It's not just AI startups. It's almost all of them. The startup world is just a massive pump and dump scheme, and so it moves from field to field finding the next hotness. Self-driving cars, VR, EVs, AI, and so on. People soon realize that you can't just take a bunch of hipster tech employees and suddenly create non-solutions that need to have problems created for them to even be solutions, much less choosing extremely hard and ill-defined non-solutions.
I truly wonder how much of the U.S. economy has slowed down because of all this wasted effort. Is it really a good idea for every company to try and reinvent what a company is? Every startup I've been spends a non-zero amount on just re-figuring out how to be organized, how to manage people and projects, and even just figuring out the logistics and tech needed to run a company. It's a lot of wasted effort, time, and money. My time at startups has been the least effective portions of my career. Real work actually moves quite slowly at startups, in my experience.
> startup world is just a massive pump and dump scheme, and so it moves from field to field finding the next hotness
But we retain the solutions that work. A slime mold tries a bunch of useless paths in a maze before it finds the right one. Then it trims and reinforces. Most start-ups failing isn't a sure sign of wasted work.
Yes, many business ideas I see are of the type "That would be a great idea if you could make it work, but given how much easier it is to fool people into thinking you have made it work, than to actually make it work, the odds aren't good".
And if you're a rich investor willing to bet on a lot of such things hoping to make it back on of the few that work, well, you just made that ratio even more lopsided.
> but given how much easier it is to fool people into thinking you have made it work, than to actually make it work
One thing that has often surprised me about HN is how people routinely upvote and boast about how they faked a demo. Idk why we refuse to call that what it is: a lie. IDK if Jobs did it, it is still a lie. It's one thing to say that this is what we imagine a technology will be like, but to call it a demo and act like it is the actual tool and that its capabilities generalize to the tasks being shown is theater. Are tech bros just theater kids?
The best way to do a completely unbiased focus group is to come up with an MVP and try to sell it to them. There's no more honest endorsement than someone handing you money to solve their problem.
If they buy, you're on the right track and are ready to scale up to a larger focus group.
Yes, what most people don't realize though is that the MVP doesn't have to be a functional product. It can be a cartoon, a drawing, a mock up, a video, whatever.
> It's exactly the opposite, they're essential and certainly do not have a "mixed history"
Focus groups are good for refining. They're absolutely terrible for new product development. If there is an OG playbook for start-ups defeating incumbents, it's in exploiting the committee-based designed process that large companies tend towards. (Owning a product decision is hard. Outsourcing it to a marketing consultant isn't.)
Where you and I probably agree is in the importance of finding the shortest path to a sale.
> It's called user focused design
You're quoting this guy [1]. He also did marketing consulting [2].
It's not really that way anymore. The vision of a loan visionary coming up with a new product or invention like being hit by a lightning bolt and then getting rich by building a factory that builds it is just a fairy tale.
In reality it looks more like a series of focus groups. They may not be focus groups per se, but showing your idea to potential customers, getting feedback, refining and showing again is how it's done.
> may not be focus groups per se, but showing your idea to potential customers, getting feedback, refining and showing again is how it's done
Sure. I’d argue both focus groups and the lone visionary are myths. But in order to show an idea properly, you have to start thinking about building it. And to start thinking about building it properly, you have to start building it. For every Juicero there are teams iterating around a local optimum.
> Most start-ups failing isn't a sure sign of wasted work.
It pretty much is, yes. Not a sure sign, but SV startups are probably the least efficient way of finding solutions that work. That's not to say that solutions to meaningful problems aren't occasionally found, of course, but the odds are very slim and the amount of money and other resources burned per success is astounding.
> SV startups are probably the least efficient way of finding solutions that work
On what metric? By any objective measure, barring the tail end of ZIRP, it’s been remarkably more efficient than many national projects.
The point is a VC ecosystem working properly should have a high failure rate. Lots of failures just tells you risk isn’t too low. To measure efficacy and efficiency you have to look at the output. And again, barring the pandemic vintages, it’s a pretty clear record of success.
This will be tough to answer because you've supplied a vague standard. According to the Unabomber, electricity is not a real solution. Where shall we sit on the spectrum?
Ceph? (Inktank was founded in 2011 by Sage Weil and DreamHost, which then it raised 1M in 2012 from Mark Shuttleworth, then RedHat bought them for 175M in 2014. Seems like a nice startup stort, no? The project itself was started around 2004 by Weil at LLNL, and then he worked on it as part of his PhD still in 2007, then the work continued at DreamHost, probably because Weil is one of the co-founders of DreamHost.)
