Someone had an idea originally and certain decisions were made about that approach.. those that best adapted to the conditions of the time were successful... rinse and repeat over several decades.
Those ideas that preserved the past were more likely to succeed because they preserved the ecosystem that already existed. Ideas that diverged too much from the existing successful tech had a huge obstacle in their way: they had to recreate all of the existing solutions in their new model. So even if they would lead to a better solution over time, they may never get past that initial roadblock. And the longer we stay on the path, the bigger that roadblock becomes.
If we are all thinking of ideas around the existing model and assuming all of the existing assumptions, then we end up with similar solutions (fashionable solutions). We've reduced the solution space, leading to a limited number of solutions.
Perhaps groundbreaking/valuable tech can be found by questioning those existing assumptions; identifying where tech is still built on assumptions that are no longer true; or reexamining past solutions to see if they can be solved in better ways today with what we've learned since then...
This kind of relentless indifference to conformity is a rare quality in people.
The expected outcome is increased human resource so it somewhat offsets the economic costs.
One caveat is that this process requires the willingness to rethink your own established models, and recognising when they might’ve been outdated or incorrect takes some self-discipline.
Even the nuclear bomb was fairly incremental, with truly a big bang in the end.... (pun intended ;)
With waterfall-kind of planning, that would work. Unfortunately it got a bad name in the '90s and early '00s. The wild-west of inexperienced developers, who needed to solve business problems they didn't understand, combined with clients who didn't understand technology.
I believe this is a great period to do bigger things, and you see this happening with for example SpaceX, Tesla and Apple: space is happening again, companies are creating super specialized and complex ICs, there's actual business value and consumer value being produced, and projects are reasonably on time with these companies. This in constract with companies who, what it seems, do more of an iterative approach: Facebook, Google, Amazon. No huge innovation going there.
What follows is my own personal views on current technological evolution, which may be wrong. Regardless:
Electric vehicles have been evolving extremely rapidly, and now they are viable enough for most privately owned passenger vehicle use cases. Through continuous evolutionary investment, they may make their way to some commercial vehicles, maybe even small planes. But the chasm is so extremely large for large commercial vehicles like airliners and cargo ships, that we would need massive (several orders of magnitude) technological improvements in energy density in order to even start considering them.
SOFCs are inferior to batteries and supercapacitors from an efficiency standpoint. They may never be the power system of choice for passenger cars or other small scale and lightly used systems. But the thing they have going for them is their ability to evolve as the ecosystem evolves. They can run on diesel fuel, JP8, even crude. They can run on biodiesel, or renewable ethanol or methanol. And they can run on pure hydrogen. They can even run on a mix of all of those fuels. At every path in the transition to hydrogen, there is a viable intermediate state. For this reason, I think you'll see fuel cell powered cargo ships and airliners long before you see battery powered.
Not really; wheels aren't an improvement in the first place so there is no need to explain why they didn't develop. Note that it's easy to make robots that use wheels, and difficult to make robots that use legs, but we make legged robots anyway so that they'll be able to handle environments other than dedicated roads.
(More recently, we make flying robots, sidestepping the issue that we don't really know how to do legs well.)
How useful do you expect a set of wheels would be on the tundra? Or are you conceptualising a theoretical monster truck human? We'd be better off with tracks...
You’re so right! There are great examples of companies that preserved the existing ecosystem of solutions (Facebook with PHP, Microsoft with C++, Google with Java and C++, maybe also Amazon with Perl 5 and C++).
There are also great examples of successful companies that in a certain sense did or had to do things in an entirely new model (WhatsApp with the Open Telecom Platform, F5 Networks with their data center FPGA-powered hardware load balancers, Tesla and SpaceX, and so on).
There are also great examples of companies that did both, where they preserved an existing ecosystem but also added another ecosystem on top to fix flaws (NeXT Computer with Smalltalk and C, Stripe with the modern version of that as Ruby and Go, Twitter with Ruby and Java, so on).
Microsoft is the best example of a company that can succeed even with what seems like a crazy choice. Almost everything being written in C++ sounds insane to the point of being business suicidal, but they worked hard enough to make it work. If Microsoft can succeed in the ways that they have with C++, and if Facebook could succeed for as long as they did with PHP, and Instagram can allegedly succeed with Django[1], then anyone can succeed with anything as long as you can endure the stress that a seemingly peculiar decision might cause you. And those standard ecosystems that they chose all had immense benefits, they just also had some pretty immense weaknesses as well (Powering 500 million daily users primarily with Python scripts?? That just seems pretty crazy to me, if it’s true).
