• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Tim Cain (co-creator of Fallout) described the problem with modern game development

Topher

Gold Member
Did you even watch the video? He described the problem, told that he wrote similar code before and a bunch of stuff had been already in place.

The truth is that the modern level of developers is quite quite low. And from my observation, a lot of good developers these days are leaving for science area (not data science but more into mathematical sciences, physics stuff like that).

As a developer (non-gaming), what Tim Cain described should be obvious to just about any coder. Certainly shouldn't be difficult for any developer who works in video games. It is literally incrementing values in a list. He even said the callbacks are already there so all the developer has to do is take the value damage and who delivered (both of which would be passed into the callback) and store those values in an enumerable data structure. The item in that data structure with the highest damage value will be the NPCs target. That's it. Four weeks? No way.

Bullshitters everywhere.

Just got to hope someone has the knowledge or intuition to call them out on it. I don’t have tech experience so what do I know if a coder says 4 weeks?

But if I’m setting company targets and a sales guy says the target of $10M I calculated for him is too high due to falling sales trends, I’d just check what’s going on. We’ll according to me your sales trend is fine, you got a new product line launching and supply chain had fully caught up so no more out of stocks. Add all that together and it looks like $10M to me. Prove me wrong or fuck off.

Nice try at getting a negative growth target.

Exactly. And this is an old tale, ain't it? Happens in every day life as well. The guy who is fixing my car could be bullshiting me to hell and back and I sure wouldn't be able to challenge him on it. He's taking advantage of my ignorance. You just hope the guy you hire is honest. But if you are not ignorant and are able to call out a guy who is trying to overcharge you then who the hell isn't going to do that?

Here, Tim Cain knows better. And that's why that developer tucked his tail and ran when Cain said he had to explain why this simple task was going to take four weeks.
 
Last edited:

StreetsofBeige

Gold Member
As a developer (non-gaming), what Tim Cain described should be obvious to just about any coder. Certainly shouldn't be difficult for any developer who works in video games. It is literally incrementing values in a list. He even said the callbacks are already there so all the developer has to do is take the value damage and who delivered (both of which would be passed into the callback) and store those values in an enumerable data structure. The item in that data structure with the highest damage value will be the NPCs target. That's it. Four weeks? No way.



Exactly. And this is an old tale, ain't it? Happens in every day life as well. The guy who is fixing my car could be bullshiting me to hell and back and I sure wouldn't be able to challenge him on it. He's taking advantage of my ignorance. You just hope the guy you hire is honest. But if you are not ignorant and are able to call out a guy who is trying to overcharge you then who the hell isn't going to do that?

Here, Tim Cain knows better. And that's why that developer tucked his tail and ran when Cain said he had to explain why this simple task was going to take four weeks.
Yup. I guess that’s the difference between coding and something like my experience of sales targets. Same self serving BS but just different sandbagging,

The cider guy will sandbag on time needed. But the people I deal with time constraints aren’t an issue. Nobody says I need 4 weeks to do something. Almost all normal day to day shit can be done fast.

But the sandbagging is financial related.

Same BS. Different shell,
 

bender

What time is it?
It is literally incrementing values in a list. He even said the callbacks are already there so all the developer has to do is take the value damage and who delivered (both of which would be passed into the callback) and store those values in an enumerable data structure. The item in that data structure with the highest damage value will be the NPCs target.
E8d1.gif
 

kiphalfton

Member
Easy to say “I could do that myself in 45 minutes” but if your boss asks you how long it will take to finish you better build in some breathing room in case something goes wrong.

The programmer is going to be responsible to make sure it doesn’t cause any issues and will need to thoroughly test their solution for bugs as well, that’s obviously gonna take more than 45 minutes.

I've had a couple bosses/professors/etc. who said "you should be able to do this in [1/10th the amount of time it actually takes]"

Guess they're a different breed, of hyper efficient beings.

Seriously though, you always sand bag the numbers when asked how long it will take you. Otherwise you're just setting a bad precedent if you're too efficient. Hit a good happy medium, and you'll be sitting pretty.
 

SHA

Member


He tells some interesting stories. Including one where a piece of code he wanted written to improve enemy AI, that he said would take himself 45 minutes to write, when he gave the task to a programmer, told him it would take 4 weeks, and when he pushed back, angrily walked out on him, and the lead programmer had to calm to situation by negotiating it down to 2 weeks.
This was for the development of Outer Worlds.
Now it makes sense why nobody can make a good game in less than 5 years.

