Hacker News Reader

There's two competing lenses of thought that bookend your struggle:

On one hand, Sam's point is basically described by Paul Graham's essay on "Narrow and Deep" ideas [1]. You want an idea to have a few passionate users rather than lots of lukewarm users, because there's an activation energy to building a habit and if you're only lukewarm you're not going to get over it. As PG put it, "Sum that reaction over a whole population, and you have no users."

On the other hand, you have Geoffrey Moore's "Crossing the Chasm" [2]. This has the critical observation that early adopter & mainstream populations are different. About 10% of the population has an early-adopter mindset: they will try anything because it's new and cool. Those are the people Sam says you should be hanging out with in this article. But your experience with them doesn't generalize, because most people don't think like that. Most people are self-interested and have short attention spans: if it doesn't deliver clear value to them with minimal hassle, they aren't interested.

The set of world-changing products really exists in the intersection of these two worldviews. Products that had lots of passionate early adopters but failed to cross the chasm include Lisp, LiveJournal, RSS, VR (both times), and cryptocurrency. Products that aimed for too mainstream a market and only had lukewarm users include WebTV, grocery delivery, and smartwatches. In between are products like the iPhone, where the PDA market had been sitting in early-adopter limbo since 1991 but technological changes finally made it practical, or YouTube, where streaming video had been in early-adopter limbo since RealAudio in 1997 (?) but cell-phone video, Flash, and easy web uploads made it practical in 2005.

[1] http://www.paulgraham.com/startupideas.html

[2] https://en.wikipedia.org/wiki/Crossing_the_Chasm

I digress, but I think the real turning point that helped YouTube was Apple requiring data plan for iPhone and convincing AT&T to offer unlimited data plan. So sometimes help will come from your rival :o

I thought YouTube's turning point was when users could make money.

Seems to me things that "come out of nowhere" switch from one to another. So if you optimize for either, you get neither.

I think that we lose a lot from a mentality that isn't adaptive, that prescribes in advance whether something is going to become popular or not. Because then something that is currently a niche gets killed off before it finds the mass market. I feel like whenever I find something new I like in the supermarket it disappears a short time later, presumably because of greedy optimization.

>grocery delivery

I think the USA implementation is poor not the idea is bad. Here Finland K-ryhmä[1] did really nailed implementation and it is like 10x better than going to the supermarket. My mom and aunt both in their 70 started using this separately and both are saying they are not going to back in supermarkets after covid19.

[1] https://www.k-ruoka.fi/kauppa

For folks in the Bay Area, I constantly recommend Good Eggs (goodeggs.com). They’re serious farm-to-consumer folks and have small producers and fishmongers that you wouldn’t see otherwise (many of the names will be familiar from restaurant menus).

After having gone through CSAs, our various decent grocers, and then realizing there were “no” retail fishmongers, a friend had recommended Good Eggs and we’ve never gone back. It’s not for everyone, as it’s basically only the priciest things you’d find at Whole Foods, but it’s also delivered to your door and without a large-scale grocery chain in between.

Covid-19 has obviously strained their availability, but overall they’ve still been amazing.

In Hamburg Germany they have Grüne Kiste (The Green Basket), which deliver high quality foods on weekly basis. Other services are also growing, but I have nothing but positive to say about Grüne Kiste.

That's interesting - I live right next to a K-Market and I'm not in a risk demographic so not had the need, but when I moved from the UK to Finland some 10 years ago one thing I missed was that the UK supermarkets at the time had excellent online delivery systems (they're buckling under the strain of the current crisis due to surge in demand, but back then they worked well enough). Finland didn't seem to have anything equivalent until recently, which I thought was more due to geography, population density and other factors that just didn't make it economical for the K and S chains.

Supermarket delivery in the UK has been fantastic, much better than Finland. I think you are right about the population density, and also the amount of competition between 8 large supermarket chains in the UK, vs the near cartel like situation in Finland.

Another example would be automated checkouts - that came much later in Finland compared to the UK. I think there's a good point there that the duopoly stifles innovation, plus there's just a much smaller market and less pressure to scale up.

Man I still miss LJ so much sometimes.

(Yes I know about Dreamwidth, I also miss having my friends all there and getting them writing essays again when Twitter is so tempting is an uphill climb.)

