Why decline in generalists leads to disjointed games and harms tool quality
gameworldobserver.com
gameworldobserver.com
The issue has never been a lack but that the industry itself crushes you into a box and insists you don’t exist. I’ve sat in meetings where generalists are described as unicorns. There is usually zero organisational impetus to utilise people’s abilities outside of their job spec. There aren’t a lack of generalists, organisations are just blind to their existence and have no understanding of how to use them.
As a consequence I’ve always looked for opportunities that let me be a generalist, working on prototyping teams, startups that need you to muck in and now leading a team of generalists making games. Quite a few of my team came from other industry jobs doing something specialised because of what we offered.
I think Tim’s diagnosis of the issue is correct for larger teams, generalists provide glue, shortcuts and important cross-cutting expertise. It’s just the actual problem is larger teams are also terrible at finding and integrating generalists because it’s anti-thetical to their working model.
I find it fascinating that the gaming industry is faced with similar issues with generalists, where I was expecting a lot more pragmatic choices and better paths for individual contributors even if they're not ultra specialized.
Sometimes the intersection point issues are due to a breakdown in communication between the two experts but sometimes its due to a lack of knowledge overlap - i.e. the lack of a generalist on either side of the divide.
It's a similar solution to other tech: generalists aren't vauled as much at huge studios, but they thrive in smaller studios. So the "exit plan" for a generalist involves hitting the ground floor early on an up and coming studio and getting a significant stake in a successful release.
And like those unicorns, those new studios are a huge risk. But hey, at least these days you have a huge rabid fanbase of hardcore gamers who will foam over each PR release as long as you do some form of advertising.
A generalist is a threat to the organization as a concept.
Organizations rely on power and information flowing in well defined ways. Information flows up, control flows down. A generalist might disrupt that by passing information and instructions sideways. Which is why, no matter how often leaders go on about cross team working and matrix organizations, things tend to revert to micromanagement every time.
I think it's simply a choice made by these big title publishers. If you want more generalists, give them more freedom, more leading roles.
Mind you, I don't work in the industry, so I could easily be wrong, but this is my impression as an outsider.
He could get underlings to execute on his vision, but he felt that if they didn't believe in it, they likely would put a lot less effort than he would.
Some industries have realized they want some people who have multiple skills, such as the "full stack engineer" roles, but it seems rather more common to prefer someone who precisely fits into the specialist role they defined.
I could have come into a job and say "okay, I know back end, let me focus on that" but instead I spent a day installing a shitty data pipeline that more or less took us through the next 4 years of data analytical needs for almost no cost. Etc.
It is just how it is. I even read a post on HN about a decade ago about how a guy was hire #1, was instrumental in the growth of a company, and when it was purchased was pushed out because they didn't need a generalist anymore, they needed a specialist. And so he decided to drop the generalist role and focus entirely on one technical aspect to the benefit of his career.
So while I agree with Tim, I also have to say, as someone in the thick of it, there's no good way out. Unless you want to keep working for Seed / Series A startups, fingers crossed they sell or you won't make that okay payoff.
Just give it some time and start by doing it on a part-time basis. It might take several months before you land your first project, but once you completed some work for some clients, more business will come if you are good.
You will know when you are ready to take your individual contracts full-time.
You can just drop it in casual conversation. "What do you do for a living?" "Oh I'm a software developer at X-Corp, and I also do some work for individual clients" Something like that. People remember that stuff. If you like, you can even tell people you offer a finder's fee. Just don't be pushy. Be willing to build the business part-time until you have enough work to take it full-time.
You probably won't be getting big projects from the get go, but if you are humble, that won't be a problem. My first project was just optimizing and cleaning a slow and messy Python script. I charged $35 an hour an threw 10 hours into it.
People flat-out don't believe that the skills enumerated on my resume could be held by a single person, and are.
Apparently, I'm clearly "over the hill", having been at this for nearly 40 years. I should state that I have 20+ years experience instead, as-if having life experience doesn't actually make you better qualified to understand the contexts of any role in an organisation.