Anyway he has a channel on YouTube now and people who are interested in game development would find what he has to say fascinating.

That's called experience, plus, it's different when you do your own thing without being told what to do, it's called self taught, being humble doesn't help the situation, that's how the indie business started.
 
Last edited:

Tomeru

Member
Easy to say “I could do that myself in 45 minutes” but if your boss asks you how long it will take to finish you better build in some breathing room in case something goes wrong.

The programmer is going to be responsible to make sure it doesn’t cause any issues and will need to thoroughly test their solution for bugs as well, that’s obviously gonna take more than 45 minutes.
Its the same if your boss writes the code. It'll still need to be tested. And he did say it was dome in two weeks. Tim didnt laugh at that developer or fired him.
 

rofif

Can’t Git Gud
Of course the dev takes 5 years now because some of these snowflake devs are just engine editor users. Something even I could do.
There are so many studios now that of course you dont have so many talented devs. Everyone makes games.
 

supernova8

Banned
From my experience, there is often a huge gulf in terms of the ability of the most skilled (and intelligent) engineers and those who are new or just not very good. I remember at an old job, one of the software teams (5 people) was stuck on something for about a week and then our CEO (who doubled up as an engineer when needed - an older, less well-off Elon Musk perhaps) confronted them about why they were taking so long and then literally fixed the issue in a single evening.

I remember seeing customers hiring cheaper (often India-based) software houses to fix their problems only to have an army of incompetent software engineers fail to do what our company could do much quicker with a fraction of the engineering headcount. Our day rates were muuuch higher (to the point that a lot of prospects would initially laugh during sales pitches) but the overall cost would end up being the same or less (for the work to be done faster). Remember my boss at the time saying something like "trying to get an incompetent engineer to fix a difficult software problem is like seeing a person struggling in the water and then expecting someone who also can't swim to jump in and not drown along with them".

On that basis, I'd be inclined to believe him if he said it would take him 45 mins vs 4 weeks for the other developer. Even if he underestimated it, let's say it takes him 2 days to fix it, that's still a huge difference and illustrates the same point. No point getting hung up specifically on the 45 min bit.
 

DenchDeckard

Moderated wildly
I have a couple of friends who work at solicitors and they were telling me how they purposefully take weeks to respond to things and sort certain elements out as this is how it's been laid out for years to cover the charges they cost. Their is an expected time been set from years of work across the world for these aspects of their workload and its basically just bollix, to justify what they charge.

I found it interesting.

They were explaining how doing some aspects of work too fast can come across as negative, like you haven't taken the time to look at it properly...or your business is too quiet and isn't busy with other things.

Thoughtful.

I see this is my work. We are losing the get up and go across many aspects of businesses I guess.

It feels good to be old school in that regard.

That developers just trying to have far too much of an easy time and wants to justify their wage.
 
Last edited:

Denton

Member
I’m not saying the developers are all rock stars, but just because the dude said he thought it should take 45 minutes doesn’t mean much. The guy immersed in the code thought it would take longer, and overestimated how long it would take on top of the rest of his work load. Do you know of any changes to a game’s AI that could be fully implemented and tested in 45 minutes? Seems like wishful thinking.
Did you watch the god damn video?

It was about 10 lines of code. It was not meant to be anything for the game build itself that would need some super serious bugfixing. Just absolute basic AI code for testing. And Tim already coded that exact thing multiple times in the past, hence his awareness about how quickly it could be done.
 

hyperbertha

Member
You guys have gotta stop blaming "diversity" for everything.

Dev teams are diverse because the developer workforce is diverse. It's a worldwide profession.
These are companies trying to make money. they're gonna hire the based on the best combination of technical skills and social skills.

If you're getting declined for jobs, it's not because of your race or gender or whatever. It's because you aren't as good as you think you are.

And if you're one of those redpill edgelords that "find it difficult to work alongside ethnic or trans people" then there's no surprise you don't get hired.
Why would anyone want to hire you? It doesn't matter how good of a programmer you think you are if you can't work with all sorts of other people.
So you don't know how diversity hires work...
 

