Which Categories of Seed Startups Are Thriving? Which Aren't?(tomtunguz.com)
tomtunguz.com
Which Categories of Seed Startups Are Thriving? Which Aren't?
https://tomtunguz.com/which-categories-of-seed-startups-are-thriving-which-aren-t/
56 comments
Fascinating that so much money is going into blockchain startups when I can’t think of that many useful applications for it.
Agreed, nearly every commercial use seems better off simply relying on a centrally trusted authority.
"This has to be on the blockchain. But! don't look behind the curtain to everything else in our company that relies (and thrives) on a single point of trust.
"This has to be on the blockchain. But! don't look behind the curtain to everything else in our company that relies (and thrives) on a single point of trust.
Including quick money transfer. FED is working on an upgrade that will allow instant USD transfers. There's absolutely no reason why the central banks of the world cannot have instant transfers using their non-blockchain databases.
BTC is only good as a deflationary pseudo-commodity, like gold. And like gold a good holding amount is probably below 10%, I'd say below 5% of the portfolio.
BTC is only good as a deflationary pseudo-commodity, like gold. And like gold a good holding amount is probably below 10%, I'd say below 5% of the portfolio.
Yes, I always look to the FED for cutting edge technology. /s
Instant money transfer is not cutting edge. Some countries have had it for decades (albeit at a premium cost).
http://www.fasterpayments.org.uk/our-achievements : not even at a premium cost.
The underlying blockchains often do not have central authority; that is, no one entity can make a rash decision for bitcoin tomorrow.
Cryptokitties can manipulate their contract, so in that case it is true a central authority is in control. Even though Bitcoin has been around 10 years there is still a lot of development and people are still figuring out how to approach this type of thing. The obvious answer is to start with a central contract and have a mechanism to slowly decentralize or remove ownership. It’s going to take time to get it right but it will happen (and even cryptokitties can upgrade their contract to include this logic).
Cryptokitties can manipulate their contract, so in that case it is true a central authority is in control. Even though Bitcoin has been around 10 years there is still a lot of development and people are still figuring out how to approach this type of thing. The obvious answer is to start with a central contract and have a mechanism to slowly decentralize or remove ownership. It’s going to take time to get it right but it will happen (and even cryptokitties can upgrade their contract to include this logic).
The underlying blockchains often do not have central authority; that is, no one entity can make a rash decision for bitcoin tomorrow.
I disagree, they have plenty of pinch points which allow de facto control.
The small group of core developers control the laws of the currency and directly tied to that its value. They can perform hard forks and change all the rules. The users have very little influence there.
Exchanges control which currencies are popular, along with huge amounts of the money in circulation, and change things like deposit/withdraw rules arbitrarily.
The miners can attack the currency to rewrite history (has been done), refuse to use new rules, or switch to a new currency.
Many of the fundamental tenets of bitcoin as it started out (that most users will mine, that most users will take care of their own wallets) have been proven too optimistic when in widespread use. People just don't want decentralised things given the massive tradeoffs and the rewards of centralisation for companies are too great.
I think blockchains like bitcoin are a very interesting experiment, and currencies will escape state control as nation states continue to wither away, but none of the current currencies is fit to take that role, nor was the pivot of startups to 'blockchains for...x' a good idea - it's a solution in search of a problem in most cases. We are unfortunately more likely to see a corporate currency or one from a superstate take that role than a decentralised one.
I disagree, they have plenty of pinch points which allow de facto control.
The small group of core developers control the laws of the currency and directly tied to that its value. They can perform hard forks and change all the rules. The users have very little influence there.
Exchanges control which currencies are popular, along with huge amounts of the money in circulation, and change things like deposit/withdraw rules arbitrarily.
The miners can attack the currency to rewrite history (has been done), refuse to use new rules, or switch to a new currency.
Many of the fundamental tenets of bitcoin as it started out (that most users will mine, that most users will take care of their own wallets) have been proven too optimistic when in widespread use. People just don't want decentralised things given the massive tradeoffs and the rewards of centralisation for companies are too great.
I think blockchains like bitcoin are a very interesting experiment, and currencies will escape state control as nation states continue to wither away, but none of the current currencies is fit to take that role, nor was the pivot of startups to 'blockchains for...x' a good idea - it's a solution in search of a problem in most cases. We are unfortunately more likely to see a corporate currency or one from a superstate take that role than a decentralised one.
If you (or anyone else) think you can effectively break bitcoin then there is a billion dollar bounty available to you in the form of a highly leveraged bitfinex short position.
Short positions on highly volatile (and probably manipulated) markets are extremely vulnerable to short squeezes. There's no good way to take out a leveraged short that's guaranteed to remain in place for, say, six months.