[1]Still very unclear about how much Python actually powers the web services behind Instagram. <https://instagram-engineering.com/tagged/python>.
People make too much of a fuss over the differences between programming language. There are some important differences (whether or not memory safe is a big one, but it is hardly critical to a succful bussiness) but by and large the difference between using say php, python, etc for your program is superficial in the extreme.
Meaning: the rareness of disruption is to be expected, because most disruptions fail in stifling silence.
Aside: This SMBC shall be long etched in my memory: "All the best work has been done over here! [...] The funding is here too!"
---
Here's a teaser challenge for someone who has access to HN comments data for analytics: How often does an early comment come to dominate the discussion, rather than a better comment which comes later? Assume that there is some IID stochastic process which generates comments of varying quality (play with your distribution of choice), and plot what the time distribution of "best" comments looks like under that model -- compared to the time distribution of most upvoted/discussed comments on HN (can try different composite metrics here too).
---
Provocative meta question for PG: Are (venture-funded) startups the "fashionable" idea that is over-subscribed? How would one know? (suppose we restrict the discussion to Silicon Valley)
Most of the time new products evolve from the current state. Sometimes there's pent-up demand/frustration/hostility that allows for a clean break to a new solution (even if that solution has just as many problems, at least they're not the same problems).
Probably nine times out of ten people are looking for a better buggy whip. Without any other information, improving what you have is statistically the better bet. You have a very strong predictor, but no guarantees.
> I've seen a similar pattern in many different fields: even though lots of people have worked hard in the field, only a small fraction of the space of possibilities has been explored, because they've all worked on similar things.
Anyone want to step in with some examples? Without them, the thrust of the essay seems to be: "If only other people understood what problems are worth working on! Especially in the well-studied areas of essays, Lisp, and venture funding! Too bad they do not. Well, goodbye."
In machine learning, there are currently a lot of people working with neural networks, but relatively fewer people exploring alternative model architectures. So much so that issues specific to neural networks sometimes get framed as fundamental to machine learning itself. I'm personally exploring an alternative class of models called tensor networks with many possibilities for research directions and lots of open questions but only a handful of people work on them. One reason for working on a popular idea is that it's nice to work on a topic where you have many colleagues and know in advance that your model is likely to give good results on challenging datasets.
I think the reasons tensor networks are unexplored is interesting: The tools and techniques for dealing with them build on those for neural networks and the theoretical benefits are not clear cut enough to gain them a foothold over the practical results of neural networks.
Great example:
"One quality that’s a really bad indication is a CEO with a strong foreign accent"[0]
The danger is that I think pg has been "right" about so many things that every time he pontificates about something it's treated as dogma. So when he's "wrong", it'll be ignored. Impressionable people (which likely fits the characteristic of many young tech entrepreneurs) will therefore be lead astray.
[0] - https://www.forbes.com/sites/knowledgewharton/2013/12/19/292...
Would be interesting to actually test that to be honest. Get Paul Graham to set up another site under a fake identity, post the next say, three essays there, then submit 'em to HN under the same identity and see the stats.
Economics: There are surprisingly little people studying very large, grand, topics such an inequality. You've probably heard of the handful of economists that do. It's unfashionable due to the influence of the Chicago school and it's focus on free market principles, as well as many other factors.
Physics: Good luck getting any interest or funding if you are not studying the currently dominant theory in your field (even if there has been no progress in decades).
The two most common fields studying it are: 1) Development 2) Labor Although there are a number of Macro-Economists who have studied the impact of inequality on growth. The reason you may not read about it is that the relationship between inequality and growth is a largely "solved" problem. It is "unfashionable" because of that.
I think being unfashionable is more viable in fields that are meritocratic and don't require popular approval to succeed.
For example it is proven there are infinitely many primes. Are there infinitely primes that differ by 2? By n for any n? Are there infinitely many palindromic primes? Are there infinitely many primes of form n^2 + 1? Is there always a prime between n^2 and (n+1)^2?
If this is true then, assuming there are 200k proven theorems, there would be >1m unproven but readily stated theorems which would mean it wouldn't be too hard to find areas no one is looking into.
Pardon the digression, but the Twin Primes conjecture was proven by a Subway restaurant worker a few years ago.
Agreed.
I think the premise is true, but PG has given no actual insight.
NGO's are a good anti-example. In developing countries we know of working solutions but fashionable rules. I like computers so I'll give them to the poor.
Elon is a good working example, rather than looking at existing tech in space craft he looked at how to make them cheaper using steel and methane.
