Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

You can't git clone a team

virtualize.sh

Post reply on HN
Print view View original

101–110 of 115 posts

Re: You can't git clone a team

#101

Post by johnnyanmac »

jandrewrogers wrote: post #70
cadamsdotcom wrote: post #68

The value of knowledge is plummeting. It can, will, and should be subsumed by LLMs. Consider London’s black cabs. Drivers pass a gruelling exam to prove they know “everything” about one part of London. But, Google Maps puts that knowledge and more on every rideshare driver’s dashboard, for free, no advance study needed. Map knowledge doesn’t make you able to drive. Driving happens in real time - steering, obeying roa…

One of the big issues with LLMs for e.g. systems software is that there are broad domains where training data effectively does not exist. Consequently, you can’t learn much from them of value. It is the blind leading the blind. The lack of training material for LLMs is of course a lack of training material for people too. Some areas of software have a long history of relying almost entirely on an oral tradition to pa…

>This has some advantages but it doesn’t scale and it makes it basically “dark knowledge” for LLMs or people without access to those that know it.

We call this "tribal knowledge" in games. I despise it. It's done for a few reasons:

- NDAs make public knowledge a landmine. And every game studio makes you sign NDAs. Even at the interview stage.

- Churn. No one gets time to really develop expertise as they work on a project for 2-3 years and then layoffs come. Only a relative few become experts, and probably not because of the studio itself

- lack of incentives. Games aren't very connected with acedemia to begin with, despite relying so much on cutting edge tech. So the best resources for sharing such techniques is shafted. This is slowly getting better as more tech conferences talk about games tech, but it's a pretty slow tricke unless you come from one of the largest studios and specifically come in for R&D.

>If you want to get into an area like this, you often need to find a way to spend a lot of time with people that already know it.

All too true. Open source development is one bastion for this, but that's overall why I keep trying to stay in this domain. You literally can't get the knowledge elsewhere. And it's knowledge that directly leads to better looking, more optimal, and less buggy games overall.

Re: You can't git clone a team

#102

Post by ipaddr »

johnnyanmac wrote: post #97
ipaddr wrote: post #52

Earlier quoted context omitted.

Good, that's the way it was until the splitting of roles for commodification. A programmer is more like the Renaissance man who makes it a goal to do everything from different disciplines than a drone who has been trained to do one thing and can only be trusted to do one thing.

It's not commodification, it's acknoledging that tech got exponentially more complex over the decades. just think of your favorite video game character in 2000 and then one in the 2020's and consider how much tech is needed to render, animate, light, and conceptualize it. in 2000 this was all done by maybe one artist and one gamedev, probably making a character with some hundreds of polys at best. now that artist has…

My point was unnecessarily they split the roles. An artist can cover concept and materials. A programmer can do gameplay, graphics, rendering and builds. In fact having people who understand the entire project makes for a better project.

It's like moving from custom built cars to the assembly line where someone's job is putting in one screw. I understand it's cheaper/faster because you can hire anyone unskilled for cheap but cars were all suppose to be identical. Software should be unique (if not just copy the last thing built) but I guess when it comes to major games things are more of factory throwing millions of pixels of characters at existing game engines while copying gameplay of successful games. That's why games are shovelware these days like a netflix original.

Re: You can't git clone a team

#104

Post by constantcrying »

The median software developer in France (similar for e.g. Germany) makes around 50k. Software development as a path to personal wealth does not exist outside the US.

The proposition by these companies for new hires is: "Put in a lot of work, longer hours and solve difficult problems, so that (after taxes) you can earn 2k more a year than someone with a far more relaxed job".

The only people you are going to find are the very few ones intrinsically motivated to forego their social life in favor of your job offer and new graduates who don't know any better and will burn out and resent you and possibly the entire industry for the rest of their lives.

I will not put in exceptional effort, unless you are offering me something exceptional in return. And no, the "opportunity" to work hard on some problem does not count.

Basically all of Europe has this problem, companies are looking for people who put in exceptional work for average salaries. The problem isn't that these people don't exist or that you couldn't find them. The problem is that everyone who fits your role knows what kind of effort it involves and knows that their time will never be adequately rewarded.

Re: You can't git clone a team

#105

Post by e63f67dd-065b »

This kind of talent is ... not rare at all? And pretty easy to find too, just wander around the hallways at a major academic systems conference or hang around the kernel mailing lists, you'll meet most of the people working at the cutting edge of these things and get connected to those working on critical systems components. And yes, most of them work for FAANG or are funded by them.

Seriously, book a plane ticket to ATC/OSDI, EuroSys, etc and talk to all the people there. The good ones are already hired by one of the big established players (FAANG, Red Hat, Intel, etc), which is why you need to offer competitive compensation to lure them away.

Re: You can't git clone a team

#106

Post by frainfreeze »

IshKebab wrote: post #54
ziddoap wrote: post #24

Earlier quoted context omitted.

The question was "what is the best way to market yourself?" and your recommendation is to keep quiet and do less marketing. If that has worked for you, that's amazing! But it seems really counter-intuitive to me.