Panajev2001a

GAF's Pleasant Genius
So you don't know how diversity hires work...
I do not think it is that really, biggest push is DevEx… in ideal terms it is good, but fundamentally we tried to hire cheaper engineers to staff bigger teams and dumbing things down and standardising them by force to try to make it work. Do we need UI rendered with web browsers in games? Do we need JavaScript in the backend? Do we need web based frameworks like Electron powering up all the apps on Desktop? Ok, so we allowed a lot less experienced devs in, people that may not care about what they are doing and look for a factory like job or code-artists… and now we pay for it too.

Not all the improvements in tooling and languages and middleware was bad mind you, but it was not all roses either.
 
One of the projects I work on is modeling systems. I can do what one person on our teams takes a month to do poorly in a few hours. These people do exist but everyone is shocked by it. It is a rare case of utter incompetence. I highly doubt that this one whiner is representative of the industry. I also assume that what Tim is talking about is in relation to his experience coding very simple games compared to what we have today. Without knowing the details, I would expect that altering the game's AI significantly would actually impact a lot of systems and while the coding change might be relatively straight forward, integrating that into the games systems would require a lot of tuning.

I am going to put this out there. I think the real problem is these old leaders have no fucking clue how much more complex the games are today and how difficult it is to make them fun and engaging to gamers that have been gaming for decades. They are still designing things like it is 2002 and think they are gods when really there design ideas suck. Tod Howard led a pile of shit that is really just a fresh layer of diarrhea on top of decades of the hardened turds that is the old Morrowind / Oblivion system. To actually make Starfield good, Bethesda would need to bring in someone who is not an arrogant prick who did something good 20 years ago when games could be rolled out in a year or 18 months. These assholes have way too much influence and no concept of what innovation or leadership is.
Did you even watch the video? He was talking about the Outer Worlds...Not exactly the game from 2002.

I'd also say that any studio who hired someone that useless for that role, would surely have made the same catastrophic error in their other hires. And if that's the case, how the fuck would literally anything be getting done anywhere? And even then it would be a failure in management more than anything.
Well, modern software works and optimized like crap for a reason :messenger_tears_of_joy:

Regarding hiring is another topic. For example in my company, HRs are literally looking for API developers. That's it. API developers. Why? Because according to some research "companies with APIs are more successful than others". Yeah, that's all. Yet HRs themselves are not developers and do not understand what they expect. "He needs to know Azure", "He needs to know how to write API". But people literally can spend years using some visual tools to build some generic APIs or just doing plain coding of endpoints, without understanding how it works. And they can pass the interview...Because they worked with API!
 
I'm fascinated by the stories about the whiteboard and programmer saying it would take 4 weeks. I'd really like to know what their reasons were. Instead of blaming diversity hires and sjws like some here do, I suspect that some corporate mandated performance metrics are in the play
 

Clear

CliffyB's Cock Holster
Well, modern software works and optimized like crap for a reason :messenger_tears_of_joy:

The obvious reason is that its more complex and feature-rich than its ever been.

Yes the tools and middleware are better, hardware and infrastructure is faster and more capable, but there's no getting around the reality that everything is more complicated and technologically advanced even if much of the nuts and bolts complexity is hidden inside of "black box" API's and applications.

Which ironically makes debugging far more challenging than if you're working with a bespoke self-written code stack!
 
Is anyone surprised at all? Back then this industry was dominated by geeks/gamers who knew and loved what they did. Now you have diversity hires and political propagandists instead. The industry will implode sooner or later, the signs are obvious.
Yep, just look at Hollywood movies since like 2006 or so. It's been downhill. There only about a handful of directors that use actual filmmaking techniques instead of fake cameras and fake sets. Not to mention the writing also sucks in a majority of them.
 

Roni

Gold Member
Including one where a piece of code he wanted written to improve enemy AI, that he said would take himself 45 minutes to write, when he gave the task to a programmer, told him it would take 4 weeks, and when he pushed back, angrily walked out on him, and the lead programmer had to calm to situation by negotiating it down to 2 weeks.
This isn't exclusive to game development, we educate children poorly in a world in which knowledge is more important than ever. This has resulted in talent erosion in all industries. The new generation can't keep up with the previous one and everything is either slowing down or decaying.
If he could do it in 45 minutes, why didn't he do it himself then?
That's a VERY slippery slope. Start doing that and suddenly you're developing the game by yourself while the team watches. He could have offered to coach the programmer if he was a direct superior. But do it himself is unacceptable if he's paying the guy.
 
