@Lensflare

As @D-4got10-01 said, it's the game optimized version of the Nvidia drivers.

I was part of the spir-V team, optimization for Vulkan mainly.

In my current company, they are offered on company time, but I've seen both cases.

In my opinion, if it's due to client necessities they should be on company time. If it's by your request and the company can gain from it, they may offer to pay for it or do in company time but not both.

If they have no stake in it, you should be on your own.

Collate. That's your second statement. The first, I can relate. Getting ready to throw fists until it's just you missing some simple-ass config

In the end, during my last job hunt, I ended just uploading old stuff from my previous company, everything either discontinued or shelved.

They have me permission because I asked but I don't think they would have found out or cared either way.

Also, at least in Spain, as long as you don't cause profit damages by publishing, you have the right to, as intellectual owner, even if you have to enforce a closed license.

Whatever, "idiot".

At least I don't have to worry about my rent.

@antigermanist

Well, I don't seem like the one having trouble in the market because they can't read the trends, nor the one wanting to build walls, so, by all means, keep hating, not like I really care XD.

@BordedDev

That's another whole can of worms, because it turns apparently continuous math into actual discrete math. And that's another whole puzzle to solve while still having wiggling room to make regular enemies non-trivial while keeping upgrades feel impactful against entities like bosses with disproportionally huge HP pools.

Well, my experience with clients is that you can tell them you changed the colour (when you didn't) and they will happily swallow it.

All so-called web developers should be what is wrongly called "full stack".

Pure frontend is just glorified graphic designers who can understand the simplest CSS or XML grammar.

Pure backend should know at least what's the difference between block and inline elements.

AI will replace them? For 80% of web work, likely.

Some will survive, the rest need to adapt fast.

You need to either implement level scaling, or rely on non mathematical mechanics like expedition 33.

Any monotonously increasing function (such as level based rpg stats) will always outpace a static value, and thus the problem.

WoW does it kinda right, in the sense that you can party with a lower level friend (or choose to scale yourself, such as in timewalking) so things stay "challenging" (nothing in wow is challenging if you have two brain cells).

Super bosses that do not rely on unique mechanics require the game allowing you to abuse mechanics (again like expedition 33).

Or having an endless difficulty scale like Diablo rifts, but that feels a little artificial since it's unnaturally boosted by ridiculous bonuses in sets.

PoE does it better, but it doesn't have endlessly scaling difficulty, and if you pay enough, you can trivialize even super bosses.

@12bitfloat

Can't really have any semi functional scene graph without composing transforms anyway...

@3dgoosee

You can not reliably store rotations with just three values (I'm assuming you are using Euler angles) because they suffer from gimbal lock, essentially losing one degree of freedom.

Rotations should be represented by unit quaternions, which don't suffer gimbal locks, and are just as easy to compose.

@CaptainRant

If you quit your job and want to start in a new stack from scratch, you go back to being a junior, and that's it. Your previous knowledge should allow you to climb relatively fast, but you are still a junior in the stack.

I'll watch sin city again.

I'll keep Noirvember to not dye anything red or yellow :)

The problem with lambdas is that they require competent DevOps. And that's hard to find nowadays.

When done right (repos integrated with CDK and SAM), good cloud watch management and correct test events, they are basically no different from working on pretty much any other repo and running shit locally.

And yeah, snap scalability is their main use case.

No matter if you have a load balancer with autoscaling, unless you can predict spikes in traffic, by having more baseline nodes (and paying for them when idle), by the time more are spun up you already crashed that spike. Lambdas have cold boot time of around 1.5 seconds, and you pay for exactly the time they run,making them great for those situations.

So, long story short. Great for their intended use case. Terrible when abused for the buzzword like with so many other aws products.

(And it doesn't help that aws "expert-"... Salesmen lure you into overcomplicated setups for really simple use cases)

@Wisecrack

Problem arises basically when you think the scheduler is deterministic.

(And by all means it is, internally), but that's not part of the "API" of a CPU.

So, as a user app, you cannot have any guarantees to the behaviour of the scheduler, and thus, it's not deterministic.

The idea had merit, but it glossed over all the technicalities of the superscalar computers of today.

Not gonna repeat what knowledgeable people already said, just that from theory to practice there's a leap.

And "lock-free" doesn't automatically mean "better".

Multithreading is like the Hydra. Cut one problem and two will spawn. Stay true to the building blocks people smarter than any of us have built.

