investments and so even though I really like a lot about this I'm going to say no”
Successful founders do not have natural conviction at the beginning when they have no customers and no working product; conviction builds over time through customer feedback and product traction, creating a positive feedback loop.
“and this is the big butt no one has this at the early stages when they don't have a good idea and they don't have customers and like it's objectively not working and so I again I know a lot of Founders like well I don't feel that way oh no maybe I'm you know maybe I'm an imposter and I shouldn't do a startup well of course you don't feel that way if you haven't talked to any customers and haven't built a product”
Successful founders that turned around struggling startups had a common trait: deep love for their customers and product and co-founders, knowing customer names and details personally, which signals psychological conviction worth betting on.
“something that a lot of the folks that turn it around have't common is they actually do love their customers and they love their product and again if in the Airbnb story again you know it really well but they really liked Airbnb like and they liked working with each other and they liked the first host that they met and they knew all their names you know what I'm saying like they were actually they loved their startup even though it was going bad”
Trucking startups became fashionable and widely attempted after initial success, then the space became crowded with mediocre companies trying to copy the winners, demonstrating how tarpit ideas form: one success creates a wave of imitation that makes subsequent attempts less likely to succeed.
“a few years ago startups around Trucking were super new and fresh because no one was doing them and they worked really well and then it became completely conventional wisdom to do like Trucking related startups”
Overfocusing on TAM size predictions when evaluating pre-seed or seed companies is not useful because early TAM projections are speculative; instead, YC focuses on whether the founders can make something people want and whether they can grow it.
“trying to be super pedantic about Market size when you when it's like a preed company or someone applying to YC is not you know it's just not something I put a lot of thought on you get whatnot oh what's the Tam of the collectible Funko Pop industry I don't know I don't think it's that big man”
When founders hire senior executives from big tech companies early, they often wake up to discover things have gone wrong—not because anyone is at fault, but because the founder took their eye off the ball by delegating away their core responsibilities.
“and then you wake up one day and you're like oh wow everything went wrong it's not really anyone's fault it's just that you you you o you took your eye off the ball and this is what happens um to first-time Founders lot”
This internal conviction and belief ('I'm the one') acts almost like an internal gravitational force that warps the world to bend to the founder's will; people believe because the founder believes so much, and this influence extends to employees and stakeholders.
“somewhere deep down in there they're like oh yeah I'm gonna make this work this is this company is gonna be big and they and they they just believe and it's almost like that internal gravitational force inside of them is so large it kind of warps the world to bend to that will and people start to believe it because they believe it so much”
Caldwell's favorite life motto is to periodically check in with yourself to confirm you are having fun and enjoying what you're doing; if not, make a change—this applies to all career decisions, not just startups.
“find useful in work or in Life share with friends just check in with yourself that you're having fun and and that you enjoy what you're doing and if you don't you should probably make a change whatever that is”
Being excited and enjoying your work drives success more than personality type, and successful people get excited about the work early on, which sustains them through difficulties.
“and then the other point there is again coming back to you just need to be really excited and enjoy the work you're doing because that'll drive you forward and make you be successful”
Segment's founders couldn't have created their successful data pipeline product (Segment) from scratch because they didn't have deep knowledge of analytics; they only gained that knowledge through multiple years of grinding on previous startup ideas, which gave them unique insights that led to Segment.
“there's no way those Founders could have started with the final idea you get what I mean there was no Universe where they would have made up the idea for Segment because they didn't know anything about how analytics worked but because they were grinding for multiple years and became experts on these things is a side effect of their earlier ideas they ended up with really good unique insights”
Nearly 100% of startup founders experience a near-death moment where they seriously wonder if the company will survive, making this a nearly universal part of the startup journey.
“I would argue that if we Define it is the company had a near-death experience where it was going poorly and the founders seriously wondered if it was all going to be over 100% of the time people go through that you know where the found like yeah I guess we're done guess we should pack it in and at least you feel that way at some point in your startup journey I me everyone goes through that”
Life is short with limited career years; spending those years forcing yourself to work on something that makes you miserable solely to avoid losing face represents a large opportunity cost on your actual life.