Last edited:
The obvious reason is that its more complex and feature-rich than its ever been.
Nah. A lot of modern software is built on top of the solid foundation, created by the people before them and just began to bloat with functionality being added without understanding how the underline system works. It is just easier to add some features, them blame weak computers for not working properly or not having enough resources.

Yes the tools and middleware are better, hardware and infrastructure is faster and more capable, but there's no getting around the reality that everything is more complicated and technologically advanced even if much of the nuts and bolts complexity is hidden inside of "black box" API's and applications.
The thing is that the same tools, middleware, networking while indeed are more capable, they are not better in a sense that their creators are also people similar to those who are unable to optimize and design software applications properly.

Yep, just look at Hollywood movies since like 2006 or so. It's been downhill. There only about a handful of directors that use actual filmmaking techniques instead of fake cameras and fake sets. Not to mention the writing also sucks in a majority of them.
Yeah. Now we have so many directors yet - due to amount of directors - we have more trash than ever.
 
Last edited:

Sorcerer

Member
Okay, if the code only takes 45 minutes to write, why all the back and forth with the programer and settle on 2-weeks? Why not just sit down for the half-hour plus and bang out in the name of efficiency and to prove his point? Maybe sit down the programer and teach him a thing or two. They spent more time argueing this than that supposed 45 minutes. LOL!!!
It doesn't add up.
 
Last edited:

FoxMcChief

Gold Member
If he could do it in 45 minutes, why didn't he do it himself then?
So you’re saying his employees are worthless and should be out of the job? Dude has more important things he should be working on.

You don’t ask the higher ups to deal with low level issues. There’s a reason higher ups don’t restock break rooms or clean the coffee maker.
 
If he could do it in 45 minutes, why didn't he do it himself then?
Okay, if the code only takes 45 minutes to write, why all the back and forth with the programer and settle on 2-weeks? Why not just sit down for the half-hour plus and bang out in the name of efficiency and to prove his point? Maybe sit down the programer and teach him a thing or two. They spent more time argueing this than that supposed 45 minutes. LOL!!!
It doesn't add up.
Did you listen to the part of the video where he both offered to do it himself and wrote out the lines of code it'd require to the manager, and the manager shot him down? It doesn't seem like you did.

Both the video and this thread are really interesting views into people and they way they respond to efficiency/work ethic, it seems.
 
Last edited:

FoxMcChief

Gold Member
Did you listen to the part of the video where he both offered to do it himself and wrote out the lines of code it'd require to the manager, and the manager shot him down? It doesn't seem like you did.

Both the video and this thread are really interesting views into people and they way they respond to efficiency/work ethic, it seems.
I wonder if the this is a generational thing here as well. Are the posters against Tim’s younger than 30? Are those that side with Tim 35 and older?

Are the ones against Tim low level employees themselves, and the ones siding with Tim in higher levels or in management?
 

sankt-Antonio

:^)--?-<
So in reality its not a 45min job, but a 2-3day job. The Dev he asked most likely has ten people on his ass with these "45min" worth of code change requests. And off course everyone is flagging their shit as super priority.

Fun fact, agile software development was created to fix that problem, but led to new set's of problems where Devs would overcalculate time needed to overpace sprints, making development not really much faster, because as it turned out there is no way to tell how long software changes need before sitting down and actually starting doing it.
 
Outsourcing is a plague, it increases budgets exponentially and there is no oversight on quality anymore. It needs to stop and publishers/game dev studios should do everything inhouse again like it was in the past with the exception for example CG movies or Music or anything else that doesn't have any coding.
 

Clear

CliffyB's Cock Holster
Nah. A lot of modern software is built on top of the solid foundation, created by the people before them and just began to bloat with functionality being added without understanding how the underline system works. It is just easier to add some features, them blame weak computers for not working properly or not having enough resources.

If you didn't build the foundation, you don't know where the cracks and stress points are. You are left to assume that everything is uniformly solid and free from weirdness... until you run up against those rare corner-cases where things don't behave like you expect.

That's the issue I was getting at.

Most teams, and certainly very few individuals have the time and wherewithal to build an engine the equal of Unreal. Its such a monumental amount of work in addition to achieving the project goals, that its a no-brainer to simply buy into an existing piece of tech. And of course hiring people who are familiar with it is a whole lot less hassle than training up people on a bespoke engine with spotty documentation.