> What startups have provided lasting real solutions and value, not just capitalistic, in say the last decade
Deel, Rippling, Redwood Materials, Convoy, Airtable, Anduril, Cerebral, Checkr and Cerebras were all founded since 2014. Pick one. We're currently seeing an explosion in creativity around satellites, autonomous vehicles (on the road, indoors and airborne), generative AI and biotechnology.
Go back one more cycle and we have SpaceX, Tesla, Uber & Lyft, Moderna, Zoom, DoorDash and other giants that entered busy fields, survived the winnowing and basically created a new industry.
Deel provides a solution to workers having too much time and patience, and the value of giving nice chunk of each salary for the pleasure of making payroll suck with some AI smeared on top.
They didn't innovate there! Safeguard Global and plenty of others were doing the same middlemanning for years before. Just because they interact with employment doesn't mean they're as important as the whole concept.
> Safeguard Global and plenty of others were doing the same middlemanning for years before
This is sort of like saying the iPod was a better Walkman. Yes, you could functionally do the same thing. But no, it wasn't as easy which meant it wasn't as frequently done. (And as a result of pressure from the upstarts, payroll across the board became smoother.)
> Just because they interact with employment doesn't mean they're as important as the whole concept
Nobody said that. I said "their impact is meaningful to the degree employment is." To the people whose roles exist on account of platforms like Deel, their existence is meaningful.
I was asked for a list of startups that “have provided lasting real solutions and value.” I gave a list and said “pick one.” You pushed back on Deel. I pointed out why they have value.
You then pivoted to the non sequitur that I “like Deel.” I pointed out that isn’t true, and never needed to be—this was never a discussion about which start-ups we want to put a sticker on. It’s a bit rich to complain about logical fallacies after throwing one out.
I wasn't convinced that Deel provided value, though it seemed that that by subjective or semantic means you had reasonably come to the opposite conclusion. My attenpt to boil that down to something pithy didn't do much good, and after your response I felt like the discussion was providing more heat than light, so I used a humorous reference to indicate that I thought we were mostly naysaysing each other.
You haven't changed my mind but you may have changed others' and you did make me think. Let's have an even better discussion next time.
> Can you quantify what you mean by "provided value"?
I took it to mean ignoring companies in sales & marketing or FIRE (finance, insurance, real estate), as well as focussing on the effects of a company's products and services versus their valuation.
Private equity has mastered this formula in the last decade.
Throw as many unrelated businesses that you can buy for pennies on the dollar into a basket. Claim it is now diversified. Decrease expenses by allowing all of your good employees to run for the hills screaming while not replacing them. Post one year of record profits by sacrificing quality in the name of quantity. Re-capitalize in order to generate a massive return. And then rinse and repeat on your next victims, I mean ventures.
I personally think this will be one of the biggest "what the hell were we thinking" ideas to come out of the still-developing hangover of the 2013-2023 startup space. From my perspective, the once-startups that are now big companies are the ones who focused ONLY on solving a real and immediate problem with happy, motivated employees. Every industry has inefficiencies to improve upon. Most employees want normal hours, a 401K, some benefits, and clear performance reviews. There's still a ton of work to do at cool-but-not-the-absolute-cutting-edge startups.
This is simply not true. Most startup founders are genuinely trying to build something that has value to the economy, and therefore society. Most fail early, most of what's left fails later, but some go on to be real, profitable, businesses.
That's how it works outside of the startup world too. Most restaurants, clothing stores, car repair shops, etc. go out of business after just a few years, while a few move on to become large chains and the like.
Your last paragraph paints a dual vision of the world that just isn't accurate.
There are the two you mentioned:
1- small businesses that go out of business after a few years
2- ...ones that become big chains
But you are forgetting #3, the one that actually makes up the largest proportion of businesses:
3- people that quietly, successfully, run their businesses in a sustainable, family and community oriented manner (not financial growth focused but personal growth focused) and then retire quietly surrounded by loved ones.
> But you are forgetting #3, the one that actually makes up the largest proportion of businesses: 3- people that quietly, successfully, run their businesses in a sustainable, family and community oriented manner (not financial growth focused but personal growth focused) and then retire quietly surrounded by loved ones.
I wasn't forgetting them, they just aren't relevant to a conversation about venture capital, at least to me. There are plenty of small bootstrapped startups out there, but their ROI isn't high enough to justify a gamble on the part of investors. Likewise with the businesses you're talking about.
