YC isn’t to blame here. It’s a microcosm of the incentives that have been set up. Unfortunately most of the game-changing tech isn’t going to be low capex moving forward. We are moving past the “Airbnb for dogs” era and into the “build a better quantum computer” era.
YC is absolutely to blame here. Their standard term deal sheets are unnecessarily greedy and massively distort the cap table for hardware startups that have to raise a lot of money before a Series A.
Their docs are predicated on startups being able to raise a small-valuation Series A right out of the program, which is only true of startups that require low capitalization.
For instance, I'm out fundraising for a molecular nanotechnology startup right now. I will not be doing YC because it would be a shite deal for both me and my existing investors.
Let me preface this by saying that on the whole I agree with you.
However hardware startups that have to raise a lot of money before a Series A are... considerably riskier, no? If you're an established name in hardware or a hardware-adjacent field then you'll typically have less of an issue getting access to capital. You can strike some pretty amazing deals with vendors if the stars align.
But otherwise there's a considerable chance that you end up out of your depth, out of runway while pre-production units sit on a loading dock in China.
Pretty much everyone prefers investing in startups that require low capitalization at this point. Everyone has been bitten by this point. We've learned.
Not to say it isn't possible but I'm not going to unbalance my portfolio towards hardware startups unless that's somehow my play.
Risk can be factored in to the terms of the deal. But there is no situation where it makes sense to offer the very first investors of an idea-stage company full, unlimited and indefinite dilution protection.
It used to be common and expected that early investors be given pro rata participation rights, so that they could maintain their ownership stake by participating in future rounds. YC’s new post-money SAFE terms effectively gift them equity in every future unpriced financing round. So if you need to do another convertible note round (which any hardware startup would), they effectively participate without having to give you a cent.
This is unheard of and ludicrous. But they get away with it because most software startups immediately raise a priced round on demo day anyway, so it’s a moot point for them. For hardware startups, this can become a poison pill.
I’m glad you took it this direction because it gets to the reason I brought up the other article. It’s not an accident that China dominates in so many technological areas. It is the result of a deliberate plan. The CCP specifically targets areas of investment. The US does this too, but usually in a less direct way (politics mean you have to keep constituencies happy which means farmers keep getting subsidies). The idea that the market is “free” is a myth.
So if we want companies to invest in certain high capex industries, we need to de-risk those investments. Change the cost-benefit calculation so that companies will invest in industries we need to grow domestically. CHIPS and Science kind of did this, but we need to keep doing it.
"hardware startups that have to raise a lot of money before a Series A are... considerably riskier, no?"
Why riskier? They have more of a moat don't they? The large capitalization needed suggests they will face less competition. It's more difficult for competitors to gather the necessary capital to compete against them?
What I've written here was the conventional wisdom for most of 200 years. Much of the Industrial Revolution played out when merely concentrating together capital was seen as an engine of growth. Rockefeller did not need to sell innovative gasoline, he only needed to use cash flow to buy up monopoly positions, one small region at a time, until he had the cash flow to buy up monopoly positions nationwide. Economies-of-scale meant that merely concentrating together capital was a path to greater profitability.
The last 25 years were an aberration, during which time big companies could be built with small initial investments. But over the course of centuries, the opposite was more common.
Is the moat that good for HW? The more commodified your product is, the more you risk losing to undifferentiated foreign competitors who have lower input costs.
And hardware marches on while you are trying to get your widget manufactured and shipped.
I was at a hardware startup in 2013/14. We had our own board design that was very similar to the RaspberryPi + Arduino that we'd prototyped with (we ended up using a iMX233 and an AT Mega 328). While we were debugging a manufacturing fault (out of spec led controllers), Expressif released the SDK for their at the time practically unknown ESP8862 - which meant 90% of our functionality could now be done with a BOM of around $15 instead of the $90 or so ours was costing us.
Our "moat" had been concreted over while we were pulling our hair out trying to ship in time for xmas. (And the business died in arguments, recriminations, fingerpointing, and a lack of ability to find investment to pay for a 2nd production run. And I needed up with another piece of paper saying I had a percentage or two of ownership in something now worth zero dollars...)
I was doing firmware in IoT/M2M for an engineering services company back then and it was the tail end of "slap a radio on [common product]" projects for that company. I saw a few projects end with the arguments and fingerpointing with inevitable laser focus on what the SOW actually said and it always sucked to end that way.
As a startup founder, all growth startup founders need to be a bit delusional — I don't mean that in a bad way. The few hardware startup founders I've met have either been very dialed into their trajectory with a clear deriskable exit plan, or extra delusional. It's fun talking to folks like that, but from an investor perspective I can see why your risk evaluation skills need to be dialed.
You're out of your mind if you think the YC pre-seed terms are unnecessarily greedy. Where in the world can you get half a million with nothing but an idea to show for it and only have to give up 7%+uncapped?
If you've already got investors, why are you looking at YC at all? YC funding is a product, you can't just go and say "ah this product has got features I don't need, so it's a terrible deal".
> Where in the world can you get half a million with nothing but an idea to show for it
In Silicon Valley. You can actually get a lot more than 500K.
> and only have to give up 7%+uncapped?
It is actually 7% for 125K and 325K uncapped, but at the best possible terms offered to other seed investors. So it might end up becoming ~10% for 500K, which is a lot of dilution.
YC created a path for 20-year-olds with no work experience to raise money for a company (and sometimes become billionaires). That requires low capital investment because the founders are completely unproven.
Hardware startups are run by people with work experience in hardware and a career track record, because you need more credibility to justify the higher investment required.
> China leads or is on par with global leaders in commercial nuclear power and electric vehicles and batteries. It lags behind for now in other key sectors, including robotics, biopharmaceuticals, chemicals, and AI. But it is making rapid progress.
I'm not sure why this is relevant. They're still behind in the primary sectors that Y-Combinator invests in: software.
Certainly China is catching up and it would be naive to sleep on them, but it is magnitudes easier to catch up than it is to lead (e.g. it is magnitudes easier to learn calculus than to invent it). It is hard to make serious predictions without understanding how to lead in certain niches and a broad overview of the economy will not tell us too much about specific sectors. There's plenty of niches that Europe leads the US (and China) in too. But the conversation is all pretty complicated. And success in one sector or niche does not seem to be a strong predictor of innovation in others. Similarly success in catching up does not seem to be a strong predictor of ability to lead. This is all related to the Middle Income trap and China has probably come out the other side but it's not clear yet and this is all very fuzzy.
FWIW, I want to see more competition on the global scale. Everyone benefits if there is healthy competition. Though this can also lead to war if things get too competitive. But if our main method of "war" is "making shit better and cheaper so everyone buys from us" then bring it on! (Maybe we can get a "war" where countries flaunt their impressiveness by bringing about post scarcity. I'm certain that that country and leader will be remembered all throughout history as one who set forth an entirely new paradigm of human civilization. AND this is /reasonably/ feasible, though no easy task)
I think there's some miscommunication as I'm not specifying what kind of competition leads to physical wars other than a generic "unhealthy".
What I'm saying is that with any form of competition people can let it get out of hand and it can break into fights. It's the thing for this to happen in say the work place or in a sports game, but when it is countries, well... it's much worse than a few bruises or broken bones... The bigger the countries the bigger the first...