There's a Just So Story about rapid prototyping that I think fits how to break the mold. I can't remember which competition, but the winner won by working out how to cheaply rapidly prototype working models rather than tackling the problem head on.
So they didn't so much not be fashionable but created a system that allowed for non fashionable ideas.
It’s a blog, not a peer reviewed scientific journal.
But the essay doesnt really seem that revolutionary more common sense. If you want to make an impact, dont work in an oversaturated field. The low hanging fruit is probably already picked and other people will probably get there before you do. But you also dont want to work in a field nobody cares about as noone will care. Working on a problem with proven demand, but seems "boring" and hasn't changed much recently is a good bet, as there is probably new insights you can apply and new contexts that have appeared since last time there was a frenzy for that field.
Perhaps another metaphor is a gold rush. Even if there really is gold in those hills, if everyone knows there is, its probably already too late to go out and buy a shovel.
That said, i agree that there is plenty of survivorship and hindsight bias when it comes to any advice on how to be succesful.
I would put two caveats on this. First, if you are globally in a good position (world's top university, good connections to get into places) you may be positioned to meaningfully get even into a mature field. Second, if you don't want to do much research and exploration, it's reasonable to reap benefits from specialist knowledge of already proven fields. Just keep an eye on other options to transition to if everything eventually crashes.
Even in crowded areas you should benefit from wide knowledge of horizons and fundamentals beyond what is now fashionable. It may mean just more ideas and perspective. I work in NLP with an interest in AI, and I've dived fairly deep into currently out-of-scope things like rules-based NLP, symbolic AI and biological neurons, their physics and simulations etc. I hope this gives me more viability that your average deep learning guy from a moderate background.
That's exactly the opposite of his point, if I understood it correctly!
https://www.bloomberg.com/news/audio/2019-08-16/josh-wolfe-d...
This makes no sense. The concepts of "conservative" and "fashionable" are almost polar opposites. Becoming a musician or an actor is fashionable and the dream of many. But it won't bring any money to most people who choose that career. It's no conservative choice. Instead, it's conservative to go into STEM, law or finance. But those fields are "boring". The pattern repeats inside a field as well. It's conservative to be a cobol coder or a DOS expert, and you'll certainly make money. But it's not fashionable. How come you aren't building a cryptocurrency using self driving car that has a drone port on the roof! It results in the woke fields being overrun with very smart and capable people and capital, while tons of fields that use slightly outdated stuff are ripe for harvest but nobody is around to do it.
That's not conservative. Being conservative means that you only follow a change if it makes sense. If you adopt some new technology, you should do it because it convinces you that it's better, not because it's new, wasn't available before, everyone else is doing it, or any other such reason. If you are conservative, you don't neccessarily end up doing what the majority is doing, as most times, the majority is following some empty hype.
That's because his use of that word is wrong: everything said in that article contradicts that statement.
For most people entering a career in tech over the last 20 years those haven't been a consideration, never mind a choice (conservative or otherwise).
Nowadays the conservative choices in the sense that PG is talking about here are Java, .NET, Python, C++, C, and possibly AWS and Amazon. But these are also what the market demands.
You will get paid, and probably quite well, but they are not adventurous choices.
Do people really think the fields of essays, LISP or venture funding are fully explored?
I don’t think pg is saying they’re being ignored, but rather that there’s a sense of “yup we figured it out, let’s just iterate slowly on it”.
(I will say I can think of much better examples, but I don’t blame pg for picking the three examples that sum up his entire career. He picked those three things when nobody else cared.)
[1]: e.g. https://vitalik.ca/general/2019/10/24/gitcoin.html, https://www.zfnd.org/blog/dev-fund-guidance-and-timeline
If you found something in a big, fashionable field with a huge surface area (cough ai/ml cough), wouldn't this multiplier apply to that fashionable field, too?
So, essentially, yes - fashionable fields will be well explored; the essay is about not getting sucked into fashionable problems in whatever field you're working in.
Particles (individual humans or small groups) move around in problem space as particles with position and velocity. Each individual's movement is influenced by its best known position locally but also the the best known global positions in the search-space. When better positions are found by others, individual change their course (take hints) and move towards them.
The problems would be the same. Too much randomness and it's just a random search. Too much convergence leads to local optimums.
The counter to this is that it'll be harder to raise capital for un-fashionable problems.
If you pitched "ML for sandwich makers" right now you could raise a million bucks because so many VCs are making fashionable bets on ML.
There are brilliant ideas out there, but very few of them are brilliant and require 0 resources to achieve.
