The idea that a static world implies monotonically increasing confidence in your beliefs is really shaky, logically. Hill-climbing in a landscape of belief-fitness could easily arrive at a local maximum. To further improve your beliefs you might need to become less certain.
Apart from that, how can something (confidence) increase in a static world? PG has clearly tossed this essay off without serious thought or discussion. A waste of time.
The idea that a static world implies monotonically increasing confidence in your beliefs is really shaky, logically. Hill-climbing in a landscape of belief-fitness could easily arrive at a local maximum.
Exactly, and it's common to see that in many fields. People are persuaded by things that were bad ideas when they were new, and aren't swayed by new evidence to the contrary.
To further improve your beliefs you might need to become less certain.
Isn't that sort of the point of the essay?
Apart from that, how can something (confidence) increase in a static world?
I don't understand. Didn't your previous point argue exactly how hill-climbing could happen, and isn't that a (bad) example of confidence changing?
Or are you really arguing over the semantics changing beliefs in a static world? I hope you aren't; otherwise, like you said: A waste of time.
This is definitely a waste of time but my point is that you don't need to have a changing (dynamic) fitness landscape for monotonicity to be insufficient in the long term - you could get stuck at a local maximum _even_ in a static world.
The second point is that drawing a line between the "world" and your beliefs about it is a pretty dopey dualistic way of thinking. But hey, Graham doesn't care that there might be real knowledge and valid experience of these matters in the accumulated history of philosophy or psychology, he just shoots from the hip with some flaky positivistic analogy with computation... Ridiculous.
This assumes you can only hold beliefs that are adjacent to your current beliefs (i.e. the graph of your beliefs over time must be continuous). I disagree with that assumption.
You can explore ideas that you don't hold, and these can be quite varied from your current beliefs. Once you find an idea that you put more confidence into that your current beliefs, then that idea becomes your belief.
The local optimum from high-climbing doesn't constrain you in a world where your beliefs can teleport from one hill to another.
> If you're sufficiently expert in a field, any weird idea or apparently irrelevant question that occurs to you is ipso facto worth exploring.
My proudest creations all came from these weird ideas. Now I listen to them very carefully and nurture them. They're starting to cohere and turn from random ideas into an ideology. It will be fun to make that real.
I work with making technology teams deliver faster. It's nothing as complex and tricky as what Paul's doing, but I also get to see lots of teams and make correlations between people and performance.
One of the things I've noticed is that over a long period of time, the more arrogant and certain team members are, the less likely they are to deliver. Folks that are always learning have a tendency to kick ass. Not so much with folks who think they've already arrived. I read in an article on HN a few weeks back that Google calls this "intellectual humility" and looks for it as a hiring trait.
I hear many startup people talk about "domain expertise"
I wish someone would write about that.
What is an expert?
What is a domain?
How do I become a domain expert?
Am I an expert at something I take for granted?
I love Paul Graham's essays but there's a lot of implied logic and jargon. Maybe he should have a few non-startup, non-business people proof read them too?
One of the essential tensions of being a software developer is the fact that software isn't really useful on its own. It has to do something. That thing it's doing is the domain. Whatever very-specific information that the software system is organizing and passing, more specifically.
For example, I maintain an e-commerce system. More specifically, I maintain an e-commerce system built for a cosmetics company. E-commerce is a field, cosmetics is a field, cosmetics e-commerce is specifically the domain I need to become an expert in. The more I can differentiate cosmetics e-commerce from other kinds of e-commerce, the more well-tailored I can make my app. And the more well-tailored I can make this app, the better I'll be able to understand the more general case of making custom e-commerce systems for other industries.
The more I go through this process of understanding the specific needs of the industry, the more domain expertise I'm building. Domains can subdivide as much as industries subdivide. Say the cosmetic industry started to split along product lines, then I might start gaining insight into how lipstick companies operate, and be able to tailor a product offering straight to them.
The domain doesn't have to mean commercial industry. You might have domain expertise, say, in XML parsing. Then you could wax poetic on the differences between Nokogiri and say, Chrome's internal parser.
So, essentially, it just means "expertise". More specifically, the type of expertise that allows you to customize a software system around it. Which can be in practice quite different from normal expertise. The guy who devises our makeup formulas is definitely an expert.
But he can't design the system to actually manufacture it. The guy who writes those systems might not understand the actual formulas, so the two types of experts have to work together.
When startups talk about the need to be a domain expert, it means getting out of just pure programming and into the actual meat. Rather than hire an expert, to become one yourself. This not only allows you to reduce staff, it also makes it easier to make better, more custom solutions. The Soylent guys have to make the system to manufacture their product, but they also must understand nutrition, food, the regulatory environment of the food market.
This not only allows you to reduce staff, it also makes it easier to make better, more custom solutions.
I would even go as far to argue some level of domain expertise is almost required to make solutions. Even with perfect requirements (oxymoron I know), a programmer still has to make a ton of tiny domain driven decisions while writing a piece of software.
As you mentioned, software on its own isn't very useful. A programmer who learns even general business is much more useful.
I really dig this post. As it happens, the blending of domain expertise and software skills is a big reason I moved into devops. I wanted to make my domain expertise as portable as I could, while relying heavily on my software skills. Your cosmetics e-commerce app scales according to the same principles as most others', using the same techniques. Which is cool, because it increases the amount of useful knowledge in my head and allows me to apply it to a lot of different problems. (I call my role "being a multiplier", and it's a lot of fun.)
I really feel this. As a founder, my biggest strength is that I'm a domain expert. Right now, I'm doing coding because I don't have money to hire someone else to do it, and I'm perfectly capable of it. But I feel it's not a good use of my time. Any decent programmer could do what I do most of the time right now. What they can't do is understand the domain as intimately as I do.
So as soon as there's money to hire programmers, I'm getting out of the coding biz and hiring others to do it. I want to focus on working with my customers, understanding their needs better than I already do, and guiding programmers to build for those needs.
examples of domain: tax law, fashion, playing guitar
expert is used in the sense of the standard definition in any English dictionary
one could be an expert at programming, and though that's a valid topic and craft to have expertise in, its not a domain in the sense of "domain expert" you're asking about
A set of disciplines, or majors, other than computer-science (or alternatively something other than core programming/software-engineering skills). The word is typically used in CS/sofware-engineering circles.
An (arguably crude) example:
Think of who designed an mp3 encoder/decoder? A programmer? Does becoming a good programmer and/or getting a typical 4 year CS degree, sufficient to come up with an audio encoder/decoder?
No. You have to understand audio signals, their properties, learn how to analyze them, etc. That requires knowledge from the disciplines of physics, and electrical engineering (e.g., digital signal processing), maybe even some biology of sound. And if you learn that, you become a part-programmer, part-acoustic-engineer. That "acoustic-engineer" bit is your "domain expertise", i.e., something you need to know beyond CS to get your work done.
You might say, "why do I have be an acoustic engineer? why can't I, as a programmer, work with another guy, who is not a programmer but an expert acoustic engineer?". Yes you can. In that case, you're being the software guy, and he's being the domain-expert.
Sometimes collaboration between a software-engineer and a domain-expert works. Sometimes it helps if the software-engineer himself develops domain-expertise (so one guy playing two roles).