And of course if you hire a coder to write a crucial component, and they subsequently leave... then you're really up against it because then somebody needs to maintain and modify this alien code-base.



The thing is that the same tools, middleware, networking while indeed are more capable, they are not better in a sense that their creators are also people similar to those who are unable to optimize and design software applications properly.

Define "optimize". What does that mean in real-terms? Optimal for performance, code-size, versatility, specificity, what?
How about optimal for work-flow, getting a quick and dirty fix in-place so as not to hold up the scripters and mission designers can get on with testing their contributions for function and playability?

You really do not have a fucking clue what you're talking about. Everything is connected and as the code gets increasingly layered and complicated those connections and contingencies multiply.

Optimal is not just running perfect with every setting jacked to the max on your 4090!
 
Outsourcing is a plague, it increases budgets exponentially and there is no oversight on quality anymore. It needs to stop and publishers/game dev studios should do everything inhouse again like it was in the past with the exception for example CG movies or Music or anything else that doesn't have any coding.
Looking at Ubisoft - even having stuff in house does not help much.

If you didn't build the foundation, you don't know where the cracks and stress points are. You are left to assume that everything is uniformly solid and free from weirdness... until you run up against those rare corner-cases where things don't behave like you expect.
But that's the thing - people should learn how those systems work. You have go into the codebase and try to understand that.

Most teams, and certainly very few individuals have the time and wherewithal to build an engine the equal of Unreal. Its such a monumental amount of work in addition to achieving the project goals, that its a no-brainer to simply buy into an existing piece of tech. And of course hiring people who are familiar with it is a whole lot less hassle than training up people on a bespoke engine with spotty documentation.
But even with UE, the developer should have a normal level of understand how game engines work. Not just "put here and there, oh it works, too slow? ask for bigger requirements then".

And of course if you hire a coder to write a crucial component, and they subsequently leave... then you're really up against it because then somebody needs to maintain and modify this alien code-base.
But that's where the problem of lack of talent comes in - there are just not enough developers that can replace the person that left.

Define "optimize". What does that mean in real-terms? Optimal for performance, code-size, versatility, specificity, what?
How about optimal for work-flow, getting a quick and dirty fix in-place so as not to hold up the scripters and mission designers can get on with testing their contributions for function and playability?
With such approach, no wonder that the game performance is becoming worse and worse. "Hey, it is 720p but we have Nanite".

And optimal performance is not mutually exclusive with code size (especially when people copy paste the same code from one place to another without thinking) or versatility and so on.

You really do not have a fucking clue what you're talking about. Everything is connected and as the code gets increasingly layered and complicated those connections and contingencies multiply.
You are trying to absolve developers from any responsibility. "oh, the system is too complex, they can't help with that". No.
 
Last edited:

Topher

Gold Member
So in reality its not a 45min job, but a 2-3day job. The Dev he asked most likely has ten people on his ass with these "45min" worth of code change requests. And off course everyone is flagging their shit as super priority.

Fun fact, agile software development was created to fix that problem, but led to new set's of problems where Devs would overcalculate time needed to overpace sprints, making development not really much faster, because as it turned out there is no way to tell how long software changes need before sitting down and actually starting doing it.

Yeah, but assigning effort to each task to be accomplished in a sprint is typically a group effort. That's why some use scrum cards in sprint planning. Devs get better at estimating level of effort the more they do it.

Now, in this case, the developer seems to have been saying that this single task would take 4 weeks all on its own. That's how I took what was said. Could be a miscommunication, however, and he was really incorporating his current workload into that estimate. Hard to say, but I've known people who tried to pull this overestimation shit before. Sucks because taking too much time on any one task means the rest of the team has to work harder to pick up the slack.
 
Could be a miscommunication, however, and he was really incorporating his current workload into that estimate
But who does really estimate the task like that? I mean if asking to do something you usually don't say that it will take 4 weeks for me to finish because "I am busy with other stuff". You give fair estimations and then just tell when it can be completed or discussed it with the team.
 
Last edited:
I wonder if the this is a generational thing here as well. Are the posters against Tim’s younger than 30? Are those that side with Tim 35 and older?