My skill-set is "threatening" to others, because, apparently, knowing things about a great many technologies means that I know more than the person hiring me, which somehow means that I'm a threat. As-if surrounding yourself by people who are smarter than you is suddenly undesirable.
I've been told to "dumb down" my resume because there's too much there, even if there's 99% chance that a computer will be the device actually reading it and selecting me. Apparently I'm supposed to "infiltrate" a company at a lower level and then "show" them that I can do more, despite 40 years experience showing that people only see the skills that they want to see.
According to some automated job technologies I have 104 years of experience. Not sure how that works.
Then there are those who want to know which of the experience is "professional" and which isn't. (I've been self-employed for the past 23+ years)
In the end, I liken it to this:
"How much experience do you have in eating food?"
"Do you mean, adding together each of the 15 minute meals, or how long have I been eating food for since I was born?"
"Okay, how much professional experience do you have eating food?"
"..."
The "system" still serves the ones who grow vertically Junior X, X, Senior X, Junior X Manager, X Manager, Senior X Manager, and so on. I remember listening to DOAC chatting with Harley Finkelstein of Shopify that was describing how he appreciates T-shaped skill-set. So you master skill X (as you 'grew up with it') but as time was going by you also developed other skills in parallel.
Modern companies & minds get it and celebrate it. Old-school-ers don't (and probably never will). There is also the thing "I want him/her to do X, so (e.g.) project management skills are not useful". And then I am happy when I hear this on an interview because I most definitely do NOT want to work with that person! :)
I've listened to Chris and am a huge fan of his. The depth, experience, and love he puts into POE clearly demonstrates a generalist who gets it. Chris is a linchpin.
This is exactly the kind of people Blizzard lacks. They used to exist but they pushed out many of the generalists for the factory model of specialists and haven't really recovered since.
That doesn't mean they don't make decent games given the talent of the specialists in art and development, but the game design itself really requires a broader individual or group to bring it together.
Thanks for share.
Like even when I make a small scale RPG, my damage formulas and scaling and so on are my bread and butter. I obsess over them. I would accept being wrong about the resulting experience and having to adjust, but to me it's really unacceptable to have oversights and mistakes there.
So something like how broken resistance was, or different types of scaling implemented wrong. That's not disagreeing about the outcome. That's clear oversight in what I consider the heart of the game.
Or for example how the mount's controls with mouse were just remapped from joystick and you have to put your mouse at the edge of the screen to go fast.
I don't understand how to reconcile such a big and well made game with many baffling oversights in the game's core systems. Feels competent and amateurish at the same time.
Consider a chair that is sturdy and load-bearing, while still being lightweight, but it wildly uncomfortable to sit on.
The art direction is great, they produced a massive amount of assets at a very high quality, the story is great by Diablo standards, the marketing was good, the tech is pretty good.
The player experience is still very good and very high quality. I usually find the only people who complain are the hardcore players. It makes sense because once you've run out of campaign content, you really have to rely on those core mechanics to keep the game going.
A lot of the people involved are competent and have a good track record. I just don't understand why the foundations are shaky. They're not even innovating so much on what an ARPG is supposed to be.
I think many of us did enjoy Diablo 4 but the systems that make an ARPG shine fell through the cracks which is surprising given Blizzard defined the category.
I am very much am aware they aren't the same people running the show now from Blizzard North, and they had some major team shake ups in D4s dev cycle, so that was part of the problem but it's hard not to have high expectations.
It's just those pieces or connections in-between the excellent departmental work that fell apart, along with the foundation of an ARPG being flawed, that a gaming generalist could have solved to bring it all together.
I imagine things would look quite different now if that generalist person or persons on D4 played ARPGs for decades, had Diablo DNA infused in them with equal experience in game design and systems, and had some authority they could have made the best ARPG all over again since D2 given how good D4 is in so many ways. They equally learned a lot from D3 and it's shortcomings but that knowledge all felt lost in D4.