People think raising is a signal for success, when it's more like kissing the dice at a casino
The paradox is that nobody really knows what will be the next fashion and similarly nobody knows what's the next worthwhile problem-area to work with.
There's a good reason why fashionable problem-areas are well-researched it is because the results so far have been useful and promising.
If I were to take on Netflix/Disney/Twitch with some new kind of video entertainment product, they'd have deep pockets to fund a competing offering.
The lever of equity might work to attract better talent, but only if you succeed. There's a lot of risk.
Scaling rapidly also means giving up control as you seek capital. It'd be hard to organically grow and go unnoticed.
However, this was justified: even the most promising-looking competitors tripped over their own bad assumptions long before becoming a threat. We saw plenty of novel ideas but they'd always be sunk by a failure to understand the basics of how our market worked - things like trying to put a complicated app with a thousand options at a point in the journey everyone is trying to simplify and time-optimise, etc.
If someone who knew the market well had gone at it seriously and solved hard problems rather than apply the usual hand-waving "tech! blockchain! magic!", by the time we'd noticed it would have been too late to respond. You'd hope more recent incumbents like Netflix or Twitch might be a bit more responsive, but corporate inertia can build up surprisingly quickly.
https://news.ycombinator.com/item?id=21854793
What if all this cloud/k8s / serverless stuff is really all piffle? What if running stuff on dedicated hardware ends up a better solution in some fashion?
I doubt that really is the case. I think most decisions are made based either on anecdotes, or whatever someone happens to have experience with.
It's rare that you can accurately predict the kind of workloads you'll have to deal with ahead of time, and it's even rarer that the people making the decision have experience with multiple completely different stacks.
And I don't really think it matters that much. Some people solve the problem with distributed document stores and key value stores, other people use a big transactional database and just keep putting extra RAM sticks in their server... I don't think there's always an obviously "better" choice.
Are we not eternally one packaging system short of Nirvana?
Edit: Oh, it was originally a tweet: https://twitter.com/paulg/status/1183687634763309056?s=20
Both are widespread.
an example: a fashionable solution for a static webpage is a fashionable static site builder, a unfashionable solution is an HTML page.
devoting more than 20 percent of your time to problems that are unfashionable but dear to you is ill-advised because: a. they don't pay the bill b. they take time and in the grand scheme of things spending time with others on things you all understand is better than being happy alone.
but of course there are exceptions...
his essays don't necessarily have to be relatable, they just have to be useful. he isn't a life coach (though I suppose that is arguable). his expertise is in tech and startups, that's where he's proven himself, and that's the area where his advice carries weight.
Everyone is trying to break the mold at various scopes.
Meanwhile, you're welcome to reproduce your late-90s success at any moment, Paul. We'll wait.
I love Joseph Campbell but his advice to his students to “follow their bliss” may not always be optimal.
I read the AI book “Mind Inside Matter” in the late 1970s, and even though I have done a ton of non-AI architecture and software development, I have also been able to work on AI problems like knowledge representation, expert system, NLP, neural networks, and deep learning starting in 1982. I definitely followed my bliss, but I have never been world class in my profession, but I have enjoyed myself.
The platitude is directed toward those living a life they feel is negative, but who slog through for whatever reason: fear of change, obligation to other people, etc.
Taken as feel-good, head-in-the-sand optimism, of course it's facile advice.
But imo there is genuine wisdom in it, and it's this: You are much more likely to do something well, and to continue to do it until you reach a high-level of expertise, if you don't have to force yourself to do it. People who naturally love working out are more likely to be fit. That's what the heuristic is.
1. The more defined and mature the problem space is, the more the assumptions that underpin that field are taken as a given.
2. These assumptions may become so deeply ingrained that people become effectively blind to the entire range of possibilities.
3. These assumptions frame how the problem/solution space is looked at. Therefore the solution space is constrained by the set of assumptions (about what the problem is , how to solve it, what to do)
4. These assumptions are recursive, in that they are contained within other assumptions. It is turtles all the way down. At some level, someone working in the space may not even understand what the core assumptions are. We have to have these assumptions though. See the next point.
5. The interesting thing about this is: The constraining of the problem/solution space is actually a positive. It enables co-ordination and incremental improvements and refinement. It allows people new to the domain to quickly get productive.
I like to think about it this way:
Picture yourself in a massive area that is pitch black. You are grasping around and can not see much. Someone figures out how to get a tiny fire started. With this tiny fire you get to see a little bit. Using this you can build a bigger fire illuminating more of the area (but still leaving the entire space unexplored). This eventual results in the ability to build a permanent light illuminating a specific corner of this space.