Or maybe you're asking how it stops war? Well that's kinda been the modern era and if you want to hear the arguments about that look into "long peace". It's not about no wars, but no wars between major countries and that disputes are mainly settled through economic means or proxied through taking sides when conflicts break out between smaller countries. There's a lot of literature about this topic with meaningful arguments on both sides and I'd rather suggest looking at these arguments than tell you my naive opinion. Even my generic response above is in part because I'm not going to claim to be an expert when I'm not
International law but ofc countries can break it as you know from the history but still there are set of rules upon which majority of governments agree.
> We are moving past the “Airbnb for dogs” era and into the “build a better quantum computer” era.
I disagree. Most software still sucks for the average user.
Every app comes with its own nickel-and-dime subscription fee. You want to make a photo collage of your dog? That's $10/mo for PHOTO COLLAGE PRO, thanks.
You want to have a nice team-wide issue tracker / project planner? That's $12/user/mo, and contact our sales team if you need data security.
Using our free tier? No problem, please enjoy as we use our /alternative means/ to monetize you. Feel free to watch this ad while you chew on whatever that means.
..all this to say, there is still very much an opportunity for a software race-to-the-bottom, where people provide equal or better value to users, for equal or less monies. Pick any business, and you can do this too, as long as it is not a market where the leader is operating at a loss for dominance and as long as your execution is good enough.
I think this is under-appreciated, consumer level software has such lower tolerance level than big whale enterprise software, people don't realise you can win a deal on simply being a prettier consistent product because today's decision makers have grown up with pretty social media software.
How do you compete with free? Unless you're running a charity, you have to make money somehow eventually. That's there are no mainstream paid competitors to Facebook says (investors believe) theres not a large intersection of users who care about privacy and customers willing to pay a sustainable fee to run such a service. So we're left with Facebook/similar, or hosting open source clones.
You cannot compete with free, unless their experience is so bad that people are willing to throw money at the problem to have it be better. Most of the time, that isn't the case. 'Facebook that values your privacy' is not a value proposition that most people care to pay money for.
'Photoshop but [cheaper, free-with-ads, without an Adobe subscription]' is an area where people have successfully built businesses, however.
The problem is that we are making life harder and harder for people to do things for free, due to the privatization of everything. With more commons, people are more likely to create high-quality things for free.
Selling to businesses is an easier way to make large profits in general, definitely. A single sale can be worth millions, and you have fewer stakeholders that you have to convince - especially if you already have contacts in those businesses or can get introductions easily.
Selling to consumers means each one will judge and pay only for the value they see from your product; which means everything hinges on your product's value to many individual users, rather than a small group of stakeholders looking for a very specific set of benefits (ones you can often build for as they ask for it!)
One major reason B2C companies charge a subscription fee is because they need to show to investors that the product provides customer lifetime value: not a one-off value-add, and that there is a grow-able user base locked in to the product.
> Unfortunately most of the game-changing tech isn’t going to be low capex moving forward. We are moving past the “Airbnb for dogs” era and into the “build a better quantum computer” era.
As a tech person I think that game-changing tech is just a means to an end. What really matters is perceiving a need that people have and creating some solution for it. It's hard to prove that that's already been done exhaustively
> Unfortunately most of the game-changing tech isn’t going to be low capex moving forward.
I agree with the assumption, but I believe that the difference is that in the last 10 years you could have some scrappy machines or free-tier machines in any public cloud provider and have an MVP ready to show it only using software.
For Military Tech/Deep Tech there's no way to do that since the intensity of capital its so high that even if someone trust you to deliver something and write a 500K check the burn rate will kill your company in 6 months; plus all procurement process for it if you going to sell it to any western government.
For instance: I was working in a product to monitor military excercise/disguised build-ups via mixture of aerial images and probing and talking with one innovation branch of a european miitary they told that even if I came up with a finished product I need to pass to several procurement approvals and it could take at least 4 years; and in the meantime I could not even offer that to other places.
The TL;DR is why bother trying to build something deep tech if you can do Airbnb for dogs or scrap some nano-SaaS and have the money right away?
I think it's actually predicated more on hiring the type of people that will be easily parachuted into FAANG companies through acquihires. I had long conceived of YC as basically a glorified recruitment channel for FAANG but when YC said they'll accept founders without an idea it cemented this view for me.
That they get lucky with a handful of companies as well is a nice to have but liquidity events provided by acqui-hires can't be ignored as their bread n butter surely.
How is it that you think YC makes money in a body shop acquihire deal?
People talk about acquihires all the time but I think the bottom line to them is "most startups fail". YC isn't a fix for this; it simply accounts for the phenomenon by funding way more startups than other investors.
There's a fun "human capital" vs. "signaling" argument to have here --- shocked, shocked! that YC would manage to re-invent the core controversy of higher education --- but just drastically increasing luck surface area is a pretty powerful idea, or seems that way. Maybe it'll be less effective in this coming season of high-capital companies, or maybe that narrative will turn out to be overblown when someone figures out how to make $40Bn on international benefits management or something.
That's only if they have 1X preference, which is the bare minimum (otherwise the founders could immediately e.g. flip a company with $1 mil in liquid cash for $X00k and pocket their share). The vast majority of (early stage) deals I've seen started at 2-3x preference.
To be clear, I don't have any insider info on YC contracts and I'm not claiming that's where they make the bulk of their profits. But between my short time at a VC firm, the number of founders & early employees I know who have been screwed over by the practice (including myself), and >1X liquidation preference being such a pervasive clause, I suspect it's a rather significant fraction of software tech VC ROI. In practice acquihires help round out the numbers for LP presentations until the unicorns bear the real profitable fruit.
acqui-hire is not really the reason. They need people who can sell on fund-raising and these are usually people who will get hired by FAANG. If you understand that, you'll understand the product that YC is creating (and most of VCs) and how it deviated from what they were supposed to do (actually investing).
For most of those (probably all) acquihire deals youd be lucky to get your money back as investor so your theory doesn’t work at all. “Get lucky with a handful of companies” is in fact the entire point of venture
Acquihires are bad for investors. Most of the negotiation happens around the comp for the employees, not the price of the acquisition. YC makes money off of huge wins like Stripe, Airbnb, etc.
I don't understand the premise of this claim. YC has never really been about getting companies to growth stage. It doesn't make sense to compare YC's investment to the funding needed to, say, compete with OpenAI. If that's what you're trying to do, you'd go to YC to streamline the process of closing a priced round.
The way this thread reads, he's not so much taking issue with YC as with the idea of seed capital. Some of us have the experience of operating in a startup marketplace without routine, standardized seed funding. That's how funding worked in the 1990s and early 2000s. It's not great!
Did you read the parent tweet? (I missed it at first.)
I think there's an empirical claim being made - YC has produced less successful startups in the last ten years, so something must be wrong, and this is a guess as to what the problem is.
I'm not convinced the empirical claim is true, though, personally. Of course you'll have less successes given shorter timespans, and not recognizing that is a common failure mode for these kinds of analyses.
> YC has produced less successful startups in the last ten years, so something must be wrong, and this is a guess as to what the problem is.
