Hacker News Reader

Yes, the idea and the debate is as old as history. I prefer the quote from Heinlein, "Specialization is for insects." Because actually, if you look at well run organizations, everybody needs to be a generalist to some degree. You don't just work in a silo and ignore every other aspect of what your organization does, otherwise you produce deficient solutions that your colleagues in their own siloes need to work around. Eventually the only people that can thrive in such an organization are the generalists because everything fits together so poorly only the people willing to tackle novel things can solve any problem. People talk about force multipliers because you need to be generally aware of what a lot of different people are doing and help them align to be as productive as possible.

> A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects.

He was a bit pro-military for my tastes, but I hope you get the idea. You'll make more money stepping up than drilling down. Of course, context and opportunity matters a lot. Success isn't dictated (possibly barely affected, given the debate) by this single dimension.

It's an evocative quote, but it's also nonsense. The selection of skills is a bit scattershot, but let's look at some examples: "butchering a hog" takes a year to learn, well, butchery to a professional standard, because someone who can butcher hogs but not cows sounds awfully 'specialised', some places even say that a butchery apprenticeship takes 24 months. (And I personally wouldn't trust an unqualified butcher to process meat I'd be eating.) Building a wall is easier, you can get a bricklaying certification in 8 weeks. Learning to "Fight efficiently" and "take orders" takes US army recruits 22 weeks (and we can throw another 12 weeks of officer training for giving orders and generously allowing this to count as 'planning an invasion'). Computer programming, sonnet-writing, account balancing, cooking and solving equations are actually part of general education, so let's ignore those. Basic first aid only takes a few days, and pitching manure can be learnt in under an hour. But then we get to "design a building" and "conn a ship" (not a boat, a ship). Conning a ship is a bachelor's level qualification in of itself, and designing entire buildings is a job for a chartered engineer, which requires years of experience and another bachelor's degree, so we're taking probably ten years of full-time training for those two alone. And I would not want to be on a ship piloted by someone who'd been on a 3-day course or live in a tower block designed by someone who'd read a couple of Wikipedia articles. Then we'll throw in a 4-year degree in whatever they're actually doing, and you're looking at a society whose adults don't start entering the workforce until their mid-thirties, and a newbie doctor or research scientist or some other field which requires postgraduate qualifications is probably in their early forties. This isn't even remotely viable as a way to run a society. And it's also implied that that list isn't even exhaustive, so the real numbers are even worse if there's some other complex safety-critical thing that Heinlein also felt should be on the list, like 'fly an airliner' or 'manage a chemical plant'.

Maybe with sci-fi anti-aging techniques that mean people can still be expected to be healthy and productive in their nineties, but in the real world, Heinlein's generalist utopia is pure fantasy.

I think they're talking about butchering a hog at the skill level of a person who grew up near people who owned hogs, ie, definitely a "I did this with somebody at a barbecue a few times" and not "24 months of training".

Giving and taking orders likewise doesn't remotely require army training, it just requires growing up around human beings and being part of anything -- theater, bands, church groups, service groups, clubs.

Fight efficiently means, I believe, fistfight, aka, take some boxing lessons and you can check that one off.

I think you've really misunderstood the quote. Or maybe I have.

Well, Lazarus Long was supposedly over 2000 years old, so he had the time to learn.

https://en.wikipedia.org/wiki/Lazarus_Long

I notice you ignored all the actually difficult items on the list there. You know, the ones where the Dunning-Kruger brigade are a liability and mistakes mean people die.

I didn't ignore all the difficult ones. But fine. Bricklaying is a few days to learn on the job, tops. Conning a ship almost certainly means captaining a boat which takes some experience for sure, probably a few months. Basic first aid, including setting a bone, is a week long course. Etc.

I can't imagine why you went to the trouble of trying to make them all sound like complicated many-year projects, but it's a ridiculous take and I doubt anyone else will read it that way.

Are you seriously suggesting that Robert A. Heinlein, former naval officer, wouldn't care about the distinction between a ship and a boat?

And you still ignored the two most difficult ones.

I think he knows the difference, but surely meant the life skill of captaining something rather than the literal many-years-to-acquire technical skill of doing so, because the former makes sense and the latter doesn't.

Which two? Designing a building I see I skipped. That seems pretty attainable, he obviously didn't mean 'be a trained architect and building engineer' like you interpreted it. People in charge of organizations sometimes have to design buildings (probably much moreso in the past though). So like, 'kitchen goes here' etc level stuff.

I'm not saying I agree with the whole quote. I just don't see the point of misinterpreting it so aggressively.