Domain expertise is not limited to one domain only. It could be no-domain (a pure CS programming, like database, web-development etc, although you could call them domains within CS), or a domain consisting of multiple disciplines (audio-processing comprising physics, DSP), or multiple domains (acoustic-engineering plus submarine-engineering, for example), and so on.
A domain expert or becoming one is what makes a startup profitable.
What is an expert: Someone or something recognized as such by other recognized experts and trusted by a significant number of those who are not.
What is a domain: A limited area of human knowledge small enough that it can fit in one person's memory but wide enough that when applied it solves multiple useful problems.
How do I become a domain expert: By doing. First learning. Then making. Finally advancing. Unless of course you are creating a new domain which skips to step 3.
Am I an expert at something I take for granted: No. Most US adults know how to drive and do it semi-(un?)consciously. They are not expert drivers.
One interesting thing about beliefs in a changing world is how you deal with contradictory beliefs.
I see "beliefs" as an optimisation that avoids having to return to first principles for every single decision.
That speeds things up, but is obviously dangerous.
Sometimes it's possible to have two beliefs that contradict each other. Usually this means they need reevaluating, but sometimes it means there is context in which each is true.
Reasoning in the face of contradictory beliefs is one of the most interesting things about "knowledge", and something that humans can do surprisingly well - at least until we realise we are doing it.
1. We have some initial beliefs
2. We use it to make decisions
3. We receive some new data
4. Given the new data, we confirm/change the beliefs
5. We can make different decisions
6. Go to 3.
The initial beliefs may or not be true, the key to get as much accurate data as fast as possible.
A better algorithm would be to ask other systems beliefs and compare it to ours too (eg: neuronal network).
They definitely are, and there shouldn't be anything wrong with having beliefs (per se). What matters is how you react when you receive new information that challenges those beliefs.
Do you adjust your beliefs or do you staunchly adhere to them?
re this : "if you write about a topic in some fairly durable and public form, you'll find you worry much more about getting things right than most people would in a casual conversation."
i wonder what the bar here is - certainly more than blog posts / self published websites as those are fairly close to casual conversation imho.
glad to see pg picking up the pace on the writing!
yeah - I saw that, but seems to me it only barely meets the "fairly durable" test if you assume cache's and other history keeping mechanisms.
I wonder if what is additionally required is : "someone you respect asks you to write on a topic". In PG's case, his audience might substitute, for others an employer or publisher might be what it takes.
I am noticing a trend in Paul's essays. The last two were about changing paradigms. The first was about how a second read of a book leads to a different understanding, and this essay is about not letting your previous beliefs close your mind about things that can change.
I like though how he circles back to two fundamentals. One, human nature doesn't really change, and Two, betting on people is a better indicator of the future than betting on ideas.
Probably the biggest mistake people make is looking for confirming evidence instead of using evidence to determine the relative likelihood of competing hypotheses.
Even a large body of confirming evidence doesn't prove anything because that same evidence could be consistent with other hypotheses you never considered. You need to evaluate the diagnosticity of that evidence. (i.e. a patients high fever has almost no diagnostic value in determining what is wrong, because so many illnesses are consistent with that evidence.
Whereas, one shred of evidence can disprove a hypothesis completely. Which is why good doctors rule out things to identify the most likely problem.
These ideas PG is talking about are, IMO, some of the most important things for founders to understand.
Don't look to confirm your theories, look to disprove them. Don't accept lore, challenge everything. Most importantly though, always evaluate competing hypothesis to prevent you from just trying to confirm what you are looking for.
> Don't look to confirm your theories, look to disprove them.
I use this framework, though learnt it by reading Fenmayn's writings, when there's a debate as to use tool A or tool B to get X. And it's been very effective so far. I have a vague sense that B is better and a colleague is particular about A. Earlier I used to look deeper and deeper into how B is good at doing X. But realized that I wasn't convincing enough. Then I remembered Feynman's saying to the effect "the best way to prove something is to try very hard to disprove it". This has helped quite a lot in coming up with objective comparisons and use right tool for the task at hand; or in general to come up with the right solution the problem at hand.
Updating one's beliefs to get more accurate priors is something we should all strive to do. There is no such thing as having two beliefs of equal weight that contradict each other - that's an illusion created by either applying a different definitions to each belief, or a refusal to admit that you believe one more than the other.
...
> "The winds of change originate in the unconscious minds of domain experts."
I would also add that domain expertise is when one's internal model / map matches the territory. When this happens, one's intuition is accurate and one can notice patterns that other people can't, or one can make predictions that turn out to be right.
I wonder if a lack of domain "expertise", but a general understanding of the domain instead, can sometimes be an advantage in generating domain-changing ideas.
There's definitely a risk as a domain expert that you'll take things for granted. Henry Ford is a good example of this, he was never an expert in horse-drawn carriages but he changed that domain drastically.
>There is no such thing as having two beliefs of equal weight that contradict each other - that's an illusion created by either applying a different definitions to each belief, or a refusal to admit that you believe one more than the other.
Of course there is: P(A) = 0.5 and P(~A) = 0.5. Don't be silly.
Great read. I would go a bit further and say that the goal is not to be an expert and is not related to start-ups. The goal of being open minded and sceptical as described in this essay, is to make sure one do not embrace sheepishly some ideas that will be considered obviously stupid twenty years in the future. For example in the 50s in Europe most intellectuals were communist or pro communist. Only a very few were neither reactionary nor communist. The goal, for me, is to be of these few. In the US history, it would mean to have been e.g.among the earliest abolitionists.
This reminds me of Arthur C. Clarke's "Hazards of Prophecy"
> When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.
and Asimov
> Those people who think they know everything are a great annoyance to those of us who do.
beliefs, like stereotypes are shortcuts in cognition. We should always be applying the most powerful tool we have developed in the history of humanity, the scientific method.
I've always wondered if Clarke's statement has anything to do with elderly scientists in particular, or if he's just commenting that most things are possible.
Almost certainly elderly scientists in particular; if he meant a general statement then the juxtaposition with the previous sentence is very odd. Also there's this (paraphrased from memory) quote in Rendezvous with Rama: "even after hundreds of years, science never solved the problem of having elderly, conservative scientists at the helm of its governing bodies".
An elderly scientist is often one who did all his work in an old theoretical paradigm, a paradigm whose death is eagerly awaited by all the young PhDs looking for jobs with a new paradigm in their thesis.
Inmy limited experience, I found (to my surprise) that most young scientist are also reluctant to think out of the beaten track (even if it's just to play devil's advocate in a discussion).
I guess it takes somekind of rebel todo it, no matter what kind of human group you are talking about.
It really isn't a paradox. Your strength being 'always right' is a strength, but relying on it without proof is a weakness. The scientific method delivers an expert to their position, but it also keeps one in their place.
We always need to show our work. The fact that an expert can crutch on the Markovian Weather doesn't mean I should trust them in any way based on their word. Trust me, says the expert, I am always right. There are those who are right and there are those who are right with proof.
The outcome doesn't make argument, the argument and the outcome does.