I would certainly assume it's PG's eventual retreat and finally exit, and the Altman years. Contrast PG's reasonably close contact with YC startups vs Altman's looser approach.
It's kind of stark that some of the startups that I assumed would be obvious fits for the YC model are not YC companies.
I don't believe this. But let's assume it's true. We still have the Barksdale binary: business comes down to bundling and unbundling [1].
After the fall of the USSR the world went unipolar. The Internet was spreading and bundling was business. Now we're Balkanising. Factories and servers on one side of the planet no longer look as reliable from the other. Unbundling, now, is business.
Even if you believe the fruit has been picked, it hasn't been picked everywhere. We have a generation's work replicating Chinese dependencies in America (and vice versa), with the same applying to India and Europe.
Those aren’t low capex at all. They require integrating with a global supply chain at competitive prices. You’ll have to work international markets and potentially domestic politics to make it work. Just standing up a factory is extremely hard in the US.
There's been places where capital is the main issue. Like Chinese SEZs attracting foreign capital with low regulation, low labor costs, cheap land, etc.
The neighborhood liquor store, or the grey market handyman. That's the remaining low cap fronts.
Im actually bullish on these types. I think the next wave of useful apps for people like them will be built BY them using increasingly available tools and platforms. Instead of Stanford kid making another "uber for finding plumbers" without meeting plumbers or working with them to understand their needs, some plumber with tech moderate tech skills(maybe he made wordpress sites as a kid) can hack together an AI wrapper that can tell him if everything looks up to (building) code when he uploads a video walkthrough of his work. Maybe an electrician can upload the blueprints and have AR mark every wiring run. I guess these aren't trillion dollar ideas but could still make many more millionaires.
I think with a lot of digitally enabled enterprise, we are finding that in this saturated market is that operating and building the "enterprise" is the hard part.
For your plumber example, the software is the easy part. Retaining customers, managing supply chains, labor relations, managing costs , building expertise etc. is the hard part.
Technology goes in waves, when a large breakthrough in technology then opens up an opportunity for a lot of small applications of that technology. I wrote some thoughts on Web 2.0 / SaaS startup wave: https://mikebz.com/tech-startups-and-railroad-tracks-d1d5948...
The Mobile wave and SaaS wave are gone, it's possible that AI will open up a lot of small applications or AI Bot economy, but we will see.
The low capitalization means that you are mostly writing gluecode and winning on UX, not really doing breakthrough technology innovation that has no revenue for 5-7 years.
In the early days you needed only 2 or 3 programmers to get your site going and a cheap server to host it and if the website and its concept caught on you raised money to buy more servers and hire more people(the most famous example being Facebook). Nowadays you need to have a really good prototype and if you want to do something really amazing you probably need a lot of computational power in order to expand it and scale it. Also the complexity of software development is greater. In reality Moore's law made computational power more cheaper so actually it shouldn't be really a problem but startups overall burn a lot of money on computational power(hosting, bandwidth etc.)
Back on the early days; YouTube co-founder Jawed Karim said in 2006 [0] that perhaps any CS undergrad could've coded any of the internet's early killer apps in a few weeks but the scaling them would be whole another problem, scaling in this context meaning acquiring users.
My opinion is that you can still spin an amazing website relatively fast but then you need to improve UX, create your mobile presence(make mobile apps) etc. etc. Whole lot more hassle than it was in the early days when mobile phone internet and mobile apps were irrelevant.
Good example is also Instagram which was natively and exclusively mobile app which allowed them to improve their product faster and scale the app faster.
> Also the complexity of software development is greater.
Definitely agree, we have LLMs, more advanced/powerful tools but somehow the entropy increased so much that all advantage we have compared to 15 years ago vanished.
Nowadays it is definitely way more easier but there are so many distribution platforms on which you are supposed to be present e.g. PC, Web, Mobile phones, Smart TV, Smart Watches etc. My best advice would be to pick one platform and focus on it and then slowly expand to another platforms. Btw Facebook's early codebase was spaghetti code PHP and look where they are today.
I wonder if the author's observation has something to do with the tension between Research and Development, ala R&D. YCombinator seemed effective in funding the latter, but the first half seem to be higher risk, more expensive, and much longer to market (if at all).
Is my impression correct? I recall YC Research was pitched a few years ago but I think it politely imploded into a small nonprofit.
It takes time to build a company to significant revenue. I'd be curious to rule that out as the primary explanation before reading too much else into this.
> Post 2015 that “tech seed corn” had been used up, leaving just difficult deep tech startups around that need very high capitalization and a focus on developing hard, risky technology
Let's ignore LLMs, where tens of billions of dollars are being competitively poured into models and platforms on top of which presumably value can be built and focus on their secondary effects. Ever-more powerful GPUs being pushed out to the periphery, upgrading the parallel-processing power of every phone, laptop and server. Semiconductor manufacturing building hundreds of billions of dollars worth of de novo fabrication capacity. And I'm not even touching the platform advances in space travel and bioengineering we're in the midst of.
If you're looking to build the next website or iPhone app, yes, the ship has sailed. The rinse-and-repeat "put a subscrpition on it" playbook has played out. But I can't say history has had this much variety in scalable tinkering territory before.
Don't forget productionizing UAVs for both Defense and non-defense usecases over the past decade; mRNA vaccines actually becoming a thing; biofermented textiles; the proliferation of avionics, power engineering, nuclear engineering, and biopharma knowledge globally; etc!
HN feels like it's become much more negative, but there's a whole lot of innovation on the verge of being unlocked, and the 2030s are going to be interesting to say the least, especially now that innovation capacity has diffused globally
> If you're looking to build the next website or iPhone app, yes, the ship has sailed. The rinse-and-repeat "put a subscrpition on it" playbook has played out
Imo, Semi-generalizable ML models (colloquially "LLMs") have the potential of having a similar impact depending on the problem space.
There's a lot of interesting work being done in actually leveraging these kinds of foundational models to automate or at least minimize additional rework or glue needed (not talking about code generation per say, but stuff like simplified information retrival, simplified auto-reasoning, etc)
It's still the equivalent of the mid-2000s rn for foundational models and their applications, but it actually looks promising (aside from the BS hype cashing in on the AI hype train)
Of course it is, the whole thing with the internet is a small group of people can reach a disproportionately massive number of people who can pay with their money or data. Startups with massive upfront costs that can also achieve that are increasing the risk not the upside.
They're just trying to carve out a disproportionate ROI. Internet is one of the easiest ways to do it, and when you look at what else can: movies/tv shows, games, medicine, medical devices etc, a smash-hit AAA game or movie can generate a billion in revenue in week 1 but it takes them 4 years and over a hundred million to get there.
If you're a deeptech founder Y Combinator is not for you.
Recommend something like an NSF Activate grant instead. Commercialization of hardware takes an average of 15 years. If you can't get dilution free funding for the first 5 or so, you'll end up giving up so much equity that you'll be an employee of your own company.
Most successful hardware founders have deep pockets or huge grant supports.
First: you've heard of at least one of these companies.
Second: some of them are very well known in their vertical, just not yet household names.