Are the ones against Tim low level employees themselves, and the ones siding with Tim in higher levels or in management?
Im in between on all of those, haha. I am in a position where I *COULD* tell people that stuff that I do has a way longer turnaround than it really does and they would believe me, but I avoid that and get it done really swiftly.

It leads to a weird position where I have a good deal of down time, which I think is the thing people are afraid of. They realize that if they don't pad out their time requirements, and do things very quickly, it'll often appear that they have a lot of free time which they aren't filling, and they don't want to be seen as dead weight.

That is exacerbated when you are in a company which has overhired (sometimes by multiple orders of magnitude) and it is difficult to find useful tasks to work on in the first place.

I'm not sure when this became an issue in companies. I suppose the movie Office Space provides a hint that it has been going on for a while.
 

Fake

Member
If you are a competent developer in it for the money you're better off going into Fintech. Nobody really goes into gamedev for the money.

I sure some folks do go to the game dev for money.

Why you guys act like this job don't pay well or something?
 

Three

Member
I sure some folks do go to the game dev for money.

Why you guys act like this job don't pay well or something?
It doesn't pay badly but there are better paying jobs for programmers. People don't go into teaching at uni for the pay either but it doesn't mean it pays badly.
 

Topher

Gold Member
I sure some folks do go to the game dev for money.

Why you guys act like this job don't pay well or something?

They do but I think the point being made is that their programming skills would be more lucrative in other non-gaming fields.
 

Fake

Member
They do but I think the point being made is that their programming skills would be more lucrative in other non-gaming fields.

Feel like is some sort of low job or something I really don't get.

If people tell to be that was a hard job, totally fine, but no paid well? No fucking way.

I do agree with your point, there are sure better places for the skill some game devs have, but I'll never gonna fall into that 'low paid' premisse people try to paint.
 
Last edited:

Clear

CliffyB's Cock Holster
But that's the thing - people should learn how those systems work. You have go into the codebase and try to understand that.

You make this sound trivial when its an extremely lengthy process to learn to ins an outs of a very complex codebase written by somebody else. Its also a highly specialized task because not everyone is an engine coder.

But even with UE, the developer should have a normal level of understand how game engines work. Not just "put here and there, oh it works, too slow? ask for bigger requirements then".

What's a "normal level of understanding"? You can hire someone with a high level of competence in graphics coding, but what if they've mainly had experience with say Frostbite or some other engine? That's not especially helpful when the issue is one of experience with the foibles of say UE.

But that's where the problem of lack of talent comes in - there are just not enough developers that can replace the person that left.

You're missing the point, if a piece of tech is basically one-of-a-kind, there is no direct replacement. You need to give someone the time to explore and dissect how it works, and this all takes time which is the main limiting factor in commercial dev.

With such approach, no wonder that the game performance is becoming worse and worse. "Hey, it is 720p but we have Nanite".

This isn't an issue, its a choice. You can trade off detail for resolution, arguably this is the smart and effective choice given the advancements made in recent ears to upscaling tech.

Your desire for higher resolutions isn't relevant.

and optimal performance is not mutually exclusive with code size (especially when people copy paste the same code from one place to another without thinking) or versatility and so on.

Actually its highly relevant. Unrolled loops for instance are way more efficient than any other method due to avoiding branch overhead but at a cost to code size.

You are trying to absolve developers from any responsibility. "oh, the system is too complex, they can't help with that". No.

No. I'm pointing out that everything -including all the small things you take for granted as being standard- has a cost. And each small loss of efficiency mounts up.

There are also cases where resource utilization carries massive penalties. If you can keep data cached (in vram for instance) you don't need to constantly shuffle data in-and-out. Its why memory leaks are so dangerous to overall performance because given enough time, even a minor leak can be catastrophic.

And the real kick in the balls is that leaks can happen anywhere, like in a subsystem like an event-script parser, but the consequences can manifest elsewhere like in the rendering pipeline. This is a perfect illustration of the cost of complexity because when so many processes are running concurrently its harder and slower to pin-down the specific source of the issue.
 

hussar16

Member


He tells some interesting stories. Including one where a piece of code he wanted written to improve enemy AI, that he said would take himself 45 minutes to write, when he gave the task to a programmer, told him it would take 4 weeks, and when he pushed back, angrily walked out on him, and the lead programmer had to calm to situation by negotiating it down to 2 weeks.
This was for the development of Outer Worlds.
Now it makes sense why nobody can make a good game in less than 5 years.