I spotted the following in my usual reading this morning:
>"Until this point, I have been describing the stereotyping process as a
negative force for individual and team functioning. However, this
process actually stems from an adaptive, often functional psychological
process. Mental heuristics and cognitive shortcuts enable us to process
information without conscious deliberation: they fill in the
informational gaps we often experience when making decisions. In other
words, habits of mind help us to save brain power for more difficult
tasks. Joseph Pieper's classic "Leisure, the Basis of Culture" sets
out this theory well. In the 1980s, people thought we could create
expert systems by interviewing experts like brain surgeons or oil
exploration specialists, and creating a rule chaining prolog environment
that would recreate their decision-making ability. The problem was the
experts did not know how they knew what they knew. That is, experts are
creating associations between disparate experiences and pieces of
knowledge, using the subconscious brain." [1]
The quote above reminded me a bit of the previous "How You Know" essay
by Paul Graham [2,3]. At times it's disheartening to realize the
discrepancy between how much I read, and how much of it I can actually
remember, but it's nice to know at least some of it manages to stay with
me somehow.
In this new essay, pg says:
>"When experts are wrong, it's often because they're experts on an
earlier version of the world."
Stereotyping and similar classification are essentially heuristic
modeling based on past experience. Some substantial part of our mental
models are based on "Unknown Knows," or better said, our mental models
are based on experiences we cannot consciously recall but still somehow
know unconsciously. Even if we keep the models updated to have a
"current version of the world" through seeking new experiences, we are
still often incapable of either anticipating or handling new surprises.
Maybe people who regularly succeed as startup investors (i.e.
anticipating and handling surprising exceptions) have learned to retain
some degree or form of suspended disbelief when applying their models
to the world?
It's trivially easy to suspend disbelief for the sake of entertainment
like enjoying a fictional story. Intentionally escaping reality for a
while is an established habit for most people. On the other hand,
intentionally suspending a bias, belief or model in the real world seems
to take more effort, but that's possibly due to my own lack of practice.
"... intentionally suspending a bias, belief or model in the real world seems to take more effort, ..."
For most people this seems to be the case, but I have found a way to do it regularly. I evoke empathy before logic.
By being empathetic and transporting yourself into another person's values, you then use your logic to deduce the answers and details from their point of view. This allows you to escape your own details.
I use this often. But I can never start with logic, I have to start with empathy. From there, I can then calculate.
You can do this as a thought experiment, and you'll find it becomes trivially easy.
I work with customers often, and when I work with them I just pretend I work for them. My desires and goals change dramatically from what I would suggest for them if I didn't pretend in my head that they hired me.
I also tend to distance myself from... caring?.. about outcomes. The less I care about what the result is, the better I am at changing my perspective.
This suspension of disbelief, and the practice of it allows you to quickly shift your entire world view. The shifting isn't really the hard part, the allowing yourself to shift is.
I hope this is clear, I feel like this is difficult for me to communicate well. I think a back and forth conversation would make more sense for this topic.
Expertise is a weird thing. In today's world, one that changes so rapidly, it's very hard to become an expert in something and have the value of that expertise remain valuable. It used to be that expertise in some field or area would last generations, and guild-like systems of master/apprentice developed in response. Now you can't even make it through a degree program before the things you started to learn about at the beginning become virtually obsolete by the end. Watching the front page of HN and you'll see entire frameworks and languages come and go virtually as quickly as the seasons. When I was starting out, it was fully possible and reasonable to consider becoming an expert in COBOL and one or two kinds of mainframes and make a career of it.
Expertise today requires that you start to think and operate at a higher level of abstraction in order to cover what you know, and where your knowledge might go. Or it might require you to make connections between old ideas that are still valid and new ideas as they appear. You might even be able to change career paths entirely, but things you learned as an expert in your old one might make you special and unique in your new one -- somebody with rare and unique talents. Like automating business processes in a non-technical career.
I've been thinking a lot about beliefs and belief systems recently. I'm not quite sure I have any meaningful or concrete conclusions yet, but I'm pretty sure I'm going down the path of thinking that beliefs should generally be treated and handled with extreme caution, that they're some kind of holdover from an earlier evolutionary stage -- a vestigial kind of proto-thought process wedged somewhere between pure animalistic instinct and enlightened reason.
They aren't inherently bad just like instincts aren't inherently bad, the instinct to pull your hand off a hot stove is good for example. Perhaps they some kind of optimized way of thinking that evolved when we had less hardware to compute with. An estimation system that got us through hundreds of thousands of long cold nights, alone in the world of beasts...when we couldn't quite think through everything. But now we can, and this old wetware technical debt kind of gets in the way now and again.
It tends to make us rigid and lazy in our thinking...why think when believing is less work?
Maybe it's possible to hack our beliefs be believing in the right things, by believing in reason, or believing that things should be changing, believing that we should evaluate and challenge ourselves from time to time, believing that we should feel like we're just on the edge of discomfort and force us to use this new better and harder to use equipment we've been gifted with.
> Watching the front page of HN and you'll see entire frameworks and languages come and go virtually as quickly as the seasons. ... Expertise today requires that you start to think and operate at a higher level of abstraction in order to cover what you know, and where your knowledge might go.
I'm glad you wrote this, because so many people bemoan the first quoted bit and don't take into account the second. Sure, FooFramework is out of fashion, but the underlying knowledge that a good programmer, one I'd consider an "expert", persists. The knowledge that makes a person a good programmer, the real skills involved rather than the minutiae and the whatevery bits involved in translating that knowledge and those skills into a particular use case, have personally served me well (so far) in web development, mobile development, devops, and game development.
> it's very hard to become an expert in something and have the value of that expertise remain valuable.
That may be true in programming, but that's not the only field. Much of my mechanical engineering knowledge dates from the 19th century and earlier and is still valuable. Even my FEA knowledge (which relates heavily to programming) is still good. Practically all my undergraduate knowledge is still valuable, let alone my Master's and PhD knowledge.
It's quite clear that you're approaching this strictly from a single viewpoint. If you really want to think, you need to broaden your view.
Even in programming there's still things that change at a slower pace. Sure there's a lot of things changing in regards to the web, but areas such as numerical methods change more slowly.
Yes, of course you are correct in many senses. But I'm hard pressed to think of a field that hasn't seen tremendous change in just the time I've been working professionally let alone my lifetime! Even the art of cuisine has changed so much that some chefs I know are struggling to keep up late in their careers - it seems like all the top chefs are new and almost impossibly young. It really used to be that you could learn your skill, unchanged from your teacher, and do it for your entire natural life, passing it along to your proteges who would then do that same skill exactly the same for their entire lives.
So while I'm pretty sure the basic physics you use is largely unchanged, I'd bet that a mechanical engineer for even a generations ago would have trouble recognizing much of what you're doing today. We have materials, measures, testing and production methods available to us today that would have looked like magic to somebody in the 60s. It's like comparing a flint-lock musket to a modern rifle. A gunsmith from the flint-lock era would know the basics, but his expertise would be obsolete.
Even in software, most of what I learned in my undergrad CS program has stayed refreshingly relevant, but the act of writing software itself has completely transformed a half dozen times since I graduated. It means I can sit down in a new environment and learn the syntax, practices and idioms of a language and dev environment relatively quickly. But in many cases you quickly start to feel like you're treading water at the same time, having not 20 years of experience but risk having 1 year of experience 20 times unless you carefully cultivate your expertise to prevent that.