Yes the make up most small businesses, but most small businesses aren't looking for private venture capital, and don't fall into the category of what OP was claiming is "pump n dump".
"According to the U.S. Small Business Administration (SBA), 99.9% of businesses in the United States are small businesses, which are defined as independent businesses with fewer than 500 employees. This amounts to around 33.3 million businesses, which employ 61.7 million Americans, or 46.4% of private sector employees. Small businesses are a vital part of the American economy, contributing to economic stability and representing 43.5% of the country's GDP.
Here are some other statistics about small businesses in the United States:
84.8% of businesses in rural areas are small businesses
43.2% of small businesses are owned by women
20% of employer businesses are owned by minorities
Small businesses account for about 96% of employer firms in high-patenting manufacturing industries
nearly every VC and accelerator in the world is going to ask you, "what's your exit plan?" and look at you like a Martian if you say you're in it for the long haul rather than exit as soon as an exit is viable.
imho that sort of "aim for the exit" thinking strongly incentivizes pump-and-dump behaviour and I think the bandwagoning we see with every hot new tech is exactly what one would expect as a result.
Well... they are investing to make money, not to hold shares in a company forever. If they wanted to do the later they'd buy index funds and sit on them. Most (I'd bet nearly all) capital sources for VCs do both actually.
Also:
1) IPO is one of those exit strategies. Tons of startups are aiming to become long lived independent companies.
2) M&A makes sense for lots of businesses, as you get to take advantage of better economies of scale as part of a larger org. Saying it's your exit plan doesn't make you a scammer.
I dunno, for all the perceived (and real) failures, the last few generations of these experiments have spawned some of the highest earning, and arguably transformative, companies in history.
Which huge companies were spawned from the last few decades of startups raising VC? I think almost all the biggest companies today?
Here's a list I grabbed of the highest market-cap companies:
1. Apple
2. Microsoft
3. NVIDIA
4. Alphabet (Google)
5. Amazon
6. Saudi Aramco
7. Meta Platforms (Facebook)
8. Berkshire Hathaway
9. TSMC
10. Tesla
Of this list, here are "startups" that raised VC money:
1. Apple
2. Microsoft
3. NVIDIA
4. Alphabet (Google)
5. Amazon
6. Meta Platforms (Facebook)
7. Tesla
So more than half of the top-10 biggest companies today were these kinds of startup "experiments".
And of course, these are just the biggest companies, and don't include many things I interact with on a day-to-day basis, which are incredibly valuable to me, e.g. Netflix, ChatGPT, YouTube (part of Alphabet but important enough to warrant mentioning), etc.
> It's not just AI startups. It's almost all of them.
The reality is that a very tiny percentage of new companies get institutional financing at all (and the majority of those that do aren't technical and aren't startups in the sense that they aren't intended to be high growth at a high margin). The loud press that makes it sound otherwise is just survivorship bias.
The core use of crypto in practice is gambling/speculation. People say that isn't a real business because they don't like it but human nature is what it is. Casinos, lotteries and bookies don't seem to be dying out, or crypto for that matter.
HN is not immune to hyperbole and hype. (And now it's pretty trendy to hate on startups.) While he truth is ... a lot more complicated.
AR/VR was hardly a pump and dump. There are massive technical and creative challenges for both, but ... I fail to see how it was a pump and dump. (Unless you mean Faceberg and Moogle spending ridiculous amounts on it, probably because Neuromancer was pretty damn influential.)
Similarly self-driving cars clearly have value (~1.3M people die each year on the roads absolutely unnecessarily, as most of them could have been prevented ... 1.3M people die of TBC, which is again preventable)
If you think EV is a bubble I would like to offer you this nice prospectus of leisure activities on Venus. (Uh, one hyperbole, thanks HN!) That said EV seems the same kind of bubble as housing. The prices are high because of the demand. Battery prices (the biggest cost currently in EVs) keep going down. Demand is going up. Heatwaves are pretty good marketing "for" global warming.
> I truly wonder how much of the U.S. economy has slowed down because of all this wasted effort.
I wonder how frequently startups and hype KILL technologies. A technology starts to show progress, everyone gets excited about that progress, tons of money gets dumped in, everyone rushes to put that technology into production, technology flops, funding for technology dries up. Meanwhile, a lot of people got very rich, so we repeat the cycle. But there was an important step missed here if we want a technology to actually succeed. Research.