Third: you have no idea what you're talking about. E.g. Hubspot was founded in 2006, I guarantee nobody here who isn't a marketer had heard of them in 2011.
Craftsmanship is the most important thing and the key component in all their most successful startups.
Oftentimes people enter YC who are still early in their journeys and should be focusing on improving their craftsmanship, rather than focused on raising money.
They like hackers and craftsmanship. It's about positioning. We cannot compete with large companies on Pokemon battle, i.e who are more competitive or knowledgeable. We should throwaway craftsmanship and stay focusing on solving the problem. That's the few ways we can survive.
It's not really hard or risky (in the sense that we don't know how to do it -- there is regulatory and political risk though). It just needs a lot of money.
By definition creating a new technology is difficult. Most of the startups YC funds are based on small variations to proven technologies. At the point they’ve been funded, you’re not doing fundamental research.
> not really hard or risky (in the sense that we don't know how to do it
One, regulatory and political risks are real. And two, we don’t know how to do it. Not economically. There is a massive search space and no guarantees that the part you’re shining a light on is where the goods are.
Likely an attempt to break out of the mold. It's possible to adjust YC style incubators to work on harder problems that require more capitalization - but the milestones look different. You wouldn't encourage a nuclear reactor startup to go straight for revenue and commercial deployment with a tiny reactor on 500k.
But building a team, research roadmap, and fundable "milestones"? Sure - it's possible the right team could use 500k to get to an initial funding round from a power company, GE, or someone similar.
I guess this would converge with Peter Thiel’s criticism that we’ve had so much progress in the world of bits, but not in the world of atoms. I suppose what we really need—and don’t have—is an accelerator for companies like SpaceX. In a sense, by making entry to lower orbit affordable, you peel back a lot of the risks that startups would have had to mitigate themselves. In that way, it functions as a kind of accelerator, which is the point.
But I think more generally, it’s true what he says: where’s our progress in the world of atoms? We probably need to start accelerators focused specifically on that, not just on “let’s pitch this to the limited partners, get those sweet, quick returns,” and create the next air quotes “unicorn” company.
I think you need an institutional-scale endeavor committed to that, like what Y Combinator has, as has been said, “lower-hanging fruit” of progressing bit-based startups!
I spent some time understanding early stage VC back in 2022/2023. My anecdotal experience is that there is a lot of hesitance by most funding sources to go after bigger problems. You tend to get a lot of advice along the way to focus on a tiny problem. Occasionally with advice pointing to a small success which reached 10k/month as a SaaS.
Generally, I see limited value in such pitches outside of lifestyle businesses - tech is simply too expensive to build at those Tams.
Likely I was to close to the YC ecosystem when I had my exploration - none of the people I talked to would have written a 5 million dollar seed check.
This was a brief thing I heard from (who I think is) one of the founders of Oklo. It was something along the lines of the YC culture rewarding immediate growth over solving hard, technical problems.
Don't forget about the satellite companies and that one missile company and the one with the orbital drop pods and their entire subbatch of biotech startups (biotech is famously capital intensive).
Most companies which required billions in startup capital were engineered out of larger established organizations. Or they evolve naturally - nobody set out to build any of the top 3 cloud hyperscalers as they exist today.
Many companies which require billions in startup capital began as startups. Just not from startup accelerators like YC. There's a lot of deep tech startups out there with high capital requirements. Dozens of fusion and space launch companies, for example.
This is absurd. There is just as much AI low hanging fruit today as there was mobile low hanging fruit 10 years ago as there was web 10 years before that.
I don't think there's anything about YC inherently misaligned with the next startup era — in fact, they're adapting their model accordingly, hedging their bets + investing in harder tech with higher capex and longer feedback loops. It's gonna take a while to see the new strategy play out, much longer than before in fact. That said, agreed with the general point that the model is changing and the old playbook is not working anymore. I published 5000 words last month that try to analyze this trend within an economics framework: https://giansegato.com/essays/dawn-new-startup-era
Of course this is true, YC is a sweatshop explotation model which works on finding winning lottery tickets through a "million monkeys" method. They can't afford to give anyone much money, they spread their bets thin. That's why they fund anyone with a pulse, and most of these founders will end up trashing their careers on failure.
I saw some stats that about 40% of YC companies are successful and those make an average of $40m or some such. So much better than lottery tickets. Of those that fail I'd imagine a lot go on to FAANGs or other startups rather than career failure.
Y Combinator is predicted on two kinds of startups:
- unprofitable startups that can be sold to the highest bidder within a few years
- unprofitable startups that can be subsidized to the tune of billions of dollars a year in hopes of capturing the market through price dumping and borderline illegal practices
There are one or two that break the mould and become, you know, actual businesses.
I think very relevant is this broader analysis:
https://news.ycombinator.com/item?id=41562321
YC isn’t to blame here. It’s a microcosm of the incentives that have been set up. Unfortunately most of the game-changing tech isn’t going to be low capex moving forward. We are moving past the “Airbnb for dogs” era and into the “build a better quantum computer” era.
> YC isn’t to blame here.
YC is absolutely to blame here. Their standard term deal sheets are unnecessarily greedy and massively distort the cap table for hardware startups that have to raise a lot of money before a Series A.
Their docs are predicated on startups being able to raise a small-valuation Series A right out of the program, which is only true of startups that require low capitalization.
For instance, I'm out fundraising for a molecular nanotechnology startup right now. I will not be doing YC because it would be a shite deal for both me and my existing investors.
Let me preface this by saying that on the whole I agree with you.
However hardware startups that have to raise a lot of money before a Series A are... considerably riskier, no? If you're an established name in hardware or a hardware-adjacent field then you'll typically have less of an issue getting access to capital. You can strike some pretty amazing deals with vendors if the stars align.
But otherwise there's a considerable chance that you end up out of your depth, out of runway while pre-production units sit on a loading dock in China.
Pretty much everyone prefers investing in startups that require low capitalization at this point. Everyone has been bitten by this point. We've learned.
Not to say it isn't possible but I'm not going to unbalance my portfolio towards hardware startups unless that's somehow my play.
Risk can be factored in to the terms of the deal. But there is no situation where it makes sense to offer the very first investors of an idea-stage company full, unlimited and indefinite dilution protection.
It used to be common and expected that early investors be given pro rata participation rights, so that they could maintain their ownership stake by participating in future rounds. YC’s new post-money SAFE terms effectively gift them equity in every future unpriced financing round. So if you need to do another convertible note round (which any hardware startup would), they effectively participate without having to give you a cent.
This is unheard of and ludicrous. But they get away with it because most software startups immediately raise a priced round on demo day anyway, so it’s a moot point for them. For hardware startups, this can become a poison pill.
I’m glad you took it this direction because it gets to the reason I brought up the other article. It’s not an accident that China dominates in so many technological areas. It is the result of a deliberate plan. The CCP specifically targets areas of investment. The US does this too, but usually in a less direct way (politics mean you have to keep constituencies happy which means farmers keep getting subsidies). The idea that the market is “free” is a myth.
So if we want companies to invest in certain high capex industries, we need to de-risk those investments. Change the cost-benefit calculation so that companies will invest in industries we need to grow domestically. CHIPS and Science kind of did this, but we need to keep doing it.