My civil engineering friends seem to have similar feelings, lots of the physics and basics they learned are still relevant, but the theory and implementation, especially the huge changes in Urban planning are having tremendous impact on their ability to cultivate expertise.
I have a landscape architect friend who graduated hoping to design golf courses and parks and who now spends most of his time working on energy efficient certifications (LEED) and planning for new construction and retrofits. Something that didn't even exist when he entered his line of work. Instead of becoming an expert in park design, he's become an expert in LEED certification.
Even my pastry chef friend left pastry school, where she intended to make traditional French pastries, and entered a world expecting innovative techniques and ingredients catering to the allergen/wheat sensitive/low carb/vegan/no GE/whatever crowd of the month. She doesn't just make scones, she has to make 6 kinds of each scone suitable for everybody's special dietary decision.
You're right in that they're all still doing mechanical physics, looking at water management, and baking things in an oven. Expertise in those subjects transcends. But being an expert in their fields is impossible because the expression of what their entire fields are are being redefined almost every day. What they're becoming experts in, and something their predecessors didn't have to do, is riding the wave of change.
This is a fundamentally different kind of expertise that maybe only ever existed in human history during times of war, or great societal upheaval. "May you live in interesting times" is a curse, not a blessing. It's to be seen if this is simply the new state of things from now on or a time of upheaval that we can't recognize and things will settle down at the other end.
Even on the web, imo a large portion of the interesting stuff done today is strongly based on decades-old expertise. For example, Google's V8 JS engine was built largely on techniques developed in the late '80s and early '90s for implementing Self and Smalltalk, by people who had developed their expertise in the Smalltalk world. That expertise didn't remain static, of course. But it was highly relevant and formed the basis of the further developments, which is why Google specifically went out and hired those people to develop V8, because having expertise in implementing Smalltalk efficiently is useful if you want to implement Javascript efficiently.
If you approach things very shallowly, learning how-to-set-up-technology-stack-X type bags of tricks, those become obsoleted quickly, but lots of technical expertise is quite durable.
"In today's world, one that changes so rapidly, it's very hard to become an expert in something and have the value of that expertise remain valuable."
In my experience, a world that rapidly changes makes it very easy to become an expert. There are so few in whatever is new, and all it takes is 1 year of effort to become an expert in that field.
Having the expertise remain valuable is more about luck and intuition, but I don't really care one way or the other if what I spend my time on remains valuable. The valuable part is being able to learn something, and developing the skill to learn tasks. If you repeat this with a different thing for every year of your life, you will become an expert at becoming an expert, a skill that far exceeds knowing any one domain.
I don't, and have never desired what you seem to desire, learning an expertise that lasts generations. Which means, learning a domain that lasts for generations.
The very act of being an expert at becoming an expert makes you impossibly valuable to any corporation that exists. The actual details of the domains is completely irrelevant.
> all it takes is 1 year of effort to become an expert in that field.
I think we may have a deep disagreement of what constitutes expertise. 1 year of effort in anything is at least an order of magnitude less effort than what I'm talking about. I would consider a smart person with a year of solid effort "functional" but not even capable of sniffing what it takes to become an expert. That level of grokkiness is simply not attainable in that short of a period of time.
But maybe you bring up a good point. Maybe the word has simply been redefined these days to a much weaker definition. Maybe a year of practice, in this sound-byte, short attention-span laden world, is good enough and our collective standards have fallen. After all I see job listings for "senior this or that" job all the time requiring 5-7 years experience in the field.
> The very act of being an expert at becoming an expert makes you impossibly valuable to any corporation that exists. The actual details of the domains is completely irrelevant.
I agree, and I think this meta-expertise really is a new thing. A new kind of paradigm of thinking that really had little analog in the past. Even the idea of a Renaissance Man, being an expert in multiple fields simultaneously, isn't quite the same thing.
"In practice "sufficiently expert" doesn't require one to be recognized as an expert—which is a trailing indicator in any case. In many fields a year of focused work plus caring a lot would be enough."
"... I think this meta-expertise really is a new thing. A new kind of paradigm of thinking that really had little analog in the past. Even the idea of a Renaissance Man, being an expert in multiple fields simultaneously, isn't quite the same thing."
I think it's existed as long as there has been a large hierarchy. I know people inside every corporation I work with that do this job, including myself.
At some level, every executive needs someone like this. The new paradigm is how much this person can be expected to possibly know. That's what's changed.
Learning has been accelerated so much that you can become fairly adept at most topics in a week or 2.
In the context of this article I think it's important to make a distinction between an expert and a specialist. I think of an expert as someone who almost always gets things right. For example, an expert pianist plays great music but it seems a bit off to think of an expert meteorologist. Weather is so variable it's just not realistic to expect anyone to get forecasts as accurate as pianist at a keyboard. But it makes good sense that a specialist will be able to better than most.
I would change "When experts are wrong, it's often because they're experts on an earlier version of the world." to "When specialists get things wrong, it's usually because the world is so complex."
So I agree that it's important to protect against obsolete beliefs it's also important to consider when the belief is really wrong and when it was just overloaded with confounding events.
There's a subtle theme in this essay which describes the thought process of how YC (well pg specifically) selects teams for YC. Perhaps in a way, bringing clarity to what is generally seen as a non-scientific and highly complex line of reasoning for hiring and selecting talent with very little limited context. However, I still don't feel like I've still fully grokked it.
> Within Y Combinator, when an idea is described as crazy, it's a compliment—in fact, on average probably a higher compliment than when an idea is described as good.
Yet, outside the walls of YC, sounding crazy means you don't get in.
> But we could tell the founders were earnest, energetic, and independent-minded. (Indeed, almost pathologically so.)
How is it possible to discern this in a YC application?
This is all reminiscent of the talk Gabe Newell, CEO and founder of Valve, gave on the hiring practices at Valve. https://www.youtube.com/watch?v=t8QEOBgLBQU#t=1291 Here he mentions the following: (paraphrased) "One of the first programmers of the company was the manager of a waffle house. He was one of the most and still one of the most creative people in the industry" So what made them take the risk? What makes Valve turn away every waffle house manager that applies today? What made YC take the risk on Chesky, Gebbia, and Blecharczyk?
Determination seems to be one decisive trait. If you're determined, people often want to help you succeed or to work with you. Determination also seems to matter more than raw intelligence, and betting on determined people tends to be a good bet.
Being willing to work for low payoff seems to be one decisive trait. If you're willing to work for low payoff, people often want to take advantage of you. Being a sucker also seems to matter more than raw intelligence, and betting on suckers tends to be a good bet.
There is an additional attribute that goes a long way in staying an expert in a changing world - humility or acknowledging that you do not or cannot know the future. PG hints at this when he says.. "...admit you have no idea what the right direction is, and try instead to be super sensitive to the winds of change...". It is also paraphrased quite well in the maxim of "Stay hungry, stay foolish" popularized by Jobs.
But humility that stems from ignorance is at odds with popular expectations of an expert. The problem is what worked in the past may not work in the future. Being sensitive to changes in the world around you, requires you to momentarily stop being an expert and look at the world from the eyes of a novice or beginner. This is not only hard but sometimes impossible to do because of how we change as we acquire new knowledge. [2]
So the trick is to be somewhat schizophrenic - being able to simultaneously view the world as an expert with the fresh thinking of a beginner or a novice.