The hype train only seems to gear up for technologies that are still in their infancies. This of course makes sense, because how else do you raise money? Chicken and the egg. But the problem here is that you're not using that money to solve the problems with the tech and get it to the point where it could then be effectively used in things, but instead rush to hamfist it into everything you can. Sometimes you gotta slow down first, and prep for the sprint. Google seems to be the worst at this, routinely killing projects that 5-10 years later are then sold as high status items. A company with that much capital doesn't need to rush to market, you can strategize and take your time.
The problem is that the details always matter and they can be a huge make it or break it for a technology. You can generate all the safe miles you want, but if you run through a red light anytime there's a child holding a green balloon a bit too high, you failed. You can build fancy games, but if you promise a holodeck and deliver a hollowdeck, people are reasonably upset. Hype is a double edged sword. You need some to get people on board, but too much and you overpromise. You can build crazy and legitimately helpful tools, but if you're (or others) are promising 10x what you can, then when they get their hands on it they have every right to be upset. Even if the thing is still useful to them and betters their lives. They are rightfully upset, but they'll unrightfully dismiss what you made. So the people who benefit from the hype cycles are the grifters. The ones willing to jump from one thing to the next. Which is why I'm always very confused why we prop up people who have "expertise" in all the last hype bubbles as if it is credentializing why they are experts in the new hype bubble. How does a giant and brilliant red flag get interpreted as green? Maybe we're not so different from self-driving cars after all.
The important thing to understand here is that you shouldn't write on a latex balloon with a Sharpie. It works well for a while, but when the balloon inevitably pops or deflates, the Sharpie ink will re-liquify and get all over the place. Then you will learn the meaning of the phrase "indelible on porous surfaces".
It should be noted, there are a variety of chemicals that will dissolve and carry away the ink from most things if you look for it. Of note, rubbing alcohol and paper towels work well.
Likewise, the paint markers which I make use of are easily removed by rubbing alcohol.
There's a useful trick many educators know: if someone uses sharpie on your whiteboard, draw over it with another whiteboard marker and you can wipe it up like normal. Also applies to glass and some other surfaces. Essentially the same process you're mentioning but different chemicals. But always be careful with solvents. If water or isopropyl alcohol (IPA) doesn't work, you should know what you're doing/using.
The weird thing with AI startups is the funding is kind of smoke and mirrors. So much of the funding is tied up in NVDA, MSFT, GOOGL, or AMZN incentives that the actual cash value is... murky.
And how many of these companies' customers are just other VC funded startups burning money and buying services adjacent to their own? Like a bigger circle.
I don't believe I've seen anything in the TV show Silicon Valley that wasn't based on something IRL. Dig around on anything that appears in that show, and I defy you to find something that the writers just pulled out of their ass.
SMBC is a treasure, just like XKCD. Zach's and Kelly's books are quite nicely, some like Soonish are just good accessible futurism/science reading, their comic collections are excellent, and they aren't afraid to offend. For example, I was part of their initial Kickstarter for The Holy Bible: Abridged Beyond the Point of Usefulness https://hivemill.com/products/pocket-bible and now I have a much-easier-to-get-through-and-considerably-more-fun Bible with Zach's signature. It has provided for much silliness at many parties in my house. I don't rotate in baptist/fundamentalist circles though, those would probably be aghast at it.
One thing I noticed from the Sili Valley startups I've been involved with in the last 20 years: exec management's parents went to school w/ the funds managers.
I have this theory that there's institutional money burning a hole in various institution's collective pockets. Those funds' managers are under pressure to invest a percentage in risky, high reward opportunities. So they have money that has to be invested. You would look pretty stupid if you missed out on investing in the next Facebook or Google.
So you get some valley VC firms managed by guys who went to MIT, UCB or Stanford in the 80s and made their cash founding 90s era startups.
Now these guys' college room-mates have kids in MIT, UCB and Stanford who are about to graduate. So they (the VC's old room mates) channel some dosh into the VC's funds. In return, the VC's fund startups started by their room-mate's kids.
For the parent, it's great. They avoid estate taxes and if there's a profit, they only pay capital gains rates.
The kids receive their pay in perfectly laundered equity which at the next liquidity event can be reified outside the IRS' jurisdiction, so they wind up paying 12% corporate tax in Panama instead of the on-shored 26% cap gains rate.
All that's really required is a greater fool to take the series A shares off your hands when you do a series B.
My take on SiliValley is it's turned into a tax avoidant generational wealth transfer system for the financier class.
Nice work if you can get it.
Academia supports the system with hype-filled pronouncements about how future tech (AI, RISC-V, Battery Tech, WISC, etc.) will completely revolutionize the world.