“we have such a short life there's only so many years we get to to to have our careers doing something that makes you miserable and the only reason you're doing it is to avoid losing face”
YC sees very few applications in enterprise resource planning (ERP) software despite its large market and strong unit economics, and putting it on the RFS list led to more founders researching and applying with ERP ideas.
“one of them I made one about erps which is you know enterprise resource planning software and I did that because I get so few applications on that and they're usually pretty good and I just would love to see more people look at that and learn about what erps are just because it's so rare that people apply with that and now I have a feeling we're going to see a lot more applications working on that and so it kind of worked as intended which is introduce this idea space to Founders that didn't even know what an Erp was”
Caldwell cold-emailed Reed Hoffman when LinkedIn had only 12 employees, and Hoffman responded casually and suggested lunch, showing that at the time these now-famous founders were accessible and not yet obviously successful.
“I remember um I cold emailed Reed Hoffman when LinkedIn was like 12 employees and he just responded and he's like oh let's have lunch it wasn't it he he was just like a guy”
When founders feel like failure is imminent, if they can identify even one more growth idea they haven't tried, that is often a reason to keep going and try it rather than immediately pivoting.
“in the Airbnb story right they they tried uh all sorts of stuff including cereal and conventions like they had a bunch of zany ideas on growth and they didn't run out of them and so I think when there's still uh gas in the tank um on an idea that might be a reason to State the course”
Many people go through life without regularly asking themselves whether they're having fun and enjoying what they're spending time on, and this lack of self-reflection prevents them from making needed changes.
“I think a lot of people go through life and they don't ask themselves this question like am I having fun am I enjoying is do I value what I'm spending my time on”
Among founders who experience serious near-death situations, a high percentage (roughly 50%+) survive through sheer will and determination to keep going despite adverse circumstances.
“Founders that actually truly got down to very very hard situations it's still a high percentage like maybe 50% I mean you can ask Founders there's a lot of Founders that come this close to it all being over and through sheer will kind of just keep it going”
When evaluating founder answers in interviews, look for evidence that they've actually thought about the problem - they've done research, have opinions, and care about the space - versus superficial answers that show they haven't put care into their thinking.
“I think evidence that they actually have thought about it like as PR said earlier that they've done research that they have opinions that they they care about it right sometimes when you when people answer questions they're like you can just tell that it's really super superficial and they haven't put much care or Soul into their answers”
YC is interested in small, fine-tuned models as an alternative to gigantic, generic large language models, suggesting there's opportunity in specialized rather than generalized AI approaches.
“small fine tune models as an alternative to gigantic generic ones yep sweet”
The most helpful thing for aspiring founders to do is apply to YC and start a startup; Caldwell loves seeing people apply with ideas inspired by the advice he shares in videos.
“I mean honestly it's just great when folks want to uh apply to YC and and do a startup and so feel free to dive into other videos and apply to YC and something this really special about my job is I get the privilege of getting to fund companies that they already know me from videos”
Caldwell enjoys rewatching old shows like The Sopranos, The Wire, and Columbo (1970s-80s police detective show) because they feel like a time machine to a different era.
“I like to watch old shows a lot and so I keep rewatching like The Sopranos and the wire and it always is different to me every time and I you know things like that I think here's a silly answer I've been really enjoying watching old episodes of Columbo which was a television show from the 70s and 80s”
Caldwell uses Oura Ring, Apple Watch, and recently signed up for Sciox (a YC company doing at-home blood testing with a small needle and mail-in analysis) to sync personal health data across devices.
“like my or ring and my Apple watch and all that good stuff like I've been a fan there's a YC compy called CCI Fox siox I just signed up for and they do at home blood testing and and basically I'm trying to sync that ad at home blood testing thing into all my other devices”