No he's saying if you are a generalist you shouldn't try to market yourself as that. Usually employers are looking for specialists that are perfectly moulded to the one exact task they need at that very moment. It's misguided of course, but that's what they think they want and if you say "I've done all sorts of things and I'm good at all of them" they'll hear "I don't have much experience with anything" and discount…

Exactly what I wanted to get across. That's why I have 3 personal pages/blogs and a separate forum. Only people very close to me know for all of them. Otherwise people get weird about you doing so many, so varied things. Here I m not talking even about work related relationships, on a personal basis people feel insulted and as if it is a competition if you list what you do when they ask you to. I have super focused, 16h days at the lab/office, of course I ll get a lot done... https://supermemo.guru/wiki/Planning_a_perfect_productive_da...

Re: You can't git clone a team

#107

Post by ajsnigrutin »

ipaddr wrote: post #102
johnnyanmac wrote: post #97

Earlier quoted context omitted.

It's not commodification, it's acknoledging that tech got exponentially more complex over the decades. just think of your favorite video game character in 2000 and then one in the 2020's and consider how much tech is needed to render, animate, light, and conceptualize it. in 2000 this was all done by maybe one artist and one gamedev, probably making a character with some hundreds of polys at best. now that artist has…

My point was unnecessarily they split the roles. An artist can cover concept and materials. A programmer can do gameplay, graphics, rendering and builds. In fact having people who understand the entire project makes for a better project. It's like moving from custom built cars to the assembly line where someone's job is putting in one screw. I understand it's cheaper/faster because you can hire anyone unskilled for c…

But we now expect a single person to design the engine, the bodywork, both aesthetically and technically, make an engine, actually make all those parts, assemble them together, pain the car and test it.

Jack of all trades, master of none. This is why we need clusters, "stacks" and "clouds" on the server side and gigabytes of ram on the client side + many megabytes transfered, just to show one simple weather forecast website that gives the user the same amount of information as a WAP site did on a five line mobile phone back in GPRS times.

Re: You can't git clone a team

#108

Post by cadamsdotcom »

johnnyanmac wrote: post #99
cadamsdotcom wrote: post #68

The value of knowledge is plummeting. It can, will, and should be subsumed by LLMs. Consider London’s black cabs. Drivers pass a gruelling exam to prove they know “everything” about one part of London. But, Google Maps puts that knowledge and more on every rideshare driver’s dashboard, for free, no advance study needed. Map knowledge doesn’t make you able to drive. Driving happens in real time - steering, obeying roa…

>It can, will, and should be subsumed by LLMs. okay. So what to do in the meantie while an LLM can no in fact help you much with kernel programming? People love talking about the future but people also have current business needs. > a motivated individual can use AI to make themselves employable as a full stack engineer faster now than at any point in history. I'd love to hear one case of that happening. I'll even ta…

Maybe have a chat to these folks in the article?

Seems they are plucking self-starters directly from universities. You could either use that funnel or build your own with the same idea.

Re: You can't git clone a team

#109

Post by johnnyanmac »

cadamsdotcom wrote: post #108
johnnyanmac wrote: post #99

Earlier quoted context omitted.

>It can, will, and should be subsumed by LLMs. okay. So what to do in the meantie while an LLM can no in fact help you much with kernel programming? People love talking about the future but people also have current business needs. > a motivated individual can use AI to make themselves employable as a full stack engineer faster now than at any point in history. I'd love to hear one case of that happening. I'll even ta…

Maybe have a chat to these folks in the article? Seems they are plucking self-starters directly from universities. You could either use that funnel or build your own with the same idea.

Sounds like a good plan. University recruitment is how you invest in society's future. I don't know how large this company is, so it's hard to determine if they can afford a training program.

Re: You can't git clone a team

#110

Post by jandrewrogers »

johnnyanmac wrote: post #101
jandrewrogers wrote: post #70

Earlier quoted context omitted.

One of the big issues with LLMs for e.g. systems software is that there are broad domains where training data effectively does not exist. Consequently, you can’t learn much from them of value. It is the blind leading the blind. The lack of training material for LLMs is of course a lack of training material for people too. Some areas of software have a long history of relying almost entirely on an oral tradition to pa…

>This has some advantages but it doesn’t scale and it makes it basically “dark knowledge” for LLMs or people without access to those that know it. We call this "tribal knowledge" in games. I despise it. It's done for a few reasons: - NDAs make public knowledge a landmine. And every game studio makes you sign NDAs. Even at the interview stage. - Churn. No one gets time to really develop expertise as they work on a pro…

I work in high-end data infrastructure, not games, but almost identical incentives and dynamics are at play. The state-of-the-art research isn’t coming out of academia for the most part. The R&D being done in private industry is slathered in NDAs that only slowly leaks out a decade or more after it was put into production. Many people don’t stay with it long enough to really master it.

There are some elegant and sophisticated techniques related to database kernels that have been passed around for a decade or two over beers that still don’t have a single reference in literature that I can find. The original researchers probably stayed quiet because it was under strict NDA but also likely retired years ago. No one writes it down because it sort of feels wrong to claim second-hand knowledge of unknown origin as your own, or to even lead people to assume as much. You show people, and they think it is amazing, but when they ask you for references or sources you have no idea where it came from.

There is a real gap here and it is getting worse.

Post reply on HN