He specifically said ‘conn a ship’. Conning is the specific highly technical act of directing the ship while in motion. Navy types are known to be sticklers for precise use of naval terms, so we have no reason to assume it doesn’t mean exactly what it says.

The task you are describing is not design; it’s specification or briefing at best. Heinlein was also an aeronautical engineer. Engineering is are also a profession where imprecise language costs lives.

The hardest one was “plan an invasion”, but I’m sure you’re about to tell me that a former naval officer and author of some of the most famous pieces of military science fiction didn’t really mean “plan an invasion” when he said it, or that despite modern history being full of poorly-planned invasions (and there’s one going on right now), actually it’s something you can learn on a three day course.

Oh, I totally missed that one.

Yeah, sounds like "conn a ship" and "plan an invasion" are actually him just being weirdly militaristic. I guess he thinks everyone should have some career in the military for a while

No matter. My complain was that for some reason you were making the other ones sound like they take a long time to learn, which is obviously absurd.

I think his intention was a spaceship/starship. He was a science fiction writer, after all, and most of the Lazarus Long stories take place elsewhere in the Galaxy. A starship can be more or less complicated than a maritime vessel, depending on the onboard computer. Maybe as simple as saying “take me to a nice planet to relax”.

If you want anyone to address the meat of your question simply state what you believe to be the two hardest tasks instead of being vague and making people guess. The conversation would be better for it.

My take is that Heinlein was trying to write literature, not documentation. So a colorful exaggeration makes sense.

People, including the poster I was replying to aren’t generally writing literature when they quote it though, and they like quoting that a lot.

How about you actually try to learn butchering a hog and stop complaining how hard it is? It took me about two hogs and a good teacher to get it sufficiently right, and I was a ten-year old kid back then. Claiming that simple things like that take years to learn is writing rich literature. Earning a living doing skill X takes a lot longer, but learning those skills is not a big trouble. If you've had an active rural childhood with good parents most of those skills are "completed" very early.

Writing a good poem and planning an invasion takes some maturity, but those should be interpreted similarly. Learn about basic logistics and doing projects and then you can plan an invasion. Making things more complicated than they are is just another way to do nothing at all. Learning takes some time and effort, if it's too much for you I understand, but don't discourage others with your weak excuses while thinking that you must become a specialist in every field to become a generalist.

You missed the forest for the trees:

> He was a bit pro-military for my tastes, but I hope you get the idea.

I was trying to emphasize the point being made to you: the particular skills listed aren't relevant to the point of the need for general skills (the idea I hoped you would get) but instead relevant to all the people around the protagonist in the novel (which has a military slant due to the author). I'm quoting literature to show evidence that the debate is old, that was the entire point of my post; what am I supposed to quote except something that someone wrote? People like quoting literature for reasons like that: evidence that an idea is general, old and timeless, that our ancestors were grappling with the same ideas, that it isn't settled and probably never will be.

The point of the quote is not that those specifically called-out skills are more valuable to know than other skills. The author could have picked 21 entirely different skills to list, and the point is exactly the same. He's also not saying people should have PhD-level knowledge or master-level training in every skill they know--quite the opposite.

That's absurd. You can learn how to butcher a hog in an afternoon (I'm not including curing ham or bacon in this statement).

> If you care about things like income and career progression, do not let yourself become one of these people. They aren't valued, and it doesn't take a lot of staring at job postings or talking to hiring managers to figure out why.

I have the opposite experience. Being a generalist allowed me to "climb" fast at my current company as I am able to lead projects holistically and coordinate cross-functional teams.

I have no shortage of attractive offers, but more importantly I like to work this way. I can't imagine limiting myself to just one narrow area and ignore the big picture and so far it looks like I do not need to.

As for job descriptions: I would love to hire more of these people but it is difficult to put a job description through the system that would not be confusing. Thus, even I have written job descriptions for one specialisation that should be dominant, but in the fine print I make it clear that I am searching for someone with broader scope and will to basically "do what is needed".

PS: Thinking about it more... do you think industry legends like Jeff Dean or John Carmack fit more into the generalist archetype or are they narrow specialists?

I think at some level specialists need to become generalists, or they will get stuck.

For a junior backend engineer, it might be okay to not know much about UX design, front end development, database administration, etc.

But at a certain level, you need to know about all the fields outside of your specialization as well. You need to understand why you are doing what you are doing, or you'll end up doing a bad job at it.

Re John Carmack. Seeing his journal, strongly a generalist that adapt.

I mean look at how they ran Armadillo. It was just a bunch of random problem solving constantly.

> industry legends like Jeff Dean or John Carmack

Those are very bad examples that prove the gp's point. Sure, flexible and multitalented people are a huge asset IF they get their big break, if they get involved in a startup at the right time in the right industry, if they are lucky to not get crushed by abusive players etc. etc.