This specific space with light illuminating it becomes highly productive, people can do all sort of things. Like read etc. Yet, there are still areas left unexplored. The light cannot simply be taken across. It takes work, and it takes turning your back towards the current "lit" up space, and taking a step back into the dark. A scary thought for some.
5. Importantly: These assumptions have been inherited from the past. So they existed and were relevant at a specific point in time. They may or may not be relevant as of today. We would have to peel several layers to get to the core.
6. While 4 is a positive, it is also a negative. The idea/areas greatest strength (maturity, constant improvements, efficiencies) is also its greatest weakness (constraining the search space)
To take a step into the dark, is to turn your back to the lit up parts. You have to question the underlying assumptions and see if they are still relevant. If you discover an assumption about the world that is no longer accurate, then you found a new space to illuminate.
To put it in another way, to explore the dark is to shift your perspective on the problem/solution. It is to see with "new eyes". Initially it may be dark, but slowly with diligent work, and passion you could light up a completely new and novel area.
How can a statement like that be made? Is there some kind of authoritative directory of 'the smartest, most imaginative people' being 'surprisingly conservative when deciding what to work on'.
Implied I guess Paul means 'who I've met or who I know of'. So then say that.
It's a big world out there. Who knows what anyone is working on or what they are thinking or have tried and why they haven't pursued it.
This is a bit like saying 'people love their dogs and will do anything for them if they are sick'. Just a general statement of opinion by one person (and generally accepted as being correct) but based on not anything even close to being scientific and/or backed up by any actual data. That part is fine. But if that is the case state it as such and not some absolute. Why does this matter? Because when someone like Paul writes something it will be taken by others to be some kind of important thought or fact.
I don't think anybody who reads his essays is looking for any scientific report based on facts. They are looking for some sort of confirmation that they are not crazy when they have similar thoughts.
It was more charming when he had to work hard to make his points known.
But hey, fame, right? Just famous people things.
There's so much more to say in this case, though! How do you avoid the traps? Waving a wand like "Just love something" leaves far too much to the imagination. Pointing at a prior essay at loving your work is helpful, but different.
Often, you have to actively offend people in order to find good problems to work on. The idea that people have devoted their lives to the wrong thing is inherently offensive to them. That's a point not covered here.
For example, I imagine that a lot of people who've studied 3D rendering for their entire lives are about to feel very outdated the moment neural network renderers displace them. And that's also a good counterexample to the point that "Often, the best place to search for new ideas is a place thought fully explored." It might often be true, but it's not always true.
And then there are the in-betweens. Bitcoin was in a field both thought fully explored (crypto + finance) and also unexplored, in a certain sense.
Some of his essays, no one else could have written them and brought a fresh, nuanced perspective.
These couple paragraphs wouldn’t get any attention if not for the name of the writer. Maybe that’s fine—great writers have their share of banality—but does reveal how susceptible we are generally to brand name over substance.
Paul is authoritative on many topics. General thoughts about life and people are great to hear what he thinks. Why not? But he is no more special than 100000 other people who have no audience. Does he know this?
I've often thought he should do some A/B posting with his thoughts. Write something and then randomly decide whether to put it on his blog or some other place and see what the interest level is.
This itself is in need of a lot of conversation. I think a really ugly part of SV startup culture is, in my opinion, "it's okay to offend and hurt people, and it's sometimes demanded of you to..."
And I think we can be so much better than that. Being "disruptive" this way is just a lazy excuse.
And some not technologies. I’m a lawyer and recently there has been movement toward letting people who aren’t lawyers do things that are traditionally considered the practice of law, like creating and filing certain legal documents. Most of the documents are ones that are already done almost exclusively by paralegals and secretaries, but I regularly see articles arguing for restricting this couched in the language of protecting people from malpractice. It seems to me to be a case of lawyers who worked hard for their status being offended by change that would lower the cost of legal services and their status. I don’t think their offense should be taken into account, rather I think we should make changes to benefit people who currently can’t afford legal services.
I also don’t personally think there is any shame in spending your life on something that becomes an anachronism, so I don’t think much of the offense taking at new technology is warranted. Once upon a time, 90-something% of people were farmers, after all...
At the other end, it is laborious to go through a well researched long-form articles.
I really like Drew Devaults blog [1] - I think he strikes a great balance between tweet-like articles (pg, daringfireball) and something extremely verbose (Stratechery).
It's not just fame, HN is PG's site. It would be almost rude if his posts didn't show up high here.