In science something similar happened. There was a time where some people knew pretty much in every science field, chemistry, physics, biology etc. but the amount of knowledge grew and it become impossible to know enough in every field.
Maybe game development isn't any different. You either become an expert in on field or your knowledge is to superficial to actually help.
And then, once you've spend that time using the tools, you can refactor the SpeechOptionNode to incorporate skill checks without attaching a script, or whatever other obscure tooling problems the writers are dealing with.
The main issue is mindset, a programmer might cut short a design to save himself time, an artist might simplify the gameplay to not put in as much effort, and a musician or writer might not focus as much on the visuals because they get caught up in how it should it sound/read etc. But they certainly can take on other roles. If you have any experience with start-ups and indies or freelancing you might find yourself take over roles you didn't expect originally. On smaller teams this goes a lot further.
Nothing wrong with them at all but the scale is just totally different, it’s like running a small bakery shop vs an industrial factory churning out bread for supermarkets
The best design solutions are usually compromises, but when the trade-off crosses into different disciplines (e.g., mechanical vs electrical), it can get tricky.
At minimum, you need people who can even recognize when such trades exist. It's better if you can find translators: the rare person with enough expertise in relevant domains to get to get two teams to work together efficiently.
In my own career, I've often found it useful to pick up cross-discipline skills, at least enough basics to fill one of the above roles.
― Motoko Kusanagi ⎯ Ghost in the Shell (1995)
These are actually at least three (likely four or more) areas of expertise: expertise in two domains, expertise to communicate well with both sides, and the expertise to "translate" from each side to the colleagues of the other one.
For example, while I do claim to have expertise in multiple areas, it - I claim exactly because of this reason - often happened in the past that both sides told me that I somehow do not speak "their" language, but what they claim to be the one of the other side. Also, I have a tendency to find "weird" connections between both areas that both sides have difficulties understanding. :-(
I'm a fella who has a smattering mix of - systems engineering, systems administration, network engineering and administration, database administration, software development, analog and digital communications systems design and administration (from the circuit level up), UI/UX design, process engineering, documentation, RF Engineering, project management, and a little bit of marketing.
The only thing I'm missing is the financial aspects of business, but from the technical, to the structural, to basically anything technology or people related - I know enough of it to step in and be competent in most situations.
To quote Heinlein "Specialization is for insects."
Some humans would find immense happiness trying to be the queen themselves. Others, to live like a cat. By themselves, with no one to tell them what to do. But our entire modern society is built upon and favours the busy, mindless drone. The education system shapes you like one, and isn't concerned to help you figure out your ideal place in the ecosystem.
The game has since sold more than 20 million copies and some publications have called it the greatest game of all time. An indie game by one person!
Generalists absolutely can succeed today!
Stardew Valley guy did pretty much everything himself. Game design, programming, music, art.
Which, it's an indie game... but clearly if he can do the whole project himself, someone on a AAA team can have cross-disciplinary skills and translate.
> Which, it's an indie game... but clearly if he can do the whole project himself, someone on a AAA team can have cross-disciplinary skills and translate.
But if one man can do 100% of an indie game, surely you can end up with a few people who are generalists on a AAA game team that are deep in one area and basically competent in another couple to act as liasions.
It's a scaling problem. Similar to how single theaded, multi-threading, and distributed computing all have different schemes to solve the same problem, so does a single/small team vs. a studio of hundreds of devs. You won't get a satisfactory output if you use a single threaded technique to solve a problem requiring distributed scale. You'll either take prohibitively long, and/or produce the wrong/redundant results as solutions clash into each other with no syncronization.
I think instead of trying to fit more generalists into a AAA pipeline, we should encourage more generalists to optimize for a small team and deliver leaner projects that can still turn heads.
Adding more edges, where artists can talk to programmers more directly and meaningfully without it happening through a mediated hierarchy, improve throughput.
Of course, having the ability for subteams to work together more effectively doesn't mean you should always use it or abandon trying to manage the project through normal means.