From a VC's point of view, for early startups, only pre-chasm matters. If many don't cross, well, it's a numbers game.

Oftentimes the VC cash is needed to actually cross the chasm. One of the difference between early and mainstream adopter populations is that a lot more product polish, and oftentimes explicit sales or advertising outreach, is needed to please the mainstream audience. That all costs money.

I wouldn't say that only pre-chasm matters, though. VCs usually refuse to invest in startups that have zero chance of crossing the chasm. The prototypical example of this is say an app for D&D players. You could have the most passionate userbase, and the app could be critical to their experience, but your total addressable market maxes out at the 13.7M D&D players worldwide and it's unlikely they'd spend more than say $10 each on it, which makes it unattractive to all but the smallest VC firms.

There's a ton of D&D apps out there. Why do you need VC funding to make a D&D app anyways? Making $137M on a single app (using your numbers) sounds amazing for an indie dev.

> Most people are self-interested and have short attention spans

Interesting way to judge those who don't hop on every new fad. Maybe they have a longer attention span and are filtering out the noise. They don't get caught in new gadgets and apps because they may be pursuing something larger scale in their life, like building stability for their family or studying deep for college or helping out the local community or training for athletic performance or whatever.

Its only in the start-up bubble that someone who is not an early adopter of Uber or Airbnb or Dropbox or whatever "huge" thing is seen as just selfish and ignorant.

Seen from a wider perspective, yeah these apps have significance but not that much in the daily lives of people. Yeah now I sometimes book my vacation through the airbnb app, instead of booking a hostel or hotel, but it doesn't radically change my vacationing habits. Sometimes I call an Uber instead of a taxi. Has upsides and downsides but it's not some new age of mankind.

Perspective, people.

Nah, I think you're asking a good question! I think your interpretation is correct - a good sign for a platform is if it has a small base of passionate users who love the product.

For some more context, the How to Start a Startup course is pretty good -- here is the lecture I think is most relevant: http://startupclass.samaltman.com/courses/lec06/

Any statement like this isn't going to apply in all cases, and you bring up some good examples. But it's a good general principle.

I see a lot of statements like this (also here in the comments), but is there actual any evidence behind those? Any research results, which has been replicated?

If yes, it would be nice to see it together with such statements, to back them, because I have the feeling most of this is just anecdotical or ideas introduced by some kind of start-up gurus...

Yeah, I agree that these statements are not awfully scientific. I wish I had studies or references to back them up, but I’m not aware of something like that.

However, as others have pointed out, this was originally observed by pg and has become a staple of YC’s advice since then. I think it would be uncharitable to say this is just a hypothetical invented by startup gurus. YC has had a lot of startups pass through their doors, and more experience than almost anyone seeing why people succeed or fail. Their findings may not exactly be science, but I think their advice is as close as we can get right now.

new platform "used by a lot of small number of people" is usually the real trend. As new platforms are quirky and buggy only people who use it a lot can power though all the issues and guide you to a product that works for a large market. They also have enough emotional connection that they can evangelize your product to the wider market.

The "used a little by a lot" people don't have the attachment or deep need to work through issues and they give up after some time. They also can't be bothered to evangelize for you.

Of course, sometimes there is no market beyond the small group. Even then, it is easier to make a business (revenue > cost) in the small "used a lot" market than the other case where people won't pay.

What doesn't make sense to me about that is that the web & email starting out as 'used a little by a lot'. Over time, these not only became better (many innovations for web, mobile access for email), but also ingrained in our life and society. That then paved way for things like social networks that eventually became standard communications. I believe most major communications products seemed to have followed that trend.

Not at all. In 1980 email had a tiny market share but really strong usage by those who had accounts.

True story, DEC included a free mail program with VMS because some developer wrote it. And they were baffled when companies started ordering VAXes mainly to use for email.

Web was similar in 1993. Very few users, yet folks like Adam Curry left his job as a famous “VJ” at MTV to start a website, buying the unused domain mtv.com for the purpose (you can guess how that turned out). I mention this because he was a total norm who got enamored with the web when almost nobody used it.

And Bitcoin, well...