But the chances of success are very slim, you likely won't be John Carmack. For every Carmack there are thousands of failed game developers who never got their break. Starting a tech company in 2022 is very different from 1992 even if you get to Carmack's skill level. Meanwhile, holding a job is something that always works and being a specialist in a high demand area is strongly correlated with financial returns.

I write this as somewhat of a "renaissance man" interested in everything from engineering to coding, to economics, to political science, to philosophy - and even holding advanced degrees in these topics. Sure, I can make my paycheck in tech, but without a managerial track record and access to the upper echelons of capital and corporate power I won't be recognized as a critical asset for an organization nor will make "Kubernetes expert" money.

Luck is always an essential ingredient in success. Can you point to any counterexamples: i.e. an industry legend who is a specialist?

The point was that you don't need any luck to become an extremely well paid and well regarded specialist, just dedication and time.

If your objective is to become a legend, then by all means avoid specialization, and be prepared to fail with a probability of 99.9%, there can only be so many legends.

Hm. I don’t know.

As a “generalist” myself, I find it sometimes hard to find a new job (as it’s hard to describe what exactly you did), but once I land a new job, I somehow become a person that just can’t be fired, and bring a lot of value to the team.

However, what I CAN recommend for such people and what did help me is to move more into management. If you are a jack of all trades, you can also quite easily tell what will take what time, and how much are people bullshitting you. And learn some JIRA/scrum shibboleths on the way, and you become quite a good manager.

I can't second this enough. After 20 years as an excellent generalist I was getting bored as an IC. I struggled with finding interesting roles until I moved into management. All those good enough skills in a lot of areas, and the ability to connect things and people, really helps you shine. Plus, you get to fix all kinds of new problems.

> 17 years ago, there was a popular essay about this sort of person, and the label at the time was "Free Electron": https://randsinrepose.com/archives/free-electron/

Also: https://www.joelonsoftware.com/2007/06/05/smart-and-gets-thi... http://steve-yegge.blogspot.com/2008/06/done-and-gets-things... and pop psych tests like MBTI

This kind of essay is good for the author, because writing an essay about "what if there were special people who were awesome and really cool?" attracts readers who think it's about them. Doubt it has much other merit though.

Personally, I try to develop the opposite talent where you don't produce anything. Otherwise they ask you to maintain it!

It works just fine for me. Even the best operations wants to stay competitive, and I have yet to meet a CEO with less ambitions than the org can handle. That is where we Wildcards thrive.

I build the departments we don’t have, make the prototypes the developers says are impossible (or unattractive technically), prove the business models that yikes the existing sales organizations as they are misaligned with common practice and so on.

That said, you must be sellable as a Generalist/Consultant/Fixer. Most people who try to go the Generalist route fails as well. And what the article fails to mention, is the hidden set of skills you need in addition to be valued in this odd position.

I've found my self in this position as well.

So far I've managed to identify a few key points which has served me well so far.

- Being able to produce results even when there is a lot of unknowns.

- Fairly good communication skills across several departments. People usually listens to what you say.

- Being known as goto kind of person. If it's about getting the gist about something or direct help of implementing any kind of service/api.

And lastly the main point of being Sellable: Being a generalist has put you into a lot of scenarios were others have no previous knowledge about. And thus making you easier to act as an early asset in assignments.

Interesting discussion. My solution has been to sell myself as a troubleshooter. So I usually see patterns and solutions that others do not and in such a case can bring something to table where people get stuck.

A troubleshooter sounds very specialist while it skills are in actuality quite generalist (on the whole)

I've sold my self as a troubleshooter before, but that only got me into positions were I was maintaining old tech. Especially things others had no idea how/why it did work.

I still find my self in these situations, but not as much like before since I stopped using troubleshooter. But your miles may vary :)

Yeah this comment thread is certainly enlightening in that regard. Maybe I should rethink my strategy! May I ask what you market yourself now?

I've kind of switched over to a more specialized direction, but I still do include all/some the general stuff as well.

Meaning: It's imperative you explain there is a focus on what you're specialized/interested in. And if they're still interested there is a lot of extra goodies you know as well. But only as long as it's part of the assignment/position you're interviewing for.

I know this answer is a bit vague, but I hope it was still to some good use for you.

Yes, thank you! This end of the year period is good to reflect on the coming year and make some changes (if necessary).

You are right in many senses. Modern organisations do value specialisation, and tend to perceive the organisation as a legible machine with specialist components. This tends to be particularly prevalent among HR, professional management, outside consultants and such.