Good to know, but just because any particular short position is not guaranteed to remain in place for six months doesn't mean there is no good way to profit from shorting Bitcoin.
The problem is that a leveraged short position that is closed out/margin called can lose more money than you initially invested. The market can remain irrational longer than you can remain solvent.
Much of this comment is straw men (exchanges, miners, wallets).
Anyone running a node can indeed decide what software they run. That’s exactly my point. There’s no such thing as “they” performing hard forks. Developers make decisions, sometimes they disagree and fork. Miners make decisions on which fork they want to run. Users are free to run a node.
Generalizing “blockchains for x” doesn’t do much good. Sure there are terrible blockchain projects as there are terrible app ideas, but there are quite a few good projects out there. I’m a fan of Keep.network for example.
Anyone running a node can indeed decide what software they run. That’s exactly my point. There’s no such thing as “they” performing hard forks. Developers make decisions, sometimes they disagree and fork. Miners make decisions on which fork they want to run. Users are free to run a node.
Generalizing “blockchains for x” doesn’t do much good. Sure there are terrible blockchain projects as there are terrible app ideas, but there are quite a few good projects out there. I’m a fan of Keep.network for example.
.. but the meaning of the blockchain and its purchasing power (or other real-world effects) are entirely determined by the software that interprets it. We've seen this a number of times, such as the Ethereum "DAO" roll back.
> Miners make decisions on which fork they want to run.
Miners are surprisingly consolidated, so this mostly determines which version "wins". And the economic value of which fork wins is determined by the markets. If you keep running the "wrong" version of a fork you find out that it's worthless and only the coins on the "right" one retain value.
> Miners make decisions on which fork they want to run.
Miners are surprisingly consolidated, so this mostly determines which version "wins". And the economic value of which fork wins is determined by the markets. If you keep running the "wrong" version of a fork you find out that it's worthless and only the coins on the "right" one retain value.
The question is why it's the right choice for the underlying blockchains not to have a central authority. There are costs and benefits to decentralization, and I've very rarely seen a blockchain startup weigh the two rather than just touting the benefits.
(I know this is a high standard for any new technology; I'm not saying I expect it on the front page of their website.)
(I know this is a high standard for any new technology; I'm not saying I expect it on the front page of their website.)
If you take a look around you'll see it's mostly non-technical people throwing their money into the crypto space (see reddit.com/r/bitcoin), unfortunately, this applies to investors too. How many years has it been and still all we have is 200 different shitecoins.
There is a very useful application for blockchain: Leverage banks desire to stay current and not loose out in the next wave of change in order to sell them platforms which generate far more interesting press releases than a regular database would..
Like no CTO ever got fired for choosing IBM - No digital transformation manager ever got hired without at least mentioning blockchain.
(Edited to add: Note this isn't a bad thing per-say. Ive seen blockchain used as ammo to do push something through thats actually useful, regardless of whether it was the most apt tech choice or not.)
Like no CTO ever got fired for choosing IBM - No digital transformation manager ever got hired without at least mentioning blockchain.
(Edited to add: Note this isn't a bad thing per-say. Ive seen blockchain used as ammo to do push something through thats actually useful, regardless of whether it was the most apt tech choice or not.)
This chart looks like it's only through 2018. My hunch would be that Blockchain is down or flat thus far this year.
Since the author doesn’t cite a source for his graphs, we are left to speculate on how recent the data is. Blockchain certainly seemed to me to be very hot in 2018. Not so much this year.
More companies and developers are entering the cryptocurrency space, though there are certainly defunct companies, too.
DeFi is the most concentrated aspect of blockchain at the moment. Assets going into DeFi are up and to the right: https://defipulse.com/
DeFi is the most concentrated aspect of blockchain at the moment. Assets going into DeFi are up and to the right: https://defipulse.com/
It says through 2018
Blockchain allows protocols for finance. Derivative contracts like perpetual futures are inherently better than old-school futures. The innovation is much more rapid. Building a new exchange becomes permissionless. Assets can flow from old exchanges to new, or be moved around to many rapidly. All of this is simply better than the current exchange situation - 24/7 trading, low fees, rapid innovation, easy capital formation. Finance will move to the crypto rails because it’s better.
How is a perpetual futures contract inherently better than "old-school futures"? It is my understanding what cyrpto calls a perpetual future _isn't_ a future which makes them difficult to compare. I'm reading this comment as if perpetual futures are going to replace our current futures system somehow, and I don't see this possibility.
Don't we have 24/7 trading without the block chain? And low fees?
It appears all the many of these benefits you mention are for those just transacting in a crypto currency, and not for the other 99% of finance transacting in other currencies (and physical goods).
Don't we have 24/7 trading without the block chain? And low fees?