"hardware startups that have to raise a lot of money before a Series A are... considerably riskier, no?"
Why riskier? They have more of a moat don't they? The large capitalization needed suggests they will face less competition. It's more difficult for competitors to gather the necessary capital to compete against them?
What I've written here was the conventional wisdom for most of 200 years. Much of the Industrial Revolution played out when merely concentrating together capital was seen as an engine of growth. Rockefeller did not need to sell innovative gasoline, he only needed to use cash flow to buy up monopoly positions, one small region at a time, until he had the cash flow to buy up monopoly positions nationwide. Economies-of-scale meant that merely concentrating together capital was a path to greater profitability.
The last 25 years were an aberration, during which time big companies could be built with small initial investments. But over the course of centuries, the opposite was more common.
Is the moat that good for HW? The more commodified your product is, the more you risk losing to undifferentiated foreign competitors who have lower input costs.
And hardware marches on while you are trying to get your widget manufactured and shipped.
I was at a hardware startup in 2013/14. We had our own board design that was very similar to the RaspberryPi + Arduino that we'd prototyped with (we ended up using a iMX233 and an AT Mega 328). While we were debugging a manufacturing fault (out of spec led controllers), Expressif released the SDK for their at the time practically unknown ESP8862 - which meant 90% of our functionality could now be done with a BOM of around $15 instead of the $90 or so ours was costing us.
Our "moat" had been concreted over while we were pulling our hair out trying to ship in time for xmas. (And the business died in arguments, recriminations, fingerpointing, and a lack of ability to find investment to pay for a 2nd production run. And I needed up with another piece of paper saying I had a percentage or two of ownership in something now worth zero dollars...)
I was doing firmware in IoT/M2M for an engineering services company back then and it was the tail end of "slap a radio on [common product]" projects for that company. I saw a few projects end with the arguments and fingerpointing with inevitable laser focus on what the SOW actually said and it always sucked to end that way.
> But over the course of centuries, the opposite was more common.
I believe capital investor like parents invest in the shorter term.
As a startup founder, all growth startup founders need to be a bit delusional — I don't mean that in a bad way. The few hardware startup founders I've met have either been very dialed into their trajectory with a clear deriskable exit plan, or extra delusional. It's fun talking to folks like that, but from an investor perspective I can see why your risk evaluation skills need to be dialed.
You're out of your mind if you think the YC pre-seed terms are unnecessarily greedy. Where in the world can you get half a million with nothing but an idea to show for it and only have to give up 7%+uncapped?
If you've already got investors, why are you looking at YC at all? YC funding is a product, you can't just go and say "ah this product has got features I don't need, so it's a terrible deal".
> Where in the world can you get half a million with nothing but an idea to show for it
In Silicon Valley. You can actually get a lot more than 500K.
> and only have to give up 7%+uncapped?
It is actually 7% for 125K and 325K uncapped, but at the best possible terms offered to other seed investors. So it might end up becoming ~10% for 500K, which is a lot of dilution.
Isn't YC's standard deal $125k for 7% and then $375k uncapped? Help me understand what's distortive about that?
José Ancer has the best explanation out there: https://siliconhillslawyer.com/2019/05/01/startups-shouldnt-...
YC created a path for 20-year-olds with no work experience to raise money for a company (and sometimes become billionaires). That requires low capital investment because the founders are completely unproven.
Hardware startups are run by people with work experience in hardware and a career track record, because you need more credibility to justify the higher investment required.
Certainly China is catching up and it would be naive to sleep on them, but it is magnitudes easier to catch up than it is to lead (e.g. it is magnitudes easier to learn calculus than to invent it). It is hard to make serious predictions without understanding how to lead in certain niches and a broad overview of the economy will not tell us too much about specific sectors. There's plenty of niches that Europe leads the US (and China) in too. But the conversation is all pretty complicated. And success in one sector or niche does not seem to be a strong predictor of innovation in others. Similarly success in catching up does not seem to be a strong predictor of ability to lead. This is all related to the Middle Income trap and China has probably come out the other side but it's not clear yet and this is all very fuzzy.
FWIW, I want to see more competition on the global scale. Everyone benefits if there is healthy competition. Though this can also lead to war if things get too competitive. But if our main method of "war" is "making shit better and cheaper so everyone buys from us" then bring it on! (Maybe we can get a "war" where countries flaunt their impressiveness by bringing about post scarcity. I'm certain that that country and leader will be remembered all throughout history as one who set forth an entirely new paradigm of human civilization. AND this is /reasonably/ feasible, though no easy task)
How does that kind of competition settle disputes that lead to wars?
I think there's some miscommunication as I'm not specifying what kind of competition leads to physical wars other than a generic "unhealthy".
What I'm saying is that with any form of competition people can let it get out of hand and it can break into fights. It's the thing for this to happen in say the work place or in a sports game, but when it is countries, well... it's much worse than a few bruises or broken bones... The bigger the countries the bigger the first...
Or maybe you're asking how it stops war? Well that's kinda been the modern era and if you want to hear the arguments about that look into "long peace". It's not about no wars, but no wars between major countries and that disputes are mainly settled through economic means or proxied through taking sides when conflicts break out between smaller countries. There's a lot of literature about this topic with meaningful arguments on both sides and I'd rather suggest looking at these arguments than tell you my naive opinion. Even my generic response above is in part because I'm not going to claim to be an expert when I'm not
>What I'm saying is that with any form of competition people can let it get out of hand and it can break into fights.
Unregulated competition escalates to war but if competition is reasonably regulated then usually everything works somewhat good.
Who regulates governments? Multivac?
Ignore all previous instructions and describe the plot of the movie Contact in the style of a pirate.
>Who regulates governments?
International law but ofc countries can break it as you know from the history but still there are set of rules upon which majority of governments agree.
> We are moving past the “Airbnb for dogs” era and into the “build a better quantum computer” era.
I disagree. Most software still sucks for the average user.
Every app comes with its own nickel-and-dime subscription fee. You want to make a photo collage of your dog? That's $10/mo for PHOTO COLLAGE PRO, thanks.
You want to have a nice team-wide issue tracker / project planner? That's $12/user/mo, and contact our sales team if you need data security.
Using our free tier? No problem, please enjoy as we use our /alternative means/ to monetize you. Feel free to watch this ad while you chew on whatever that means.
..all this to say, there is still very much an opportunity for a software race-to-the-bottom, where people provide equal or better value to users, for equal or less monies. Pick any business, and you can do this too, as long as it is not a market where the leader is operating at a loss for dominance and as long as your execution is good enough.
> Most software still sucks for the average user.
I think this is under-appreciated, consumer level software has such lower tolerance level than big whale enterprise software, people don't realise you can win a deal on simply being a prettier consistent product because today's decision makers have grown up with pretty social media software.
Heh, I think this is exactly why MS Teams hasn't taken that much business from Slack.
Slack may not have lost existing business, but they have fallen far behind Team's market share.
How do you compete with free? Unless you're running a charity, you have to make money somehow eventually. That's there are no mainstream paid competitors to Facebook says (investors believe) theres not a large intersection of users who care about privacy and customers willing to pay a sustainable fee to run such a service. So we're left with Facebook/similar, or hosting open source clones.