That sense, I think it's a false perception. Organisations don't really run on specialisation. Generalists are what make the organisation run, with the mist important specialisation being specialisation in the company itself.

Realistically, very few of the supposed specialists really are specialists. There are, of course, a few roles where specialists reside. Roles that require deep or intricate knowledge, but these tend to be few.

IMO, many of the failings in modern organisations are related to this tension between theory or ideal and reality.

People are handed specialised roles, and then expect to be handed tasks that map to this speciality. Very often, the real problems don't fit this mold. This is where we get team where no one can do anythibg unless everyone is doing. everything. It's where we get busywork, because middle managers job becomes finding tasks that fit neatly into roles.

Is this simply an apriori argument? Do you have any evidence to back up your claim beyond your reasoning?

I don't find it very persuasive, and based on my own experiences being a generalist for the majority of my career, I can say it certainly doesn't apply to me. My compensation is >$500k a year and I've been a software engineer since 2010.

What search terms would you recommend a generalist plug in to indeed.com to maximize their chances of getting hired at a competitive salary?

Remember, they are a generalist, so specific languages, frameworks, and platforms don't count; they won't be more experienced in any of those than competing applicants, by their nature.

We generalists find work through our network of contacts. People we've worked with previously reach out to us when their current company needs a person like us. In 20+ years as a generalist, I've never been without work.

Exactly: if you're talking to people who already understand your value, then you don't have to pitch your value to them. Specialists also benefit from this, with the additional benefit that they can look for employment outside their network.

There are some survivorship biases responding in this thread, but the fact remains that employment opportunities favor specialists over generalists, all other things being equal.

How are you with the Layer 8 issues? I find in large organizations, that wildcards who can effectively relate and communicate, either directly or through an intermediary, with DRIs whose organizations they depend on, will succeed. The reason is that when a generalist/wildcard/"tiger team" is introduced to a mature organization that has specialized, the reason usually is that there is something not working as desired with the system as a whole, a system that significant investment has been placed into. and that takes as much human as technical finesse to unravel. If you're good at that, I can imagine you've done well. But effective communication is not a common trait among wildcards I've known. Most have been protected by excellent managers.

> Well-run organizations -- that is, organizations that can afford to pay good money for talent -- have clearly defined roles for people, and need only a few generalists to handle the nitty-gritties of cross-department communications and management.

> The organizations that most need generalists are the ones that are constantly struggling to get their crap working right, so they need people who can do a little bit of everything, and as a consequence, they also can't afford to pay such people a lot of money.

> It is much easier for a specialist to argue their value than a generalist.

Yes, but... when these organizations need to staff a small weird project or straighten out a mess, they often hire a consultant. Consultants get paid a lot of money and their work changes frequently. This is a great place to be if you're a "wildcard". You can work for yourself (high risk, high ceilings on pay, lots of sales time) or a consultancy (low risk to you, modest employer risk, lower ceilings on your pay, little sales time).

If you have to market yourself as a "wildcard", remember that just like you never want to be a "programmer who can do X", you want to be an "X who can program", so it goes with being a generalist. No one really wants to hire someone who doesn't have some specific talent. So you're a "generalist X", not just a "generalist". (Probably, hopefully, you do have some area or other of more-detailed knowledge!)

Sure, but who gets the business to that well oiled machine kind of state? When you’re starting off, I think having someone fitting the description of this article can be pretty valuable even if the end result doesn’t really require it

For example Google hires mostly generalists and lots of people who care about income and career progression work there.

Google does not. It hires a lot of specialists and fungible people.

Specialist and fungible are contradictory. It's true that they don't hire specialists - but there are only 7-8 different software profiles they look for.

1. ML 2. Systems 3. Product 4. Data engineer 5. SRE ....

Are you implying most companies have 37 profiles? Even this article was giving multiple examples or a little Web shop that probably has less profiles than 7.

Obviously no.

What I meant is that they don't specifically try to hire nodejs/computer vision/graphics/kubernetes specialists etc.

Lots of startups and mid level companies are looking to hire specialists in computer vision, nlp, scala expertise etc.

So startups can have 4-5 highly specialized profiles.

From my personal observations, Google indeed hires a lot of generalists.

Source: been an SWE in a hardware org for the past year there, and my team is a mix of about 1-2 people with plenty of previous hardware experience and focus, with the rest being generalists without much previous related experience (including me) who can pick things up and resolve them quickly, whatever they are. Observed quite a similar pattern on other teams in the org as well, with the only exception (to a degree) being a few research teams filled with PhDs.

Or they run their own business.