Certainly, one can be a jack of all trades, and certainly, there's value in being so, but you're giving up mastery which has its own value, and comparatively, having multiple specialists collaborate is better than having multiple generalists.
On the indie side, a generalist indie needs less of what a big game needs. You probably don't need a graphics programmer at all for an indie game, for example.
Going in the other direction, the increasingly absurd size of companies is one thing I really just can't wrap my head around at all. The final output just doesn't seem to justify these massive teams and budgets when contrasted against vastly smaller projects, like Valheim.
Within a year HR approached me to change it because “it confuses people”.
When asked who was confused, … they didn’t answer.
They came back every year with the same request until I left.
Once a certain size is hit in an organization, the org chat and job titles develop an inescapable gravitational pull that only very senior / unreasonably successful people can escape over time.
The heart of the problem is that at some point, you need to make decisions and solve problems where disciplines interact. In the end, you cannot really make an informed decision, or reliability solve problems in these situations without some party actually being able to grasp the cross-disciplinary picture.
In the absence of actual generalists, you usually end up with ad-hoc education - crash courses where each discipline tries to educate the other (and the PM), so that they can try to get something hammered out. And this can actually be quite successful - this happens all the time, and works out very often. In fact, this is probably where many generalists (or dual-hatters) organically arise - you get a crash course in some other area while trying to figure out a problem, and then the next time you get tapped because you already have some knowledge, and you find it kinda interesting and the next thing you know, you're a generalist of sorts.
I think it's quite true that in large projects you can't possibly have a generalist who actually knows -everything-. But it's also quite possible to have a good bench of dual-hatters (or dunno... 3-4 hatters), who can go a long way to relieving issues.
And back when I worked at Microsoft, the PMs were call-pattern experts, too, or at least the great ones were.
If I had to guess, generalists look very unfocused to managers and recruiters and are often not considered for TL positions.
I've worked in medium to large companies (3k-200k employees) and the smaller companies what is the responsibility of a team would be an entire org in the larger ones.
Even getting all the experts in a room is hard when it's the TL and Manager from half a dozen teams.
Games require a unique combination and overlap of creative art and software engineering, something that is hard to measure and quantify... and that is why I think it becomes evident when the leadership of a company is blind to the creative element, maximizing short term business value instead of long term creative value (which arguably has more business value, but it requires passion and love for the craft).
Or how instead of designing emergent systems, most games just slap modifiers on everything and "balance" that way. +20% bow reload speed and +25% speed, anyone?
The 16- through 64- bit Zeldas, for their clunkiness, were compellingly immersive. I don't think it's just nostalgia; I've played indie games where the UI is part of a whole that sucks you in. This is true even if the UI has a modern feel, like Hollow Knight, which felt overpolished in its presentation and interface, to me at least, in a way that was off-kilter -- yet something about it sucks you in instead of pulling you out, being part of the world's presentation and in an aesthetically transitive sense part of the world itself. There's something about these modern AAA games where you feel like you're not playing a game but rather a Skinner box containing a game. It's just not the same. I can't exactly put my finger on it.
The game UI should feel part of the game (good examples: SimCity 4 and Sid Meier's Pirates) and not feel out of place.
In-game advertisements are also off-putting unless they fit thematically and stylistically into the game.
edit: your comment was much better when it was more verbose!<3
When I read a book in a modern typeface, it does not remove me from the story in the book. I do not need to read a world war II history book in Fraktur to get into it. Maybe, like typography, it has more to do with what one is used to?
It is debatable what happens in our heads. When we play one of the earlier pixel Zeldas, one might argue our mind is filling in all the details with higher resolution images. I can still agree with this argument, though.
In terms of the UI, the game controller is also a soulless interface, and I guess one could claim it gets in the way of experiencing the game.
My opinion is simply that the UI being disjointed from the game world experience might make it more or less enjoyable, depending on experience.