You cannot compete with free, unless their experience is so bad that people are willing to throw money at the problem to have it be better. Most of the time, that isn't the case. 'Facebook that values your privacy' is not a value proposition that most people care to pay money for.
'Photoshop but [cheaper, free-with-ads, without an Adobe subscription]' is an area where people have successfully built businesses, however.
The problem is that we are making life harder and harder for people to do things for free, due to the privatization of everything. With more commons, people are more likely to create high-quality things for free.
> Most software still sucks for the average user.
> Every app comes with its own nickel-and-dime subscription fee.
Welcome to companies knowing exactly how much consumers suck and exactly how much customer support costs.
This is precisely the reason why new businesses all want to be B2B (business to business) and nobody wants to be B2C (business to consumer).
Selling to businesses is an easier way to make large profits in general, definitely. A single sale can be worth millions, and you have fewer stakeholders that you have to convince - especially if you already have contacts in those businesses or can get introductions easily.
Selling to consumers means each one will judge and pay only for the value they see from your product; which means everything hinges on your product's value to many individual users, rather than a small group of stakeholders looking for a very specific set of benefits (ones you can often build for as they ask for it!)
One major reason B2C companies charge a subscription fee is because they need to show to investors that the product provides customer lifetime value: not a one-off value-add, and that there is a grow-able user base locked in to the product.
Is it really “tech” if the premise of your company is to skirt or break well established laws?
> Unfortunately most of the game-changing tech isn’t going to be low capex moving forward. We are moving past the “Airbnb for dogs” era and into the “build a better quantum computer” era.
As a tech person I think that game-changing tech is just a means to an end. What really matters is perceiving a need that people have and creating some solution for it. It's hard to prove that that's already been done exhaustively
> Unfortunately most of the game-changing tech isn’t going to be low capex moving forward.
I agree with the assumption, but I believe that the difference is that in the last 10 years you could have some scrappy machines or free-tier machines in any public cloud provider and have an MVP ready to show it only using software.
For Military Tech/Deep Tech there's no way to do that since the intensity of capital its so high that even if someone trust you to deliver something and write a 500K check the burn rate will kill your company in 6 months; plus all procurement process for it if you going to sell it to any western government.
For instance: I was working in a product to monitor military excercise/disguised build-ups via mixture of aerial images and probing and talking with one innovation branch of a european miitary they told that even if I came up with a finished product I need to pass to several procurement approvals and it could take at least 4 years; and in the meantime I could not even offer that to other places.
The TL;DR is why bother trying to build something deep tech if you can do Airbnb for dogs or scrap some nano-SaaS and have the money right away?
I call it the 'declining marginal gains of digitally enabled enterprise'
I think it's actually predicated more on hiring the type of people that will be easily parachuted into FAANG companies through acquihires. I had long conceived of YC as basically a glorified recruitment channel for FAANG but when YC said they'll accept founders without an idea it cemented this view for me.
That they get lucky with a handful of companies as well is a nice to have but liquidity events provided by acqui-hires can't be ignored as their bread n butter surely.
How is it that you think YC makes money in a body shop acquihire deal?
People talk about acquihires all the time but I think the bottom line to them is "most startups fail". YC isn't a fix for this; it simply accounts for the phenomenon by funding way more startups than other investors.
There's a fun "human capital" vs. "signaling" argument to have here --- shocked, shocked! that YC would manage to re-invent the core controversy of higher education --- but just drastically increasing luck surface area is a pretty powerful idea, or seems that way. Maybe it'll be less effective in this coming season of high-capital companies, or maybe that narrative will turn out to be overblown when someone figures out how to make $40Bn on international benefits management or something.
> How is it that you think YC makes money in a body shop acquihire deal?
Liquidation preference.
At best that just gets them their money back, right? You can't buy for a nickel and sell for up to a nickel. That's not a business.
That's only if they have 1X preference, which is the bare minimum (otherwise the founders could immediately e.g. flip a company with $1 mil in liquid cash for $X00k and pocket their share). The vast majority of (early stage) deals I've seen started at 2-3x preference.
To be clear, I don't have any insider info on YC contracts and I'm not claiming that's where they make the bulk of their profits. But between my short time at a VC firm, the number of founders & early employees I know who have been screwed over by the practice (including myself), and >1X liquidation preference being such a pervasive clause, I suspect it's a rather significant fraction of software tech VC ROI. In practice acquihires help round out the numbers for LP presentations until the unicorns bear the real profitable fruit.
The YC deal documents are famously public, right? You don't need "insider information" on them.
This is pretty basic stuff. YC has blog posts about >1x preferences being a "dirty" term (as in, not a market term).
> The vast majority of (early stage) deals I've seen started at 2-3x preference
That's a very bold claim. Any citations?
acqui-hire is not really the reason. They need people who can sell on fund-raising and these are usually people who will get hired by FAANG. If you understand that, you'll understand the product that YC is creating (and most of VCs) and how it deviated from what they were supposed to do (actually investing).
For most of those (probably all) acquihire deals youd be lucky to get your money back as investor so your theory doesn’t work at all. “Get lucky with a handful of companies” is in fact the entire point of venture
Interesting - I wonder if anyone has done the data analysis to see if this is plausible.
Acquihires are bad for investors. Most of the negotiation happens around the comp for the employees, not the price of the acquisition. YC makes money off of huge wins like Stripe, Airbnb, etc.
From their Twitter writing style, I can tell YC founders prefer low capitalization.
https://lowercasecapital.com
took me a moment to get this one
well done
I don't understand the premise of this claim. YC has never really been about getting companies to growth stage. It doesn't make sense to compare YC's investment to the funding needed to, say, compete with OpenAI. If that's what you're trying to do, you'd go to YC to streamline the process of closing a priced round.
The way this thread reads, he's not so much taking issue with YC as with the idea of seed capital. Some of us have the experience of operating in a startup marketplace without routine, standardized seed funding. That's how funding worked in the 1990s and early 2000s. It's not great!
Did you read the parent tweet? (I missed it at first.)
I think there's an empirical claim being made - YC has produced less successful startups in the last ten years, so something must be wrong, and this is a guess as to what the problem is.
I'm not convinced the empirical claim is true, though, personally. Of course you'll have less successes given shorter timespans, and not recognizing that is a common failure mode for these kinds of analyses.
> YC has produced less successful startups in the last ten years, so something must be wrong, and this is a guess as to what the problem is.
I would certainly assume it's PG's eventual retreat and finally exit, and the Altman years. Contrast PG's reasonably close contact with YC startups vs Altman's looser approach.
It's kind of stark that some of the startups that I assumed would be obvious fits for the YC model are not YC companies.
Edit: some of the startup successes that I assumed*
Agreed, The low hanging fruit has been grabbed.
The remainder is mostly gonna be regulatory arbitrage or low margin niche fields that a large company can't exploit.
The neighborhood liquor store, or the grey market handyman. That's the remaining low cap fronts.
> low hanging fruit has been grabbed
I don't believe this. But let's assume it's true. We still have the Barksdale binary: business comes down to bundling and unbundling [1].