Analysts who should have been taught skepticism when corporate leaders imply their startup will have a market value in excess of western Europe's GDP, jump on board to avoid looking like idiots on the off chance the firms they cover do take off.
Tech investing isn't a scam, there are definitely corners that are producing solid tech. But they're few and far between and very often not involved in the VC economy.
> I have this theory that there's institutional money burning a hole in various institution's collective pockets.
That was very much the case before the Fed started raising interest rates.
The term "zero interest rate phenomenon" was created to describe it.
Now that rates are up, people sitting on piles of money can just put it to work in more traditional lending (e.g. buying bonds) instead of needing to look around for weird startups to put it in.
While some people can raise a ton of money for ideas that don't seem to make sense, the majority of others won't raise a dime. It's a matter of access (which could be a signal for early investors ability to cash in for later investors.)
For investors in the art market, they will invest in some artists and not others, irrespective of how appealing the art is, as provenance is the signal for ROI.
I would watch a remake of "The Producers" set in Silicon Valley with startups instead of Broadway productions.
For those unfamiliar, it's about guys who fraudulently resell the same shares in the profit of a Broadway production 25 times over, and then try to ensure that it's a failure so there won't be any profit to pay, but accidentally create a success and are ruined.
Hi. I’m an artist. If you have a transparently idiotic startup idea and contacts with a lot of dumb money looking for unicorns to fund, drop me a line, I can use AI to create logos and promotional imagery for only a few thousand dollars up front and a small percentage of the VC investments you get.
The AI in question is Adobe Illustrator but if it gets us in front of more dumb money to say it’s the other AI then I do not care.
I thought this was going to be a long form essay/prank/experiment, where someone launched a crowd funding campaign for a "transparently idiotic startup", and pulled off a $1M raise.
Ironically however, if you were to do that well, it would show high intelligence and creativity, which would probably make it an actual good investment.
Like a lot. Didn't YC even fund that startup that wanted to turn middle class suburban homeowners into slumlords by building rental poverty shacks in their back yards?
I would say phrased uncharitably, but factually right on the money IMO. I strongly suspect that if this scaled it would have caused an explosion in the number of landlords while also making property even more expensive because now all of a sudden you can turn every PPoR into a rental as well. Absolute nightmare scenario.
Thankfully the market took care of this idea and they only made 10 "houses" before going pop apparently.
I don’t know what startup you are mocking, but this is a very bad faith way to characterize building ADUs cheaply. You don’t know the first thing about housing, besides making drama about it.
Multi-family units and rezoning single family plots is great. Trying to get poor people to live in the back yard like a family pet is a humiliating proposition.
Poor people are not children or animals. You can allow for them to just choose not to live in the backyard. For the right price I might have gladly done this.
But this is the great debate of our generation, I suppose. Are people capable of making choices or not?
> Trying to get poor people to live in the back yard like a family pet is a humiliating proposition
I would have loved to have had this option in college. Glad my landlord saved me from humiliation! (And let us be honest, for every useful idiot who thinks they're helping the poor by constraining their options there are ten homeowners who don't want poor people in their neighbourhood.)
Can’t help but think that all of FANGS would have seemed to be “transparently idiotic” early on. I mean, creating yet another search engine with no business plan is pretty idiotic. As is creating yet another social network or a book store. Come to think of it, all of the “v1.0” projects I worked on seemed pretty idiotic too and I thought they’d never make money. A good number of them have $1B+ yearly revenues today. To me the takeaway is: aside from the truly clinica cases, you can never tell what will take off and what won’t.
because it has positive economic externalities and avoids the moral hazard of enabling hucksters.
i have no problem with investment, but investing in an obviously stupid startup just cause they were launched by a friend's kids doesn't seem like a way to allocate capital that Adam Smith would appreciate.
> vast majority of investment is inside of public firms into themselves for ai
FAANG spent $114bn on R&D in 2022 [1]. VC funded nearly 4x that amount [2]. There are valid complaints about the state of venture capital. Lack of money isn't one of them.
I truly wonder how much of the U.S. economy has slowed down because of all this wasted effort. Is it really a good idea for every company to try and reinvent what a company is? Every startup I've been spends a non-zero amount on just re-figuring out how to be organized, how to manage people and projects, and even just figuring out the logistics and tech needed to run a company. It's a lot of wasted effort, time, and money. My time at startups has been the least effective portions of my career. Real work actually moves quite slowly at startups, in my experience.