It appears all the many of these benefits you mention are for those just transacting in a crypto currency, and not for the other 99% of finance transacting in other currencies (and physical goods).
I explained more in the other comment, it isn't inherently better as you say, but it's different and innovative.
> Derivative contracts like perpetual futures are inherently better than old-school futures.
Can you elaborate on this? Physically settled futures are important to producers and consumers of commodities. I can imagine that replacing these with a cash settled contract might increase spot volatility.
Can you elaborate on this? Physically settled futures are important to producers and consumers of commodities. I can imagine that replacing these with a cash settled contract might increase spot volatility.
You are right, it is a different product, it isn't inherently better simply because it's perpetual. But it is quite interesting and provides different trading strategies to the standard futures contract. See https://www.bitmex.com/app/perpetualContractsGuide for a good explanation for how Bitmex's implementation works which is the most popular version.
I think a lot of exotic derivative contracts will be innovated in the crypto exchange networks, and those that become popular will eventually applied to traditional commodities and equities once the bureaucracy comes up with regulations in 5 to 10 years.
I think a lot of exotic derivative contracts will be innovated in the crypto exchange networks, and those that become popular will eventually applied to traditional commodities and equities once the bureaucracy comes up with regulations in 5 to 10 years.
we are probably still early in the hype cycle
https://en.wikipedia.org/wiki/Hype_cycle
https://en.wikipedia.org/wiki/Hype_cycle
What about dfinity.org?
Blockchain has many useful applications: art, insurance, property inference. imo the problem is the systemic changes are not worth the gains (inertia is strong, therefore not that useful?).
Trustless/trust-lite blockchains are a strong fit for many applications, but far from proportionate to the number of applications being attempted.
Take, for example, property titles on the blockchain. Sounds promising. But property rights are ultimately enforced by governments (namely, the government that projects control over the area on which the property sits) through their legal systems.
In this case, for the majority of reasonable threat models, blockchains add complexity without adding value. Say that your wallet is holding a title token. I hack your computer and use your private key to transfer the token to my wallet. According to the blockchain, I am now the rightful of the property. However, it's very unlikely that your government would recognize me as the owner. The government wouldn't let me move all my stuff over and start living there just because the property token is in my wallet. The blockchain state and de facto/de jure state have diverged.
Similarly, say the government becomes tyrannical and begins seizing property. The property token in your wallet will, once again, not save you. You will be jailed. The "code is law" of blockchains becomes redundant to the "law is law" of governments.
Take, for example, property titles on the blockchain. Sounds promising. But property rights are ultimately enforced by governments (namely, the government that projects control over the area on which the property sits) through their legal systems.
In this case, for the majority of reasonable threat models, blockchains add complexity without adding value. Say that your wallet is holding a title token. I hack your computer and use your private key to transfer the token to my wallet. According to the blockchain, I am now the rightful of the property. However, it's very unlikely that your government would recognize me as the owner. The government wouldn't let me move all my stuff over and start living there just because the property token is in my wallet. The blockchain state and de facto/de jure state have diverged.
Similarly, say the government becomes tyrannical and begins seizing property. The property token in your wallet will, once again, not save you. You will be jailed. The "code is law" of blockchains becomes redundant to the "law is law" of governments.
I hadn't thought about this and it's an excellent point, thanks for making it.
You could say the same thing about SQL databases, but all billion-dollar companies are built on them.
In 50 years there won’t be a single Fortune 500 company that isn’t built on DLT.
In 50 years there won’t be a single Fortune 500 company that isn’t built on DLT.
I didn’t downvote you, but that’s a pretty extraordinary statement that could use some explanation. How do you see it being used such that it reaches that level of ubiquity?
Any time two or more different parties enter into any sort of business relationship, they need to spend tons of money on third-party service providers tracking who owns what. If DLT can reduce the cost of that by, say, 60 - 80% then that's a savings of several trillion dollars per year.
If you read a book about the impact that double entry accounting had on the world then I think it's easier to imagine all the ways that triple entry accounting (DLT) will transform the economy.
If you read a book about the impact that double entry accounting had on the world then I think it's easier to imagine all the ways that triple entry accounting (DLT) will transform the economy.
Interesting, thanks for explaining. Any recommendations for good books on the impact of double entry accounting?
There is a book called Double Entry. The first few chapters are worth reading to understand how it actually changed the world and created wealth. It gets kind of boring after that.
I think understanding DLT as a 10x less labor-intensive version of double entry accounting is key to understanding its ability to enable new and qualitatively different types of business relationships.
I think understanding DLT as a 10x less labor-intensive version of double entry accounting is key to understanding its ability to enable new and qualitatively different types of business relationships.
I'll check it out, thanks!