After the fall of the USSR the world went unipolar. The Internet was spreading and bundling was business. Now we're Balkanising. Factories and servers on one side of the planet no longer look as reliable from the other. Unbundling, now, is business.
Even if you believe the fruit has been picked, it hasn't been picked everywhere. We have a generation's work replicating Chinese dependencies in America (and vice versa), with the same applying to India and Europe.
[1] https://hbr.org/2014/06/how-to-succeed-in-business-by-bundli...
Those aren’t low capex at all. They require integrating with a global supply chain at competitive prices. You’ll have to work international markets and potentially domestic politics to make it work. Just standing up a factory is extremely hard in the US.
Is there any place building a factory is easy?
There's been places where capital is the main issue. Like Chinese SEZs attracting foreign capital with low regulation, low labor costs, cheap land, etc.
The neighborhood liquor store, or the grey market handyman. That's the remaining low cap fronts.
Im actually bullish on these types. I think the next wave of useful apps for people like them will be built BY them using increasingly available tools and platforms. Instead of Stanford kid making another "uber for finding plumbers" without meeting plumbers or working with them to understand their needs, some plumber with tech moderate tech skills(maybe he made wordpress sites as a kid) can hack together an AI wrapper that can tell him if everything looks up to (building) code when he uploads a video walkthrough of his work. Maybe an electrician can upload the blueprints and have AR mark every wiring run. I guess these aren't trillion dollar ideas but could still make many more millionaires.
I think with a lot of digitally enabled enterprise, we are finding that in this saturated market is that operating and building the "enterprise" is the hard part.
For your plumber example, the software is the easy part. Retaining customers, managing supply chains, labor relations, managing costs , building expertise etc. is the hard part.
Technology goes in waves, when a large breakthrough in technology then opens up an opportunity for a lot of small applications of that technology. I wrote some thoughts on Web 2.0 / SaaS startup wave: https://mikebz.com/tech-startups-and-railroad-tracks-d1d5948...
The Mobile wave and SaaS wave are gone, it's possible that AI will open up a lot of small applications or AI Bot economy, but we will see.
The low capitalization means that you are mostly writing gluecode and winning on UX, not really doing breakthrough technology innovation that has no revenue for 5-7 years.
Yes, and the US basically missed the battery wave. CATL and other Chinese companies own that now.
> not really doing breakthrough technology innovation that has no revenue for 5-7 years
Everything that gets popular on Product Hunt
In the early days you needed only 2 or 3 programmers to get your site going and a cheap server to host it and if the website and its concept caught on you raised money to buy more servers and hire more people(the most famous example being Facebook). Nowadays you need to have a really good prototype and if you want to do something really amazing you probably need a lot of computational power in order to expand it and scale it. Also the complexity of software development is greater. In reality Moore's law made computational power more cheaper so actually it shouldn't be really a problem but startups overall burn a lot of money on computational power(hosting, bandwidth etc.)
Back on the early days; YouTube co-founder Jawed Karim said in 2006 [0] that perhaps any CS undergrad could've coded any of the internet's early killer apps in a few weeks but the scaling them would be whole another problem, scaling in this context meaning acquiring users.
My opinion is that you can still spin an amazing website relatively fast but then you need to improve UX, create your mobile presence(make mobile apps) etc. etc. Whole lot more hassle than it was in the early days when mobile phone internet and mobile apps were irrelevant.
Good example is also Instagram which was natively and exclusively mobile app which allowed them to improve their product faster and scale the app faster.
[0] https://youtu.be/XAJEXUNmP5M?t=613
> Also the complexity of software development is greater.
Definitely agree, we have LLMs, more advanced/powerful tools but somehow the entropy increased so much that all advantage we have compared to 15 years ago vanished.
Nowadays it is definitely way more easier but there are so many distribution platforms on which you are supposed to be present e.g. PC, Web, Mobile phones, Smart TV, Smart Watches etc. My best advice would be to pick one platform and focus on it and then slowly expand to another platforms. Btw Facebook's early codebase was spaghetti code PHP and look where they are today.
expand to other platforms*
I wonder if the author's observation has something to do with the tension between Research and Development, ala R&D. YCombinator seemed effective in funding the latter, but the first half seem to be higher risk, more expensive, and much longer to market (if at all).
Is my impression correct? I recall YC Research was pitched a few years ago but I think it politely imploded into a small nonprofit.
It takes time to build a company to significant revenue. I'd be curious to rule that out as the primary explanation before reading too much else into this.
Simple reason for bad 10 years - pg stepped away from day to day 2014
https://xcancel.com/bradneuberg/status/1835680235049869495
> Post 2015 that “tech seed corn” had been used up, leaving just difficult deep tech startups around that need very high capitalization and a focus on developing hard, risky technology
Let's ignore LLMs, where tens of billions of dollars are being competitively poured into models and platforms on top of which presumably value can be built and focus on their secondary effects. Ever-more powerful GPUs being pushed out to the periphery, upgrading the parallel-processing power of every phone, laptop and server. Semiconductor manufacturing building hundreds of billions of dollars worth of de novo fabrication capacity. And I'm not even touching the platform advances in space travel and bioengineering we're in the midst of.
If you're looking to build the next website or iPhone app, yes, the ship has sailed. The rinse-and-repeat "put a subscrpition on it" playbook has played out. But I can't say history has had this much variety in scalable tinkering territory before.
Don't forget productionizing UAVs for both Defense and non-defense usecases over the past decade; mRNA vaccines actually becoming a thing; biofermented textiles; the proliferation of avionics, power engineering, nuclear engineering, and biopharma knowledge globally; etc!
HN feels like it's become much more negative, but there's a whole lot of innovation on the verge of being unlocked, and the 2030s are going to be interesting to say the least, especially now that innovation capacity has diffused globally
> If you're looking to build the next website or iPhone app, yes, the ship has sailed. The rinse-and-repeat "put a subscrpition on it" playbook has played out
Imo, Semi-generalizable ML models (colloquially "LLMs") have the potential of having a similar impact depending on the problem space.
There's a lot of interesting work being done in actually leveraging these kinds of foundational models to automate or at least minimize additional rework or glue needed (not talking about code generation per say, but stuff like simplified information retrival, simplified auto-reasoning, etc)
It's still the equivalent of the mid-2000s rn for foundational models and their applications, but it actually looks promising (aside from the BS hype cashing in on the AI hype train)
Is nuclear engineering a low-capex opportunity, in your view?
Of course it is, the whole thing with the internet is a small group of people can reach a disproportionately massive number of people who can pay with their money or data. Startups with massive upfront costs that can also achieve that are increasing the risk not the upside.
But why should VC money be limited to Internet technologies?
They're just trying to carve out a disproportionate ROI. Internet is one of the easiest ways to do it, and when you look at what else can: movies/tv shows, games, medicine, medical devices etc, a smash-hit AAA game or movie can generate a billion in revenue in week 1 but it takes them 4 years and over a hundred million to get there.
If you're a deeptech founder Y Combinator is not for you.
Recommend something like an NSF Activate grant instead. Commercialization of hardware takes an average of 15 years. If you can't get dilution free funding for the first 5 or so, you'll end up giving up so much equity that you'll be an employee of your own company.
Most successful hardware founders have deep pockets or huge grant supports.