Anyway he has a channel on YouTube now and people who are interested in game development would find what he has to say fascinating.

Most people now work for the money and don't care for anything else but themselves
 
You make this sound trivial when its an extremely lengthy process to learn to ins an outs of a very complex codebase written by somebody else. Its also a highly specialized task because not everyone is an engine coder.
It is obvious that people should have the basic awareness and understanding of the whole application (or the game). You don't need to be an expert in that, but it is important to know how it works - as by and large, all the engines are pretty similar architecture wise as most of the concepts literally came from 90s (for games I guess, for software it is more like 60s at this point) etc. Nobody is asking AI developer to go and fix issues in the renderer. You make it sound as if somebody is asking person from one area to be responsible for another area. Obviously Tim Cain did not come to web developer and asked him to implement the change.

It is really appalling that asking for some additional knowledge is some "extra complex challenge". The truth is simple - it is hard to find competent people these days. And the reason for that is that people just don't want to learn. That's all to it.

What's a "normal level of understanding"? You can hire someone with a high level of competence in graphics coding, but what if they've mainly had experience with say Frostbite or some other engine? That's not especially helpful when the issue is one of experience with the foibles of say UE.
What stops the developer to learn and dive deeper into the UE details in order to understand how it works? It is a basic requirement when you need to work with a new technology or technology you are not familiar with. You have to learn it. And I think that - just like with software development - UE is not doing anything that it is a mystery knowledge only the gentlemen club at Epic knows...When I was hired to develop for Azure stack, I did not walk around saying that "I only worked with AWS so I can't do this task or that". You just need to go and learn.

By and large, game development is suffering from the same issue as software development - people flocked there to gain money, bringing more incompetent people who have no interest in what they are doing. (just a regular thing that with quantity, quality drops). And those who did not want to work in that environment, switched to something else and left.
 
Last edited:
They do but I think the point being made is that their programming skills would be more lucrative in other non-gaming fields.
If that’s true, and I don’t know that it is, if would imply gaming programmers are the worst programmers and the left overs of other tech companies.
 

Yoda

Member
As someone who's been in the tech industry for a bit, I can relate. The overall quality of programmers/developers/software engineers has gone down as more people flock to the profession given the low barrier to entry and high incomes. Projects that I know I can do in a month, which I'm told by senior mgmt to delegate out will end up take 2 to 3 months with an "engineer" who has the same YoE I do. Tech used to largely be solely comprised of people who had a natural proclivity towards it (I'd count myself in this group). Now a lot of people who'd have tried getting into finance in the previous decade (I'll admit the thought did cross my mind and I even interned as a quant dev on Wall St) are coming into tech roles. I'm certainly not one to gatekeep people from living prosperous lives, but the hard truth is your average developer is less talented and charges twice as much as the average dev just 5 years prior.
 

Killer8

Member
I sometimes wonder if lowering the bar for people to get into game development via easier to use engines, AI and automation is a double edged sword. I suspect that the time saving measures and the shortcuts might actually end up lengthening development time, because it lessens the need to organically learn the skills to solve complex problems when they pop up. Saving half the time in X and Y area will be for naught if it's making them incompetent enough to spent quadruple the time in Z area. There's just no substitute for hard work and experience.
 

404games

Banned
what Tim is describing (WRT game devs)
is imho "institutionalized" game developers
it's not that they are inexperienced,
but they are just institutionalized to work within a given tool or framework
(and anything outside of that is a huge effort)


to me, this is why Starfield is the way it is, (just a new minor iteration on the same thing)
the whole team had to work hard to build Creation Engine
the designers and artists know how to use the tools
and "innovation" outside of that is "all risk" because
Invented Here

(but that's why they are being surpassed in the industry after years of
going back to the well rather than building something new)
 
I sometimes wonder if lowering the bar for people to get into game development via easier to use engines, AI and automation is a double edged sword. I suspect that the time saving measures and the shortcuts might actually end up lengthening development time, because it lessens the need to organically learn the skills to solve complex problems when they pop up. Saving half the time in X and Y area will be for naught if it's making them incompetent enough to spent quadruple the time in Z area. There's just no substitute for hard work and experience.
I am actually curious AI will affect the development.
 
Top Bottom