> There is an additional attribute that goes a long way in staying an expert in a changing world - humility or acknowledging that you do not or cannot know the future.
Sadly there's one rather important setting where this does not work at all :
Interviewing with a non-technical interviewer. Like one has to do for pretty much any non-individual contributor job at all but the very largest companies.
I keep having this impulse and I find it actually convinces people you don't know what you're talking about. The more algorithms you've seen fail spectacularly at prediction, the more capable you are (especially if you figured out why), but the less capable you appear.
Furthermore people always believe they can do better by just looking at the data and going with their gut feeling. This is not true, I've only seen people do that in a few cases. And I've seen people screw it up royally in more than a few cases. I think the successes were lucky guesses, because while they looked like they knew what they were doing, I could never get a decent response to the simple question : your arguments work just as well for 5 other predictions one might make, so why this one ? So I'm forced to doubt there. Either they knew more than they were willing to admit (one was also a political figure, so ...), or they were lucky guesses.
If you are always becoming an expert at a different thing, the second you step into a new domain you "look at the world from the eyes of a novice" without becoming "somewhat schizophrenic."
I feel like we both took the same things from the article, but we come at them from very different paths.
This assumes that the goal is to be Jobs, and I completely agree with the assessment that you need to be schizophrenic. I would like to point out, the expectation is to be detail oriented(execution/expert) while leading a larger org macro goal (learning/strategy). In the right contexts with appropriate resources, this can & does magnify personalities and outputs!
> So the trick is to be somewhat schizophrenic - being able to simultaneously view the world as an expert with the fresh thinking of a beginner or a novice.
You already mentioned this, but this is extremely at odds with how experts are valued, especially by non-experts. People want answers -- not necessarily painful truths -- and ostensibly that's what they pay for. "What do you mean you're not sure? You're supposed to be an expert!"
This trick might work when you're a Jobs figure (or a Paul Graham, or a YC company) and you don't have non-experts to answer to. For everyone else, staying hungry and foolish means you'll be seen as hungry and foolish, and the market values you as a hungry fool, not an expert.
Beliefs like JEE is a good thing? Like object oriented development is a good thing? Like Portal Server is still relevant? Yeah, I hit the wall with those all the time. Lots of ego in the way of changing those beliefs because the big guys have been selling this for the last two decades. How would you like to be the guy trying to convince enterprises that loosely coupled is the way to go when they've been sold on "silos are bad we need integrated systems"? I feel a little bit like Chicken Little at times. :)
The human nature is not static as well.
Humans are able to think and act very differently.
If I look back what I was thinking and doing 20 years ago, I have to shake my head in astonishment.
But a population is much slower. To the point where "human nature doesn't change" is an extremely useful heuristic. If you look for counter examples, you'll find a few changes that took only a century or two, but in practice you'll probably never see one of those changes.
I wonder if being open minded is enough. The "winds of change" aren't always very strong, and accordingly new opportunities aren't (at least initially) better alternatives to the status quo, they just happen to have incredible future potential. To fully take advantage of this, the expert would have to move backwards, quite possibly endangering himself, while a new entrant to the field has no risk to assume by doing so.
To use the hill-climbing metaphor described by other commenters, a new climber can begin climbing the new potentially tall mountain immediately, while the "expert" will first have to descend his peak.
@pmarca has an interesting thread on this on Twitter [1]. In Peter Thiel’s terminology [2], PG’s view is optimistic nondeterminate, which might be fine for a VC but might not be the best choice for a founder.
Unfortunately Thiel's views don't really map to a "hard" 2x2 matrix--the axes should be considered to represent spectra or gradients instead. (I'm not presuming to be smarter than Thiel here, just extrapolating rationally... nobody believes the future to be entirely determinate or entirely indeterminate.)
In light of that, I suspect PG and Thiel actually live in the same quadrant of determinate optimism. I haven't read PG suggest that we should just "wait" for the future to happen--he's merely suggesting that the most common path toward building the future is on a wave. (And Thiel built PayPal that way--he recognized a version of the future that was on the cusp of being realized and took it upon himself to make it happen.)
What a dumb assertion that someone can be an expert on changing the world. Paul graham has lived in the startup world bubble for too long and is a bit off his rocker when it comes to having perspective on the larger picture of changing industries and global trends and dynamics. The next big thing is not some social media garbage app for sharing cat pictures. That much is clear. Take a look at the kind of low impact garbage YC specializes and funding to get a clear look at how paul graham views successful startup ideas and teams. It needs no reminder that the vast majority of YC companies, fail, have no impact, dissolve and get aquired by other teams or if they are lucky get a talent aquisition by a bigger fish with a bloated mergers and aquisitions budget. My advice to Paul, write essays about what you KNOW, not what you feel like rambling on about. You have lots of experience in the world of seed staging and getting small teams to make the leap into legitimate hard working business entities with growth potential. You don't have a ton of wisdom to share outside of that domain so stop acting like you do.
PG does not in fact talk about being an "expert on changing the world" but of being an "expert IN a changing world". And indeed the essay talks about... well, seed funding, investment etc. Not random "out-of-bubble" assertions.
You seem to have read the title and just spouted your vomit onto your keyboard then hit send. That's not really a good path to constructive criticism there, is it?
Your comment is completely offtopic, pointless and undeserving. You are bickering about YC while not knowing basic facts about success and impact of YC. And you are advising pg on what he should do after making a a judgement that he knows nothing outside startups so I hope you have a good set of diverse accomplishment and experience in your resume to back it up. It's his website and his essay - he can ramble all he wants in anything he wants.
121 comments
[ 3.6 ms ] story [ 185 ms ] threadApart from that, how can something (confidence) increase in a static world? PG has clearly tossed this essay off without serious thought or discussion. A waste of time.
Exactly, and it's common to see that in many fields. People are persuaded by things that were bad ideas when they were new, and aren't swayed by new evidence to the contrary.
To further improve your beliefs you might need to become less certain.
Isn't that sort of the point of the essay?
Apart from that, how can something (confidence) increase in a static world?
I don't understand. Didn't your previous point argue exactly how hill-climbing could happen, and isn't that a (bad) example of confidence changing?
Or are you really arguing over the semantics changing beliefs in a static world? I hope you aren't; otherwise, like you said: A waste of time.
The second point is that drawing a line between the "world" and your beliefs about it is a pretty dopey dualistic way of thinking. But hey, Graham doesn't care that there might be real knowledge and valid experience of these matters in the accumulated history of philosophy or psychology, he just shoots from the hip with some flaky positivistic analogy with computation... Ridiculous.
You can explore ideas that you don't hold, and these can be quite varied from your current beliefs. Once you find an idea that you put more confidence into that your current beliefs, then that idea becomes your belief.
The local optimum from high-climbing doesn't constrain you in a world where your beliefs can teleport from one hill to another.
My proudest creations all came from these weird ideas. Now I listen to them very carefully and nurture them. They're starting to cohere and turn from random ideas into an ideology. It will be fun to make that real.