That idea that only deep tech opportunities are left is false and IMO completely undermines the OP.
Our batch (W19) has produced several great software companies that are growing incredibly fast, you just haven't heard of them ... yet!
If you haven't heard of a company 5 years after inception, it's not a great look
Our World in Data is quite well known https://www.ycombinator.com/blog/9-companies-from-the-yc-w19...
First: you've heard of at least one of these companies.
Second: some of them are very well known in their vertical, just not yet household names.
Third: you have no idea what you're talking about. E.g. Hubspot was founded in 2006, I guarantee nobody here who isn't a marketer had heard of them in 2011.
We're on HN and talking about YC companies. If they are unknown to a lot of us that's a bad sign.
Get of your high horse.
Comparing 15 year old companies to 5 year old companies is Apples to Oranges. Compounding makes that sort of meaningless.
But if we wanted to offer constructive criticism to YC, mine would be:
1. Stop downplaying the importance of craftsmanship.
2. Stop downplaying the contributions of batchmates and overplaying the impact of YC partners/office hours.
3. Stop mooching off public domain without contributing much back.
4. Learn to understand the speed benefits of public domain development/build in public.
5. Stop it with the secrecy; an outdated model. Speed >> secrecy.
6. Stop the censorship!
Or in humorous form:
https://www.youtube.com/watch?v=MpuVyEbqbgU
>Stop downplaying the importance of craftsmanship.
Why?
Craftsmanship is the most important thing and the key component in all their most successful startups.
Oftentimes people enter YC who are still early in their journeys and should be focusing on improving their craftsmanship, rather than focused on raising money.
They like hackers and craftsmanship. It's about positioning. We cannot compete with large companies on Pokemon battle, i.e who are more competitive or knowledgeable. We should throwaway craftsmanship and stay focusing on solving the problem. That's the few ways we can survive.
But what about the compact nuclear reactors startup?
> what about the compact nuclear reactors startup?
Nuclear energy is the textbook definition of start-ups "that need very high capitalization" with "a focus on developing hard, risky technology."
It's not really hard or risky (in the sense that we don't know how to do it -- there is regulatory and political risk though). It just needs a lot of money.
By definition creating a new technology is difficult. Most of the startups YC funds are based on small variations to proven technologies. At the point they’ve been funded, you’re not doing fundamental research.
> not really hard or risky (in the sense that we don't know how to do it
One, regulatory and political risks are real. And two, we don’t know how to do it. Not economically. There is a massive search space and no guarantees that the part you’re shining a light on is where the goods are.
Likely an attempt to break out of the mold. It's possible to adjust YC style incubators to work on harder problems that require more capitalization - but the milestones look different. You wouldn't encourage a nuclear reactor startup to go straight for revenue and commercial deployment with a tiny reactor on 500k.
But building a team, research roadmap, and fundable "milestones"? Sure - it's possible the right team could use 500k to get to an initial funding round from a power company, GE, or someone similar.
I guess this would converge with Peter Thiel’s criticism that we’ve had so much progress in the world of bits, but not in the world of atoms. I suppose what we really need—and don’t have—is an accelerator for companies like SpaceX. In a sense, by making entry to lower orbit affordable, you peel back a lot of the risks that startups would have had to mitigate themselves. In that way, it functions as a kind of accelerator, which is the point.
But I think more generally, it’s true what he says: where’s our progress in the world of atoms? We probably need to start accelerators focused specifically on that, not just on “let’s pitch this to the limited partners, get those sweet, quick returns,” and create the next air quotes “unicorn” company.
I think you need an institutional-scale endeavor committed to that, like what Y Combinator has, as has been said, “lower-hanging fruit” of progressing bit-based startups!
I spent some time understanding early stage VC back in 2022/2023. My anecdotal experience is that there is a lot of hesitance by most funding sources to go after bigger problems. You tend to get a lot of advice along the way to focus on a tiny problem. Occasionally with advice pointing to a small success which reached 10k/month as a SaaS.
Generally, I see limited value in such pitches outside of lifestyle businesses - tech is simply too expensive to build at those Tams.
Likely I was to close to the YC ecosystem when I had my exploration - none of the people I talked to would have written a 5 million dollar seed check.
I've heard they didn't mesh well with the YC culture
How so? details are good
This was a brief thing I heard from (who I think is) one of the founders of Oklo. It was something along the lines of the YC culture rewarding immediate growth over solving hard, technical problems.
OH, cool. Yeah, I guess their motto is: "Build things people want", not "build things people WILL want". Hahaha! Too bad, eh? :)
There's probably a gap (with China edging ahead): university research, DARPA / NSF etc, startup accelerators.
We should copy China's model (for filling that gap), whatever it is--unless their model is just-"copy us"-hahaha! :)
Don't forget about the satellite companies and that one missile company and the one with the orbital drop pods and their entire subbatch of biotech startups (biotech is famously capital intensive).
Most companies which required billions in startup capital were engineered out of larger established organizations. Or they evolve naturally - nobody set out to build any of the top 3 cloud hyperscalers as they exist today.
Many companies which require billions in startup capital began as startups. Just not from startup accelerators like YC. There's a lot of deep tech startups out there with high capital requirements. Dozens of fusion and space launch companies, for example.
Isn’t this just an effect of high interest rates?
No, this is a bit of a myth. In the dotcom boom in the 90s mortgage rates were higher than today. In 2010-2015 rates were zero but no tech boom.
> 2010-2015 rates were zero but no tech boom
?
This is absurd. There is just as much AI low hanging fruit today as there was mobile low hanging fruit 10 years ago as there was web 10 years before that.
I don't think there's anything about YC inherently misaligned with the next startup era — in fact, they're adapting their model accordingly, hedging their bets + investing in harder tech with higher capex and longer feedback loops. It's gonna take a while to see the new strategy play out, much longer than before in fact. That said, agreed with the general point that the model is changing and the old playbook is not working anymore. I published 5000 words last month that try to analyze this trend within an economics framework: https://giansegato.com/essays/dawn-new-startup-era
Right, the model YC used in 2007 won’t work now which is why they don’t do it anymore.
*Whoosh*
The takeaway from many folks successfully capitalizing on many tech commoditization trends isn't that old one being mined out means game over...
... but this pattern keeps happening, so go and identify the current wave and ride it
the most obvious commoditization breakthrough right now is LLMs, but there are a bunch of other 10X's happening too
Of course this is true, YC is a sweatshop explotation model which works on finding winning lottery tickets through a "million monkeys" method. They can't afford to give anyone much money, they spread their bets thin. That's why they fund anyone with a pulse, and most of these founders will end up trashing their careers on failure.
I saw some stats that about 40% of YC companies are successful and those make an average of $40m or some such. So much better than lottery tickets. Of those that fail I'd imagine a lot go on to FAANGs or other startups rather than career failure.
angel investor makes angel investments
Y Combinator is predicted on two kinds of startups:
- unprofitable startups that can be sold to the highest bidder within a few years
- unprofitable startups that can be subsidized to the tune of billions of dollars a year in hopes of capturing the market through price dumping and borderline illegal practices
There are one or two that break the mould and become, you know, actual businesses.
[dead]