The charts look lovely but is this even worth discussing without knowing were the data comes from? From all I know the author could have made those in Microsoft Paint O_o.
> From all I know the author could have made those in Microsoft Paint O_o.
It's R package ggplot2. The title isn't center by default and those colors are the color ggplot2 tend to use by default.
But I do share your concern in term of lack of clarification in where the poster got his data from.
It's R package ggplot2. The title isn't center by default and those colors are the color ggplot2 tend to use by default.
But I do share your concern in term of lack of clarification in where the poster got his data from.
Agree. They did themselves no favors with those graphs looking like comic sans, almost.
Very difficult to make anything in MS Paint
This is a bit of a personal plug, but as for examining trends of funding and startup activity in the “AI” realm, I wrote up a brief summary of some CB insights investment data earlier this year [1]. I thought it was interesting to examine the difference between investment activity and entrepreneur activity, as there is always some inherent information asymmetry between entrepreneurs and investors. Not a deep analysis, but a few basic charts :)
[1] https://medium.com/@kevinconnolly/the-top-100-ai-startups-91...
[1] https://medium.com/@kevinconnolly/the-top-100-ai-startups-91...
As others have pointed out, this article really needs to be backed up with sources.
> Software - up more than 3x, Software is a perennial category. Automating expensive internal processes in novel ways is Levi’s 501 jean of the startup world: a stylish classic that will never go out of style.
I agree with the above text but the graphic for software shows a clear plateau and doesn't quite bear this out.
> Software - up more than 3x, Software is a perennial category. Automating expensive internal processes in novel ways is Levi’s 501 jean of the startup world: a stylish classic that will never go out of style.
I agree with the above text but the graphic for software shows a clear plateau and doesn't quite bear this out.
Yes, but it is a high plateau, the highest on the chart. Perhaps, rather than Levi's 501, Valentino Rockstud.
I loved the graphs in this article. Would be interesting to plot some of the most successful companies in each category and see if there was any correlation to whether they were formed at the start, peak, or tail of a "wave".
Also, some categories had multiple waves, e.g. "pre-mobile social media" (early '00s) and "post-mobile social media" (early '10s).
Also, some categories had multiple waves, e.g. "pre-mobile social media" (early '00s) and "post-mobile social media" (early '10s).
But this uses funding as a metric for thriving. That's seems misleading. Lots of startups raised boatloads of cash and crashed.
Funding doesn't mean it's succeeding. It just means it's really hot right now. I would much rather see which category of startups are actually succeeding: as in getting revenue, profits and user traction.
Revenue numbers for early stage startups might be hard to come by.
I was in an accelerator with a bunch of startups and one in particular was having a lot of trouble putting together a series A. And strangely, they were one of the few that had decent revenue numbers.
At the suggestion of a savvy fundraiser, they took all mention of their actual revenue out of their pitch and they immediately started getting real interest.
They were then able to raise the funds.
It seems that actually having revenue just really dashed the pie in the sky dreams of many investors. With no revenue or profit, they were free to pretend that when the money started flowing, it would be hockey stick growth.
With revenue numbers, they could crunch the numbers and see the gradual growth curve and that would scare them away.
I was in an accelerator with a bunch of startups and one in particular was having a lot of trouble putting together a series A. And strangely, they were one of the few that had decent revenue numbers.
At the suggestion of a savvy fundraiser, they took all mention of their actual revenue out of their pitch and they immediately started getting real interest.
They were then able to raise the funds.
It seems that actually having revenue just really dashed the pie in the sky dreams of many investors. With no revenue or profit, they were free to pretend that when the money started flowing, it would be hockey stick growth.
With revenue numbers, they could crunch the numbers and see the gradual growth curve and that would scare them away.
Consumer fintech definitely feels like a bubble. There are a few promising entries in the space, but the overarching trend is price wars in various sectors (trading/investing, loan underwriting, wage advances, bank accounts) competing for the same (scarce) millenial dollars. The TAM is certainly large, but margins will quickly go to zero, and with the sheer amount of contenders in each space, I don't see the majority making it out alive.
This visualisation would be more useful if all the graph followed the same scale on the Y axis, IMO.
Appreciate the charts, but what is "Software" as a category? Looks like a container for anything that doesn't fit into one of the other categories. Can someone explain?
Interesting, but it would be more interesting to see these charts alongside a 5-10 yr returns graphs in the same said categories
I think the question should be more like "Which categories of seed startups are getting FUNDED" rather than "Thriving". Thriving means they will develop into sustainable businesses, and while most of the shown categories will, it probably omits startups in less trendy categories that are thriving but are not necessarily attracting funding at the same level as these. I understand it's great to chase billion dollar unicorns, but most of us here will probably only ever build million dollar companies. Would be interesting to learn more about their dynamics as well.