For actual numbers, here’s a graph of computer users since 1980:

https://pbs.twimg.com/media/EZI1TsJXYAAo9ES?format=jpg&name=...

> It's better to have 100 users love you than 1 million kinda like you. The true seed of scale is love, and you can't buy it, hack it, or game it. A product that is deeply loved is one that can scale

Interview with Reid Hoffman, founder of LinkedIn, Altman goes more into his reasoning [0]

https://mastersofscale.com/sam-altman-why-customer-love-is-a...

If they're all paying the same, I'll take the million thanks.

Obviously 'love' implies more attachment and longevity of subscription, but a million x even just once is still 10,000 hundred xs.

But maybe that's just my 'if I started a company I wouldn't be thrilled at the thought of taking VC money' perspective talking.

You want to be Turbotax then!

Linkedin is literally used a little by a lot of people.

At first I thought this was the perfect counter-example.

Maybe LinkedIn works because it’s used a lot by the small group that matters and pays – recruiters! I imagine they’re on LinkedIn all day long.

Paid LinkedIn Premium (which is what matters to founders and VCs) in contrast is used a lot by few people - headhunters, HR and PR people.

A key differentiator is if the new platform is used a lot by a small number of people, or used a little by a lot of people.

I found that statement ambiguous as well. For a while I wondered if he meant to imply either of those cases can be a differentiator.

Dapps haven't gone away. I'd say that trend has evolved into DeFi, which seems to be growing quickly.

Genuine question: what apps/technology is growing quickly with DeFi that can't otherwise be practically executed (for whatever reason) outside of a blockchain? I've yet to see a killer example.

A perfect killer example is globally accessible prediction markets. They always get shut down by governments. Isn’t it crazy that you currently can’t place a bet on a horse race or a presidential election in the US?

Do you have any killer examples that aren’t, um, totally illegal?

If it's legal, people prefer credit card over Blockchain.

Having reliable, auditable, public protocols which can operate without the blessing of the world’s existing governments is precisely the point.

Killer examples include any products or services which would provide a net positive value to society but are currently outlawed within various geographic territories for superstitious reasons and/or to protect existing industries from competition.

Prediction markets are an excellent example here: a substantial body of research now indicates that 1) prediction markets are superior to all other strategies in predicting any quantifiable event (because they are “meta tools” which inherently incorporate the best information from all strategies) and 2) non-private markets can’t be used for manipulating elections, sponsoring assassinations, or incentivizing real-world action of any sort. (In any prediction market with a public order book, any attempted “action-incentivizing position” can be consumed by arbitrage until all that remains is a measure of the event’s real probability). For several decades, researchers from dozens of universities have lobbied for relaxing the ban on prediction markets in USA, with little progress to show. Until that ban is lifted (which could take decades longer), blockchain-based prediction markets will have no domestic competition.

Also worth noting: even in the presence of legal, centralized prediction markets, decentralized prediction market protocols may still be competitive because they require far less counterparty-risk, allowing them to safely support higher volumes and larger positions.

Decentralized prediction markets also offer global liquidity, so the counterparty can be located anywhere on earth. Also the friction and fees should be much lower. If someone in Mongolia wants to bet a million dollars against someone in Colombia today, it’s extremely hard to do without a blockchain (think about the fees and time to wire the money, who would you be wiring it to, etc). Prediction markets are not just for gambling, by the way. They could be used for earthquake insurance, for example.

> non-private markets can’t be used for manipulating elections, sponsoring assassinations, or incentivizing real-world action of any sort. (In any prediction market with a public order book, any attempted “action-incentivizing position” can be consumed by arbitrage until all that remains is a measure of the event’s real probability)

Source(s) for this?

I know prediction markets will have far-reaching consequences when they aren’t or can’t be constrained anymore.

The mere presence of a payout for low-probability is a broad change. Maybe I’m not being paid to kill someone via the market directly, but I am naturally incentivized to collect information about the world, gain power, and then use my power to bet on and make low-probability events happen.

Another question - How do we solve the decentralized oracle problem without relying on decentralized votes that could be biased due to past voting behavior or their vested interest in the outcome of a truth vote?