Never big on writing music since I have too much of a kartoffenkopf for it, but playing, yes. Every day.

Well, at least in Europe, your identity is your national id document. Whatever is there, flies.

You can't have two, well, because you can't. You can modify your current one, sure, but that's it.

At least in Spain, even if you claim theft or loss, they'll check your fingerprints again and find out.

@TeachMeCode

Base 64 uses uppercase and lowercase English alphabet, digits, + and / XD

Like, seriously, converting an int to base 32 (dunno why but ok). Just fucking take modulo 32^n repeatedly (and figure out how you wanna represent 32 values).

Also asking how to solve CORS errors in the frontend.

Or 3D programmers asking how to calculate the angle between two vectors... (They failed to even know what a vector is, they kept repeating the unity class names).

Why do I know? Because they asked me! I told them like "take the vector from your object origin to your target origin, (a simple substraction), normalize, and acos() the coordinate of the angle you want."

They were utterly baffled. I'd think someone working on 3D experiences in Unity would have at least very basic algebra knowledge....

Guess not...

@Lensflare Fovel, not stretched... zen ve can Talk.

I don't know why people would say direct 3d is verbose or horrible.

I mean, their main competitor literally copied the style when designing Vulkan so...

And while you can kinda get by with open gl and Vulkan nowadays, of you wanna do real shit on windows, it's directx or the highway.

Well, appendicectomy is usually done by laparoscopy nowadays, so you do get a small scar in your left side as well.

@hjk101

It's basically a glorified linter, but it has context from the internal issue tracker and chat so it's trained on what trips other devs in the team and kinda checks for that too. Not really more annoying than your usual code review, except it's a machine, doesn't have feelings, doesn't hold grudges, and can trust it's not out to get you and annoy you no matter what.

@12bitfloat
@Lensflare

I said this before. Will say again. We all can read code. We can not read intent. Comment intent, not just blindly repeating what code evidently does.

Sure, if you use some obscure technique like quake fast sqrt by all means comment.

In the end, it's all about being able to put yourself in the shoes of anyone new to your project.

In nVidia, it's enforced that anything that is pushed must build, and our AI assistant can actually tell (because it analyzes frequency of use and such) whether what you implemented is easy enough to maintain.

It only flags when it doesn't understand what we try to do, and forces us to comment (again, intent) the AI has a compiler, and therefore knows better than us what code is trivial and what is doing esoteric shit

@glowFX

Not that I want to defend php, but the situation you described is exactly what php-fpm was made for.

Rust vs C++ is like Ubuntu vs debian.

They are the fucking same. Just different philosophies.

Ubuntu is essentially debian unstable. Debian based, but not wanting to conform to the rigidity of debian release cycles.

Rust is one (of many) attempts to re-do C++ without adhering to it's two core principles... Keep already written code working (C++ is oh so much held back by this) and having zero cost abstractions.

Rust breaks the first by existing. It's new, doesn't have the baggage. The second one it breaks by forcing you to adopt what they say is secure.

Sure, there are a lot of things you, as a developer, can mess up, but it shouldn't be enforced by the language. Because in doing so, it's no longer zero cost. Bounds checking, for example, *does* have a runtime cost. (As do many other "security" measures.)

Contrary to popular belief, C++ *does* have these mechanisms. They are called debug layers, and, following the core principle, are opt-in. Don't pay for what you don't use.

Know JavaScript since it was the thing it should never have tried to outgrow. A DOM manipulating language.

I started dealing with node at around 2014 or so.

Of course, it helped that I know what asynchrony is from way before Promise was a thing, and that I'm way too familiar with V8 and libuv.

I dare say I know all the inner black magics of JavaScript and typescript, and have used them to pull crazy stuff.

Yet still the decisions the ecmascript committee makes will forever keep js as a clown language.

So, all in all, considering that previous to node there wasn't any depth to really explore, I'd say 4 years or so.

Dead element detection is simple enough. Log your tasks. Create a meta object when stuff is supposed to happen.

Upon completion, delete original video and meta object.

Your cleaner can periodically check for videos without meta object and flag then or delete them outright

Why do you control shit from the *client*?

S3 has hooks for successful uploads. Don't rely on client code. (A malicious user could probably replay your "process video" message and DoS your app)

If you are worried about the webhook, s3 uploads can trigger lambdas. Use those.

Also, in your stated case you can use redis TTL and key space notifications to do what you want without relying on polling.