One of the things I've noticed is that over a long period of time, the more arrogant and certain team members are, the less likely they are to deliver. Folks that are always learning have a tendency to kick ass. Not so much with folks who think they've already arrived. I read in an article on HN a few weeks back that Google calls this "intellectual humility" and looks for it as a hiring trait.
http://en.wikipedia.org/wiki/Carol_Dweck#.22Mindset.22
I disagree with the first statement, and agree with the second one.
Arrogance has a lot of variations. I find I know arrogant, certain people that deliver, and some that don't.
But anyone who always learns, does well.
I wish someone would write about that.
What is an expert?
What is a domain?
How do I become a domain expert?
Am I an expert at something I take for granted?
I love Paul Graham's essays but there's a lot of implied logic and jargon. Maybe he should have a few non-startup, non-business people proof read them too?
For example, I maintain an e-commerce system. More specifically, I maintain an e-commerce system built for a cosmetics company. E-commerce is a field, cosmetics is a field, cosmetics e-commerce is specifically the domain I need to become an expert in. The more I can differentiate cosmetics e-commerce from other kinds of e-commerce, the more well-tailored I can make my app. And the more well-tailored I can make this app, the better I'll be able to understand the more general case of making custom e-commerce systems for other industries.
The more I go through this process of understanding the specific needs of the industry, the more domain expertise I'm building. Domains can subdivide as much as industries subdivide. Say the cosmetic industry started to split along product lines, then I might start gaining insight into how lipstick companies operate, and be able to tailor a product offering straight to them.
The domain doesn't have to mean commercial industry. You might have domain expertise, say, in XML parsing. Then you could wax poetic on the differences between Nokogiri and say, Chrome's internal parser.
So, essentially, it just means "expertise". More specifically, the type of expertise that allows you to customize a software system around it. Which can be in practice quite different from normal expertise. The guy who devises our makeup formulas is definitely an expert.
But he can't design the system to actually manufacture it. The guy who writes those systems might not understand the actual formulas, so the two types of experts have to work together.
When startups talk about the need to be a domain expert, it means getting out of just pure programming and into the actual meat. Rather than hire an expert, to become one yourself. This not only allows you to reduce staff, it also makes it easier to make better, more custom solutions. The Soylent guys have to make the system to manufacture their product, but they also must understand nutrition, food, the regulatory environment of the food market.
I would even go as far to argue some level of domain expertise is almost required to make solutions. Even with perfect requirements (oxymoron I know), a programmer still has to make a ton of tiny domain driven decisions while writing a piece of software.
As you mentioned, software on its own isn't very useful. A programmer who learns even general business is much more useful.
So as soon as there's money to hire programmers, I'm getting out of the coding biz and hiring others to do it. I want to focus on working with my customers, understanding their needs better than I already do, and guiding programmers to build for those needs.
expert is used in the sense of the standard definition in any English dictionary
one could be an expert at programming, and though that's a valid topic and craft to have expertise in, its not a domain in the sense of "domain expert" you're asking about
A set of disciplines, or majors, other than computer-science (or alternatively something other than core programming/software-engineering skills). The word is typically used in CS/sofware-engineering circles.
An (arguably crude) example:
Think of who designed an mp3 encoder/decoder? A programmer? Does becoming a good programmer and/or getting a typical 4 year CS degree, sufficient to come up with an audio encoder/decoder?
No. You have to understand audio signals, their properties, learn how to analyze them, etc. That requires knowledge from the disciplines of physics, and electrical engineering (e.g., digital signal processing), maybe even some biology of sound. And if you learn that, you become a part-programmer, part-acoustic-engineer. That "acoustic-engineer" bit is your "domain expertise", i.e., something you need to know beyond CS to get your work done.
You might say, "why do I have be an acoustic engineer? why can't I, as a programmer, work with another guy, who is not a programmer but an expert acoustic engineer?". Yes you can. In that case, you're being the software guy, and he's being the domain-expert.
Sometimes collaboration between a software-engineer and a domain-expert works. Sometimes it helps if the software-engineer himself develops domain-expertise (so one guy playing two roles).
Domain expertise is not limited to one domain only. It could be no-domain (a pure CS programming, like database, web-development etc, although you could call them domains within CS), or a domain consisting of multiple disciplines (audio-processing comprising physics, DSP), or multiple domains (acoustic-engineering plus submarine-engineering, for example), and so on.
More:
https://www.google.com/search?q=what+is+a+domain+expert
What is an expert: Someone or something recognized as such by other recognized experts and trusted by a significant number of those who are not.
What is a domain: A limited area of human knowledge small enough that it can fit in one person's memory but wide enough that when applied it solves multiple useful problems.
How do I become a domain expert: By doing. First learning. Then making. Finally advancing. Unless of course you are creating a new domain which skips to step 3.
Am I an expert at something I take for granted: No. Most US adults know how to drive and do it semi-(un?)consciously. They are not expert drivers.
Ok its sucks if English is a second language for you but that is what dictionary's and thesauruses are for.
I see "beliefs" as an optimisation that avoids having to return to first principles for every single decision.
That speeds things up, but is obviously dangerous.
Sometimes it's possible to have two beliefs that contradict each other. Usually this means they need reevaluating, but sometimes it means there is context in which each is true.
Reasoning in the face of contradictory beliefs is one of the most interesting things about "knowledge", and something that humans can do surprisingly well - at least until we realise we are doing it.
A better algorithm would be to ask other systems beliefs and compare it to ours too (eg: neuronal network).
Do you adjust your beliefs or do you staunchly adhere to them?
i wonder what the bar here is - certainly more than blog posts / self published websites as those are fairly close to casual conversation imho.
glad to see pg picking up the pace on the writing!
I wonder if what is additionally required is : "someone you respect asks you to write on a topic". In PG's case, his audience might substitute, for others an employer or publisher might be what it takes.
I like though how he circles back to two fundamentals. One, human nature doesn't really change, and Two, betting on people is a better indicator of the future than betting on ideas.
Probably the biggest mistake people make is looking for confirming evidence instead of using evidence to determine the relative likelihood of competing hypotheses.
Even a large body of confirming evidence doesn't prove anything because that same evidence could be consistent with other hypotheses you never considered. You need to evaluate the diagnosticity of that evidence. (i.e. a patients high fever has almost no diagnostic value in determining what is wrong, because so many illnesses are consistent with that evidence.
Whereas, one shred of evidence can disprove a hypothesis completely. Which is why good doctors rule out things to identify the most likely problem.
These ideas PG is talking about are, IMO, some of the most important things for founders to understand.
Don't look to confirm your theories, look to disprove them. Don't accept lore, challenge everything. Most importantly though, always evaluate competing hypothesis to prevent you from just trying to confirm what you are looking for.
I use this framework, though learnt it by reading Fenmayn's writings, when there's a debate as to use tool A or tool B to get X. And it's been very effective so far. I have a vague sense that B is better and a colleague is particular about A. Earlier I used to look deeper and deeper into how B is good at doing X. But realized that I wasn't convincing enough. Then I remembered Feynman's saying to the effect "the best way to prove something is to try very hard to disprove it". This has helped quite a lot in coming up with objective comparisons and use right tool for the task at hand; or in general to come up with the right solution the problem at hand.
...
> "The winds of change originate in the unconscious minds of domain experts."
I would also add that domain expertise is when one's internal model / map matches the territory. When this happens, one's intuition is accurate and one can notice patterns that other people can't, or one can make predictions that turn out to be right.