One of the best papers I've seen is Crime Markets: Non-Usefulness of Prediction Markets for Assassination and Crime: http://bitcoinhivemind.com/papers/6_Crime_Markets.pdf Summary:

> Some worry that censorship-resistant prediction markets will be used to encourage assassinations (and other crimes); this concern does not hold up to a sober examination. “Assassination markets” (AMs), as originally proposed by Jim Bell, are irreconcilably different from Prediction Markets (PMs). My experimental method for funding public goods with PMs has features which render it incompatible with crime. Furthermore, markets would generally present an excessively-complex, risky, and convoluted form of criminal financing. Truthcoin presents a (peaceful) alternative for accomplishing ideological goals, which features greater persuasiveness as well as lower cost. I conclude with a short discussion covering [1] recourse for those affected by AMs (of any kind), [2] features of Truthcoin designed to amplify the inherent impracticalities of AMs, and [3] the (necessarily relevant) total net effect of Truthcoin on political assassinations, general crime and general human welfare.

The linked paper does a thorough job of explaining this scenario, but to summarize: yes, in the absence of other market participants you might try to bet on low-probability events and make them happen. But because it's a public market, anyone can see the anomalous "predictions" you have made, and "free ride" on or even "front run" your position, including your accomplices (who can also sabotage you). On net, your position will be mostly consumed by arbitrage, your payout will be inconsequential compared to the capital required, and the probability "returned" by the market will remain accurate.

If you want to know how the decentralized oracle problem is solved, scroll down to “dispute round” in this paper https://www.augur.net/whitepaper.pdf

Flash loans are one example. These are zero-collateral, low-fee loans that are paid back at the end of the transaction. If the loan fails to be paid back, the whole transaction that made it unwinds, which means the risk to the lender is zero despite the possibility of loans being millions of times greater than the borrower's assets.

They've been used mostly for questionable purposes (largely exploiting oracle attacks on poorly-written DEXes), but there's something very interesting about completely removing capital availability from the constraints on a market participant. It means any arbitrage opportunity can be exploited by anyone, regardless of how much it needs to be levered up to be worthwhile, and so basically democratizes financial profiteering.

> I feel embarrassed to have to ask this > which one is the real trend and which one is the fake trend

As is usual with these kinds of YC posts, a lot of clarifying detail is handwaved away in the name of expediency and clarity of expression. "used a lot by a small number of people, or used a little by a lot of people." -- what a great heuristic! And certainly not useless.

But it doesn't help you figure out that learning how to figure out that difference when the entities are not at rest but in motion is the whole challenge and art in the first place. Sadly, I guess this is a good way for you to answer your question. Nobody outside of an early adopter ever had a practical usage for a Dapp, which is how you know the idea is a solution in search of a problem.

I think usage pattern is too idea-dependent to be a key metric, returning customers vs impulse buyers are both valid targets to aim at and often are just phases within a product lifecycle, like impulse buyer are those pushing the initial growth, returning customer those that sustain it.

I think a good idea moves on two different coordinates around usage patterns (actually more, but two in the context of users):

a) the idea has some value before the network effect

b) the idea has value even for those users that don't contribute any data

and yeah that rules out most low hanging fruit social network, but yeah, social network aren't about ideas, are about how much marketing you can push behind them anyway.

That's a good question, but it probably depends on the state and speed of adoption along with the diversity of cohorts? Like Beanie Babies became popular quickly, but combusted quickly. Facebook became popular quickly, and obviously stayed popular. To me, it seems better when multiple groups with little overlap are slowly building momentum rather than a gravitational pull quickly causing everyone to pile into using something.

at least for your specific example, I would argue that "facebook apps" never took off. facebook games did, but they are a qualitatively different class of application that just happened to use the same platform that the facebook apps did. multiplayer online games were always popular, and facebook happened to implement them well enough to catch on within the platform.

The quote from the mom test comes to mind; “before we can serve everyone, we have to serve someone”

"Dapps were used a lot by a small number of people as of a few years ago (think the cryptokitties trend),"

Genuine question, were they? I've yet to see any data on the FREQUENCY of usage of these dApps. There was certainly a lot of buzz.

Yes, CryptoKitties was so popular that it brought the whole Ethereum network to its knees. It was very short-lived however.

I still use dapps almost every day