As an example, I prefer a file > save checkpoint on a game, that gets out of the way, as opposed to one of those games where you have to go to a special place to save. In that case, for me, the immersion is ruined because I am constantly reminded that I cannot save the game state. Other players might think a File > save checkpoint menu is a flat soulless ui and they prefer the in-game integration.
The "soulless UI" effect happens when developers try to use flat design in a game where it doesn't work with the aesthetic. A game set in the future? of course. A game set in 1700? hell no. and also if the game uses too many numbers. If there are many "+20% this, +10% that" that also feels unimmersive so the player is dragged out of the game world.
I do love the idea of being able to go up and down though it feels a little awkward to me as I normally have to go way up to go down into the underground due to a lot of level design. I kinda wish they put some more cliffs nearby to save me the time. I like the underground a lot though and find that part the most fun.
I kind of wonder if something internal happened that the original article is describing. Side NPCs are designed by game designers, so probably written by them or people close to them. The game's concept is probably worked on closely with the top people in charge. But for the main story's execution, one might imagine there's no organizational requirement that the guys in charge have to have a clue how to write or direct a game or even any clue how to write or direct at all.
Everyone says that to get ahead you need to specialise, but this does not come naturally to me.
I think the points people have raised about CVs of generalists often losing out to those of a specialists is true, mostly in terms of getting through the initial screening by the non-technical HR person.
I know a lot of people who have almost zero understanding of what their colleagues are doing and that creates a lot of unneccesary friction during collaboration.
Gamedev has become so highly specialized due to the lackluster quality games used to have? (There are plenty of ps1 to ps2 games that handles like garbage despite them being good games).
That has I think resulted in more "quality" of AAA games but everything is at best a very predictable bland grey goop.
* Generalist, has worked in many different areas on different types of software
* Thinks that's a bad thing
And I do think that's a bad thing, from a career-marketability standpoint anyway. Back in the 80s or 90s, you could get away with being a "guy who knows computers" and they would pretty much stick you anywhere they needed someone who knows computers. And you had to sort of muddle through, figure out what your colleagues needed to do and how to apply the technologies you had on hand to make it easier. Starting in about the 2000s, the enterprise world especially started demanding specialists. They had a J2EE application that needed maintenance and if you didn't know your J2EE backwards and forwards, I'm sorry, there's just no place for you, BOLIYFE[0].
I had spent four years in a robotics company, and then some time at a server appliance company, and the hot thing when I found myself in need of work after those stints was "cloud". If you weren't up on cloud, the industry just couldn't use you. Résumé driven development became a must if you wanted to make a living in the industry.
And now those chickens have come home to roost in the gaming industry. Because game development isn't six guys in a strip-mall office messing around, figuring out a way to create something that thrills them at least as much as Mario did. Not anymore. I'll come out and say it: in terms of complexity, budget, and team size, games are enterprise software. Each game is an enormous project, often a distributed system, involving the contributions of hundreds. Games have passed the complexity threshold beyond which no one person can hold even a decent approximation of bigger than a tiny sliver of the whole system in their heads. The dev team is likely to be divided into smaller squads, each of which specialize in a narrow area of the game -- because it has to be. Because each of those areas, alone, is going to be more complex than can be handled by a single person, let alone a small team. And in order to not lose productivity onboarding new people, those people have to have demonstrated expertise in the narrow subset of the whole game whose team they aspire to join, typically by being credited on at least one AAA game title in that area. And generalists, unless they are literally John Carmack or at his level, don't have the deep knowledge it takes to be useful on, say, a AAA game's sound team, or level design team, or gameplay mechanics team, or scenario team.
And so we have today the situation where a Hackernews tells a promising kid, put down that homebrew retro console you built out of a Raspberry Pi and learn Unity and Unreal, because that's where you can make a real contribution to the field[1]. And he's right. The industry has changed how it works since the 90s, so in order to stay relevant in the industry you have to change too. Change with the times, or get left behind. Adapt, or die. It was always ever thus.
[0] "Best Of Luck In Your Future Endeavors"