Of course there is: P(A) = 0.5 and P(~A) = 0.5. Don't be silly.
Of course there is and the situation in which you find it is called Cognitive Dissonance [1]
[1] http://en.wikipedia.org/wiki/Cognitive_dissonance
> When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.
and Asimov
> Those people who think they know everything are a great annoyance to those of us who do.
beliefs, like stereotypes are shortcuts in cognition. We should always be applying the most powerful tool we have developed in the history of humanity, the scientific method.
I guess it takes somekind of rebel todo it, no matter what kind of human group you are talking about.
Well, yeah. My strength is that I'm always right. I thought I was wrong once, but that was a mistake.
(note the paradox)
We always need to show our work. The fact that an expert can crutch on the Markovian Weather doesn't mean I should trust them in any way based on their word. Trust me, says the expert, I am always right. There are those who are right and there are those who are right with proof.
The outcome doesn't make argument, the argument and the outcome does.
>"Until this point, I have been describing the stereotyping process as a negative force for individual and team functioning. However, this process actually stems from an adaptive, often functional psychological process. Mental heuristics and cognitive shortcuts enable us to process information without conscious deliberation: they fill in the informational gaps we often experience when making decisions. In other words, habits of mind help us to save brain power for more difficult tasks. Joseph Pieper's classic "Leisure, the Basis of Culture" sets out this theory well. In the 1980s, people thought we could create expert systems by interviewing experts like brain surgeons or oil exploration specialists, and creating a rule chaining prolog environment that would recreate their decision-making ability. The problem was the experts did not know how they knew what they knew. That is, experts are creating associations between disparate experiences and pieces of knowledge, using the subconscious brain." [1]
The quote above reminded me a bit of the previous "How You Know" essay by Paul Graham [2,3]. At times it's disheartening to realize the discrepancy between how much I read, and how much of it I can actually remember, but it's nice to know at least some of it manages to stay with me somehow.
In this new essay, pg says:
>"When experts are wrong, it's often because they're experts on an earlier version of the world."
Stereotyping and similar classification are essentially heuristic modeling based on past experience. Some substantial part of our mental models are based on "Unknown Knows," or better said, our mental models are based on experiences we cannot consciously recall but still somehow know unconsciously. Even if we keep the models updated to have a "current version of the world" through seeking new experiences, we are still often incapable of either anticipating or handling new surprises.
Maybe people who regularly succeed as startup investors (i.e. anticipating and handling surprising exceptions) have learned to retain some degree or form of suspended disbelief when applying their models to the world?
It's trivially easy to suspend disbelief for the sake of entertainment like enjoying a fictional story. Intentionally escaping reality for a while is an established habit for most people. On the other hand, intentionally suspending a bias, belief or model in the real world seems to take more effort, but that's possibly due to my own lack of practice.
[1] http://cacm.acm.org/magazines/2014/11/179827-the-data-on-div...
[2] http://paulgraham.com/know.html
[3] https://news.ycombinator.com/item?id=8753526
For most people this seems to be the case, but I have found a way to do it regularly. I evoke empathy before logic.
By being empathetic and transporting yourself into another person's values, you then use your logic to deduce the answers and details from their point of view. This allows you to escape your own details.
I use this often. But I can never start with logic, I have to start with empathy. From there, I can then calculate.
You can do this as a thought experiment, and you'll find it becomes trivially easy.
I work with customers often, and when I work with them I just pretend I work for them. My desires and goals change dramatically from what I would suggest for them if I didn't pretend in my head that they hired me.
I also tend to distance myself from... caring?.. about outcomes. The less I care about what the result is, the better I am at changing my perspective.
This suspension of disbelief, and the practice of it allows you to quickly shift your entire world view. The shifting isn't really the hard part, the allowing yourself to shift is.
I hope this is clear, I feel like this is difficult for me to communicate well. I think a back and forth conversation would make more sense for this topic.
Expertise today requires that you start to think and operate at a higher level of abstraction in order to cover what you know, and where your knowledge might go. Or it might require you to make connections between old ideas that are still valid and new ideas as they appear. You might even be able to change career paths entirely, but things you learned as an expert in your old one might make you special and unique in your new one -- somebody with rare and unique talents. Like automating business processes in a non-technical career.
I've been thinking a lot about beliefs and belief systems recently. I'm not quite sure I have any meaningful or concrete conclusions yet, but I'm pretty sure I'm going down the path of thinking that beliefs should generally be treated and handled with extreme caution, that they're some kind of holdover from an earlier evolutionary stage -- a vestigial kind of proto-thought process wedged somewhere between pure animalistic instinct and enlightened reason.
They aren't inherently bad just like instincts aren't inherently bad, the instinct to pull your hand off a hot stove is good for example. Perhaps they some kind of optimized way of thinking that evolved when we had less hardware to compute with. An estimation system that got us through hundreds of thousands of long cold nights, alone in the world of beasts...when we couldn't quite think through everything. But now we can, and this old wetware technical debt kind of gets in the way now and again.
It tends to make us rigid and lazy in our thinking...why think when believing is less work?
Maybe it's possible to hack our beliefs be believing in the right things, by believing in reason, or believing that things should be changing, believing that we should evaluate and challenge ourselves from time to time, believing that we should feel like we're just on the edge of discomfort and force us to use this new better and harder to use equipment we've been gifted with.
I'm glad you wrote this, because so many people bemoan the first quoted bit and don't take into account the second. Sure, FooFramework is out of fashion, but the underlying knowledge that a good programmer, one I'd consider an "expert", persists. The knowledge that makes a person a good programmer, the real skills involved rather than the minutiae and the whatevery bits involved in translating that knowledge and those skills into a particular use case, have personally served me well (so far) in web development, mobile development, devops, and game development.
That may be true in programming, but that's not the only field. Much of my mechanical engineering knowledge dates from the 19th century and earlier and is still valuable. Even my FEA knowledge (which relates heavily to programming) is still good. Practically all my undergraduate knowledge is still valuable, let alone my Master's and PhD knowledge.
It's quite clear that you're approaching this strictly from a single viewpoint. If you really want to think, you need to broaden your view.
Even in programming there's still things that change at a slower pace. Sure there's a lot of things changing in regards to the web, but areas such as numerical methods change more slowly.
So while I'm pretty sure the basic physics you use is largely unchanged, I'd bet that a mechanical engineer for even a generations ago would have trouble recognizing much of what you're doing today. We have materials, measures, testing and production methods available to us today that would have looked like magic to somebody in the 60s. It's like comparing a flint-lock musket to a modern rifle. A gunsmith from the flint-lock era would know the basics, but his expertise would be obsolete.
Even in software, most of what I learned in my undergrad CS program has stayed refreshingly relevant, but the act of writing software itself has completely transformed a half dozen times since I graduated. It means I can sit down in a new environment and learn the syntax, practices and idioms of a language and dev environment relatively quickly. But in many cases you quickly start to feel like you're treading water at the same time, having not 20 years of experience but risk having 1 year of experience 20 times unless you carefully cultivate your expertise to prevent that.
My civil engineering friends seem to have similar feelings, lots of the physics and basics they learned are still relevant, but the theory and implementation, especially the huge changes in Urban planning are having tremendous impact on their ability to cultivate expertise.
