I still think that OpenGL is a better place to start learning 3D because there is so much less bookkeeping to do regarding memory and concurrency vs. Vulkan. I'm working on a "Post-Modern OpenGL" tutorial that exclusively teaches the most "modern" (merely a decade old) APIs and practices of GL 4.6. But, writing is slow going...
If you are going to use Vulkan, check out https://howtovulkan.com/ Vulkan started with a lot of compromises to make mobile hardware happy at the expense of making everything overly complicated on desktop. Over the past decade, desktop devs have managed to get a lot of features added to make Vulkan on desktop more sane. How To Vulkan covers that newer approach. This recent video "It's Not About the API" https://www.youtube.com/watch?v=7bSzp-QildA shows how simple it can be if you let it.
And, if you are on a Mac, folks who use Metal like it a lot. Don't worry about lock-in. Once you learn the basics, knowledge is easily transferable to DX12 and Vulkan. You should plan to write 3 or 4 renderers to throw away anyway :P
> I'm working on a "Post-Modern OpenGL" tutorial that exclusively teaches the most "modern" (merely a decade old) APIs and practices of GL 4.6. But, writing is slow going...
Is it the AZDO stuff? Quite interested, since there isn't really much information on the Internet about it rather than some slides and GDC videos.
Yep. It's my thesis that early GL APIs feel beginner-friendly because they are so hand-holdy, one-step-at-a-time. But, they end up more complicated when you get serious and they instill bad practices.
Instead, the "modern" APIs allow you to leverage your pre-existing knowledge of "allocate arrays of structs and start indexing them." That's not trivial. But, it is already familiar. And, it ends up a lot better in the end compared to "Invoke whole lot of functions to manipulate a hidden state machine."
I'm the opposite. Between work and family, I don't get a lot of time to hole up in a cave with an IDE for fun. Prior to LLMs, it was a struggle to get much of anything done. So, I rarely tried. But, now I can get ideas fleshed out and even get code down in the small breaks I get to myself. Now I'm making progress where I wasn't before.
Sometimes I'm not interested in the inner workings of a technology, I just want a one-off tool. Ex: No interest in web tech. But, vibe-coded a tool to scrape and cross-reference a couple of online reports. Now I have the info I need for other projects.
Sometimes I know exactly what I want and it's quicker to hand-hold the AI through making it for me. Ex: Put together yet another SIMD math lib recently. SIMD intrisics are an obnoxious API. But, a non-SIMD implementation is easy to verify and many SIMD implementations are easy to validate vs. the non-SIMD reference. It's just a huge amount of obtuse code.
Sometimes I don't know exactly what I want and it's great to have a always-online partner to bounce questions off of, do rapid-prototyping for me, do code reviews for me. AIs aren't perfect. But, being instant, patient and often right makes them a huge improvement over online forums.
I feel the divide is between people who built something for the joy of building vs people who built something for the joy of consuming (like you). For you LLM’s are great, you can spend less time, effort and still enjoy all the stuff you wanted to enjoy.
For someone who actually liked the process of building, LLMs don’t do anything except at worst muddy the market with AI generated slop that needs to be compared against, which will eventually be figured out I guess.
For the joy of building, I'd say LLMs are still great for questions, rapid-prototyping and code reviews. And, beyond that: Who cares what others do? Build what brings you joy.
As @petecordell put it: "Telling a programmer there's already a library to do X is like telling a songwriter there's already a song about love."
Lots of people start out with a non-programmable config format. But, as their situation becomes more complicated, they end up shoehorning in programmable-ish features until they realize they are running straight into https://en.wikipedia.org/wiki/Greenspun%27s_tenth_rule and decide to do it properly.
I mean, fair enough I guess. I've accidentally more-or-less reinvented interpreters on two separate occasions (both times using JSON as a quasi-AST), so I understand it.
I guess I just feel like if you feel like you need programming constructs in your configuration, its approaching it from the wrong angle, and it should instead be handled via plugins.
I feel like configuration itself should be a "dumb" format like JSON, and when you want to do anything that can't be accomplished by plain JSON, you should allow for easy development of plugins.
Classic SSAO is very similar to performing an "unsharp mask" based on the depth buffer. This has been proposed "a simple and efficient method to enhance the perceptual quality of images that contain depth information" in the paper "Image Enhancement by Unsharp Masking the Depth Buffer" which came out a year after SSAO was published by Crytek.
Yes, game development is one of the use cases I had in mind. Since Clx is fully ahead-of-time compiled, it avoids the JIT restrictions present on iOS and relies on the platform's normal native toolchain.