I disagree, with caveats. There's the idea of a "T-shaped person": A little bit of knowledge in a lot of things, with deep knowledge in one area. This is what organisations tend to hire for.

The wildcard person eventually develops into a shape more akin to the differencing mark seen in heraldry for first-born sons. Much broader and deeper general knowledge, and several areas of very deep knowledge.

As one of these people, I can stand by the fact that, yes, you will be sought out by organisations which are struggling with serious challenges, but that does not mean that they can't pay good money. They can frequently pay very good rates, if you're working as a consultant, because you'll be paid for out of CapEx.

You are getting into a metaphor I saw here on HN some years back: "The Paint Drip" model of skilling.

The more you paint over a spot, the more it drips as paint accumulates. The things you spend the most time on, have the longest drips. Some may be in a T-shape, but as you say, there may be many longish drips in a person's skillset.

For the first 10 years of my career, I was more T-shaped. But also pretty good at jumping in and figuring things out.

I'm now a lot more "generalist" - I just tell people, "Try me; who knows what I'll know or be able to figure out". I still have some of those really long drips in my paint drip chart, a few really really long, but the nature of tech means that at some point, some tech you learned really well is dead and you'll never need to use it again. Such is life!

>"Try me; who knows what I'll know or be able to figure out". This mindset is usually highly toxic in workplace. At minimum it's a nice way to get yourself branded the official a-hole smartypants of the company.

Nah. I am definitely not that guy.

Toxic, perhaps for me. In that if I see something not getting done and I cannot get anyone to do something about it, I will try and handle it. I learn a lot along the way, but at times have too heavy a workload as a result.

A big part of this is that specialist can be fake, often are, and nobody can tell. They just need to know slightly more about kubernetes or whatever specialty it is, than other people around them. I often come across “experts” that aren’t really experts in anything besides self-branding. Our university system is built on this. It’s also the reason why we haven’t made much progress in the last 50 years or so. Everybody is just enough of a specialist to demand a wage, not more. What’s the saying, “they promised flying cars, but gave us Facebook.”

> You'll just never make the same kind of money as the person that knows Kubernetes backwards and forwards and not much else.

is there a statistics on this?

I laughed when I read this because being a kubernetes expert implicitly means having lots of experience on a wide array of subjects. Cloud services, networking, containers, etc. DevOps is literally one of the most generalist jobs out there.

A better example of a (typically) highly specialized job title might X Developer (where X is any programming language).

I'd be interested in seeing if other wildcard types actually care about this? Beyond living well of course, need cash for that.

My drive is dopamine, not money. As long as I'm fed and not in debt I couldn't give a rat's tail how much I'm earning. I'd take a hefty pay cut to jump on a "hell yes" project.

> I'd take a hefty pay cut to jump on a "hell yes" project.

Very easy to say. Let's get back to it when you have actually taken a hefty pay cut.

In November I quit my full time job to go freelance while building my own business because I got bored of the Groundhog Day I was living in (7 year itch?).

So far it's the biggest pay cut I've taken, although admittedly it is looking like it'll stabilise to more than I was on FT over the next couple months. Does that count? Am I allowed to get back to it? Haha

Figured now's the best time to try. No dependents and worst case I crash and burn and move back in with mum for a bit :)

I don't think taking a risk with a expected pay off qualifies as "taking a big pay cut so I can work on interesting stuff".

Fair enough.

I think a lot of technical people care more than a rat’s tail what they earn. But I agree, it seems like a of HNers say they aren’t interested in any more money than to just barely physically survive. I mean, they say that, I don’t know. I haven’t many of these types in person. It is amusing, though.

I've seen those types in person, and it is indeed amusing, but they exist. Typically, they fall in one of the two categories.

1. Recent college grad or someone who hasn't worked at a company where there is significant pay involved yet. They tend to abandon their take on this, as soon as they get their first well-paying job. At least until they potentially become a part of group [2] I describe below, but that's not necessarily always the case.

2. Those who have been in the industry for a long time and amassed a savings nest/net worth large enough, they can afford not to work for quite a long time without taking any noticeable hit to their finances at all.

For HN specifically, my guess is that we have quite a solid mix of both, with the latter probably being better represented here due to the average age and experience.

Disclaimer: all of this is pure speculation and just my personal take.

Get this — if you’re a generalist, you can actually decide to become a specialist in one area for the sake of money and career advancement, the same way you’d pick up any other piece of knowledge to solve a novel problem.

I would love to just be able to "not let myself become one of these people". Unfortunately I can't change how my brain reacts to stimuli (and I've tried).

The generalist’s solution to income is to have multiple jobs. If done correctly, the sum of those, even if individually lower paying, jobs should exceed a specialist role.