I have a landscape architect friend who graduated hoping to design golf courses and parks and who now spends most of his time working on energy efficient certifications (LEED) and planning for new construction and retrofits. Something that didn't even exist when he entered his line of work. Instead of becoming an expert in park design, he's become an expert in LEED certification.
Even my pastry chef friend left pastry school, where she intended to make traditional French pastries, and entered a world expecting innovative techniques and ingredients catering to the allergen/wheat sensitive/low carb/vegan/no GE/whatever crowd of the month. She doesn't just make scones, she has to make 6 kinds of each scone suitable for everybody's special dietary decision.
You're right in that they're all still doing mechanical physics, looking at water management, and baking things in an oven. Expertise in those subjects transcends. But being an expert in their fields is impossible because the expression of what their entire fields are are being redefined almost every day. What they're becoming experts in, and something their predecessors didn't have to do, is riding the wave of change.
This is a fundamentally different kind of expertise that maybe only ever existed in human history during times of war, or great societal upheaval. "May you live in interesting times" is a curse, not a blessing. It's to be seen if this is simply the new state of things from now on or a time of upheaval that we can't recognize and things will settle down at the other end.
If you approach things very shallowly, learning how-to-set-up-technology-stack-X type bags of tricks, those become obsoleted quickly, but lots of technical expertise is quite durable.
In my experience, a world that rapidly changes makes it very easy to become an expert. There are so few in whatever is new, and all it takes is 1 year of effort to become an expert in that field.
Having the expertise remain valuable is more about luck and intuition, but I don't really care one way or the other if what I spend my time on remains valuable. The valuable part is being able to learn something, and developing the skill to learn tasks. If you repeat this with a different thing for every year of your life, you will become an expert at becoming an expert, a skill that far exceeds knowing any one domain.
I don't, and have never desired what you seem to desire, learning an expertise that lasts generations. Which means, learning a domain that lasts for generations.
The very act of being an expert at becoming an expert makes you impossibly valuable to any corporation that exists. The actual details of the domains is completely irrelevant.
I think we may have a deep disagreement of what constitutes expertise. 1 year of effort in anything is at least an order of magnitude less effort than what I'm talking about. I would consider a smart person with a year of solid effort "functional" but not even capable of sniffing what it takes to become an expert. That level of grokkiness is simply not attainable in that short of a period of time.
But maybe you bring up a good point. Maybe the word has simply been redefined these days to a much weaker definition. Maybe a year of practice, in this sound-byte, short attention-span laden world, is good enough and our collective standards have fallen. After all I see job listings for "senior this or that" job all the time requiring 5-7 years experience in the field.
> The very act of being an expert at becoming an expert makes you impossibly valuable to any corporation that exists. The actual details of the domains is completely irrelevant.
I agree, and I think this meta-expertise really is a new thing. A new kind of paradigm of thinking that really had little analog in the past. Even the idea of a Renaissance Man, being an expert in multiple fields simultaneously, isn't quite the same thing.
"In practice "sufficiently expert" doesn't require one to be recognized as an expert—which is a trailing indicator in any case. In many fields a year of focused work plus caring a lot would be enough."
I think it's existed as long as there has been a large hierarchy. I know people inside every corporation I work with that do this job, including myself.
At some level, every executive needs someone like this. The new paradigm is how much this person can be expected to possibly know. That's what's changed.
Learning has been accelerated so much that you can become fairly adept at most topics in a week or 2.
So I agree that it's important to protect against obsolete beliefs it's also important to consider when the belief is really wrong and when it was just overloaded with confounding events.
> protecting yourself against obsolete beliefs is exactly what you have to do to succeed as a startup investor
This statement is not only true for startups, but for life in general.
> Within Y Combinator, when an idea is described as crazy, it's a compliment—in fact, on average probably a higher compliment than when an idea is described as good.
Yet, outside the walls of YC, sounding crazy means you don't get in.
> But we could tell the founders were earnest, energetic, and independent-minded. (Indeed, almost pathologically so.)
How is it possible to discern this in a YC application?
This is all reminiscent of the talk Gabe Newell, CEO and founder of Valve, gave on the hiring practices at Valve. https://www.youtube.com/watch?v=t8QEOBgLBQU#t=1291 Here he mentions the following: (paraphrased) "One of the first programmers of the company was the manager of a waffle house. He was one of the most and still one of the most creative people in the industry" So what made them take the risk? What makes Valve turn away every waffle house manager that applies today? What made YC take the risk on Chesky, Gebbia, and Blecharczyk?
They interview applicants in person, too.
But humility that stems from ignorance is at odds with popular expectations of an expert. The problem is what worked in the past may not work in the future. Being sensitive to changes in the world around you, requires you to momentarily stop being an expert and look at the world from the eyes of a novice or beginner. This is not only hard but sometimes impossible to do because of how we change as we acquire new knowledge. [2]
So the trick is to be somewhat schizophrenic - being able to simultaneously view the world as an expert with the fresh thinking of a beginner or a novice.
[2] https://news.ycombinator.com/item?id=8754334
Sadly there's one rather important setting where this does not work at all :
Interviewing with a non-technical interviewer. Like one has to do for pretty much any non-individual contributor job at all but the very largest companies.
I keep having this impulse and I find it actually convinces people you don't know what you're talking about. The more algorithms you've seen fail spectacularly at prediction, the more capable you are (especially if you figured out why), but the less capable you appear.
Furthermore people always believe they can do better by just looking at the data and going with their gut feeling. This is not true, I've only seen people do that in a few cases. And I've seen people screw it up royally in more than a few cases. I think the successes were lucky guesses, because while they looked like they knew what they were doing, I could never get a decent response to the simple question : your arguments work just as well for 5 other predictions one might make, so why this one ? So I'm forced to doubt there. Either they knew more than they were willing to admit (one was also a political figure, so ...), or they were lucky guesses.
I feel like we both took the same things from the article, but we come at them from very different paths.
You already mentioned this, but this is extremely at odds with how experts are valued, especially by non-experts. People want answers -- not necessarily painful truths -- and ostensibly that's what they pay for. "What do you mean you're not sure? You're supposed to be an expert!"
This trick might work when you're a Jobs figure (or a Paul Graham, or a YC company) and you don't have non-experts to answer to. For everyone else, staying hungry and foolish means you'll be seen as hungry and foolish, and the market values you as a hungry fool, not an expert.
But a population is much slower. To the point where "human nature doesn't change" is an extremely useful heuristic. If you look for counter examples, you'll find a few changes that took only a century or two, but in practice you'll probably never see one of those changes.
[1] https://twitter.com/pmarca/status/546533509922160640
[2] http://blakemasters.com/post/23435743973/peter-thiels-cs183-...
In light of that, I suspect PG and Thiel actually live in the same quadrant of determinate optimism. I haven't read PG suggest that we should just "wait" for the future to happen--he's merely suggesting that the most common path toward building the future is on a wave. (And Thiel built PayPal that way--he recognized a version of the future that was on the cusp of being realized and took it upon himself to make it happen.)
You seem to have read the title and just spouted your vomit onto your keyboard then hit send. That's not really a good path to constructive criticism there, is it?