But Clx currently only supports Linux, Windows and macOS, including Apple Silicon. I haven't tested it on iOS yet, so I can't claim iOS support at this stage.
This is a very important paragraph from quern's article
> There are few to no side-effects to melatonin use in adults (there is uncertainty about the risks & benefits in children & adolescents28), and it is not addictive or habit-forming like caffeine is. The usual dose for a night is 0.5-3 mg and I take 1.5mg [29]; my dose is highly likely to be too high. High doses may well be responsible for why some people try melatonin and report that it does nothing or hurts them, since in one study, the best dose for old people was 10x smaller (0.1mg or 0.3mg) and for one blind person, 0.5mg [30 31 32] . Zhdanova et al 1996 found 0.3mg & 1.0mg to affect sleep onset similarly. A study of delayed-release melatonin found with their high dose of 4mg (but not 0.4mg) elevated melatonin levels 10 hours after bedtime ( Gooneratne et al 2011) - potentially interfering with waking time.
It is difficult to find doses as small as 1mg sublingual (dissolve under the tongue) in my area. Everyone is trying to sell you 5-30mg chewables. And, I expect everyone is buying them under the assumption that more is better. But, here it explicitly is not.
The way melatonin works well for me is to wait until I'm already settled into bed and should be asleep, but I'm not. Do a couple body scans to relax for real. Try to think about something mindless. Then if I'm still awake, pop 1mg under my tongue. I'll usually wake in the morning with half of it drooled on my pillow :P
From your description it may be acting as a placebo -- the act of putting the pill into your mouth is what tells your subconscious to activate a sleep cycle.
I've had the same problem finding low-dose melatonin in brick-and-mortar stores, but Natrol sells a 1mg dissolving tablet on Amazon. If I need a sleep aid, I break it in half for a ~0.5mg dose, chew it with my front teeth, then hold it under my tongue for a bit before swallowing what's left.
It usually takes 15-30 minutes before I get noticeably drowsy, but I feel no more groggy the next day than if I slept sober.
For low dose melatonin, try liquid melatonin. Around here Sprouts carries it, but you gotta look a bit to find it. The bottle says 30 drop equal 3mg. So you can take small doses easily, and it absorbs under your tongue pretty quickly.
Yes. I agree. Liquid melatonin is the way to go for micro dosing. I also have good experiences with micro doses of 0.1-0.2 mg when waking up too early and having difficulties of falling asleep again. Taking a full 1 mg leaves after 3 or 4 am you very croggy in the morning.
Natrol makes a 1mg gummy and Vitafusion has a peach flavored 1.5mg gummy (sold as “3mg” but the dose is “two gummies.” My kids use these. I get them on Amazon.
Way too high in my experience. Can leave you croggy and dependent. Micro dosing using drops is more effective for me without feeling groggy in the morning. Anything above 1mg is way above the natural melatonin level of our body.
Interesting. Can you share any research tips that helped you arrive at that conclusion? I don't want to overdo it. Dosages in mainstream products are often even higher than the 1.5 so it's obviously easy to go way, way too high if what you're saying is true.
It's a bit harder to find, but I've had a good experience with liquid melatonin. Each drop contains like 0.1mg, so it's easy to dial in the small dose that you need.
I’m not a scholar of the fall of SGI. But, I’m sure it has been documented in detail.
AFAICT, SGI was a textbook Innovator’s Dilemma case with an expensive enterprise product that’s hard to give up in the face of cheap, low-margin competition.
This is true. I was at SGI, and their entire business was optimized to serving the needs of very sophisticated customers who were themselves pushing the envelope. Absolutely great customers to work with. But SGI’s DNA couldn’t adjust to the low margin high volume consumer space.
They built an incredible Windows NT system (for the time) but couldn’t keep up with the 6 month release cycle their competitors were on.
SGI was an incredible place to work while it lasted.
Spot on. They had the tech advantages but the high margins of full work stations blinded them to the changing winds in the industry.
I remember at the time seeing some folks blown away that they could do SGI like stuff on a PC with a $199 add on card. It wasn't identical but it was close enough and you didn't have to switch to out of the Windows ecosystem. That kind of scaling and software inertia is just too hard to compete against.
They were offered one mass-market opportunity on a silver platter, which they took: When Nintendo asked them to design the N64 GPU. It didn't seem to be very profitable for them.
It's very unclear in that era that there is a big market for 3D graphics at home. So their big customers would buy the cheap cards but in low volumes -> bankruptcy. And maybe there's either no big consumer market, or it grows too slowly to replace the loss of their main business.
reply