Was ray tracing a scam?

 So Hardware unboxed came out with a video asking "was ray tracing a scam?" It's an interesting video I largely agree with but I wanna give my own thoughts. My opinion shouldn't be a surprise if you search my blog for previous articles on GPUs and ray tracing, but I do wanna reiterate some thoughts here.

I don't think it was out and out, full on a "scam." However, I would trace the end of the old GPU market as going back to RT becoming a thing. 

I remember when it came out I kinda did think it was scammy in a way. It felt very forced. Like, it was sold as this holy grail that was gonna change everything, and I was just thinking "who actually wants this?" I never heard of  ray tracing before this. The tech was always like something that was oversold, like "it was the future." Kinda like how AI is now "the future." And much like AI, I feel like it was pushed WAAAAY before its time. Like, as steve points out in the above video, the performance wasn't there, and many gamers still dont use it today, me being one of them. And I'll be blunt as to why. The performance hit was too much. 

It's not the 2000s any more. Back in the day, it was acceptable to play games on an old CRT monitor at 1024x768, 800x600, or 640x480, and run the game at 30 FPS. Hell, if you go back to the 90s, you had N64 games with framerates dropping down in the teens and running at like 320x240. And it sucked. And no one wants to play games like that any more. I've watched gaming evolve over time, and as gen 8 became prevalent, things kinda caught up to PC where now we were going up to 1080p/60 FPS. And that's where I'm comfortable playing games. And some wanna do even more. I like to keep my FPS consistently above 60 to keep it stable, often running minimum FPS around 70-80 with averages around 90-100 these days if i can make it work. Of course, with issues...stemming from the introduction of RTX cards, that's becoming difficult these days. 

RT requires TONS of horse power. And unimaginable amount of horsepower. And it's not a tool that's been better for the end consumer. There was an LTT video a while back asking people to pick between the RT and non RT lighted versions of a game and most couldnt tell the difference. What RT DID do was make things easier for developers where rather than implement their own lighting, they could just, BOOM, DONE. They hit the RT button and the RT simulates REAL light instead. It's kinda like me being able to BOOM, hit a button and simulate election outcomes in a way that I used to have to do manually, which would save hours of time. 

But that automation has also led to laziness among devs. It's led to higher system requirements, less focus on optimization. I know Asmongold had a video a while back that I think I discussed on here that mentioned how developers wanna spend as little time optimizing as possible. So that just leads to bloated messes of games. 

And then DLSS and other upscaling acts as a crutch. It's no mistake those technologies were sold along side ray tracing. To make ray tracing usable, they had to reduce the image quality resolution wise to reduce the load on the GPU, functionally regressing us back to the late 2000s in acceptable resolutions, which...look like dog crap on modern monitors. And to compensate, AI upscaling. Which did help extend the life of the old 1060 to some degree. I mean, its better to run FSR than it is to run at a lower res natively. But CLEARLY, neither are preferable. And devs have since focused on pushing graphics at an insane rate, and then expecting upscaling tech to compensate for their lack of optimization. Oh, a game only runs at 720p/30 FPS on a fricking 6650 XT or 3060? Just upscale it bro, what do you want me to do, actually make it playable for you LIKE WE ALWAYS USED TO BEFORE THESE TECHNOLOGIES CAME OUT?! Uh...yeah? yeah. We want games to run well on the hardware available. Again, not against progress, but against progress at all costs.

Like AI, Ray tracing seemed to come way too early, and is pushed way too hard. Its demands are just too high for the hardware available. Even then, ever since ray tracing came out, that's when things started going wrong for the graphics card market. We were told, oh, we cant have sub $300 60 cards any more, RT is too expensive! We gotta put all this extra stuff on the cards and blah blah blah. And that's why the low end GPU market functionally died. Because with the introduction of this crap, it no longer became possible to produce low budget GPUs that were capable of running games. Not only did they not have the horse power, but they also lacked the extra RT/AI components to make that stuff work. The LOWEST END RTX card was $350 at launch. And even now, a 2070 type performance only went from around $500 to around $250 before rampocalypse, and $300 with rampocalypse. And that can BARELY run RT games even now. 

I mean, we could've made GPUs just have more stream processors. It used to be that GPUs advanced in such a way where you would double performance in 3 years at the same price point. But if you look at where the market is in 2026, we're literally only at 2x 2016 performance per dollar. And most of the 2020s has been a massive stagnation. It's like all of these investments are going into RT and all this nonsense instead of just giving us more stream processors and cuda cores. So now we're adding ridiculous demands to games with ray tracing, and then compensating for it with all of these technologies, while the cost of GPUs just keeps going up. People act like this is just the way things are now, that we cant grow at the rate we used to, but MAYBE IF WE JUST PUT MORE CUDA CORES ON A CARD INSTEAD OF THIS EXTRA BS, MAYBE WE'D SEE MORE PROGRESS! 

And you know what? Maybe that progress would scale down too. I think one of the reasons the lower end market died is because they STILL cant even achieve a 2060 baseline of performance with RT. And again, that's just a terrible experience for ray tracing. 

This stuff only really helps the affluent. Those with large budgets willing to dump tons of money into GPUs. But if youre a normal gamer trying to game on like a 3060 tier card (about average these days), it's completely useless. Who the hell uses ray tracing at our price range? If anything, just like steve showed, a lot of gamers are more interested in LOWERING quality settings to run games at higher frame rates. And also, to simplify the graphics to make things more visible, like in the fortnite part of the vid. And I can attest to this. Shadows tend to KILL frame rates, and they tend to make it harder to see enemies. When I optimize games to run on my PC, I turn DOF nonsense off, film grain, all that crap, and then i go for shadows if i need higher frames. And it's just useful to play with lower graphics on competitive games.

So most of us want the industry to go in the direction of simplified visuals with higher frame rates, and these billionaires with a god complex who control the industry are trying to force us to go the other way. And again, there is a market for that. Since then gaming has become increasingly an upper class thing, and more and more out of reach of the masses. It's led to higher costs, poor performance, poor visuals, and people like me feeling like things were just better 10 years ago. 

If I could hit a button that would eliminate ray tracing from the timeline, and instead go in the direction of improving raster, I bet we would have better, cheaper GPUs with cleaner visuals in games. Not necessarily gonna say higher FPS because i think some developers are always gonna take whatever level of hardware is available and then turn around and say "30 fps is good enough", but I think that it would lessen it.

Much like with AI, I feel like these billionaires just tried to replace a lot of old ways of doing things with new workarounds that just arent as good in some ways. Like they are better in some ways, but in other ways, they're just not. And I feel like the gaming industry has just gone in the wrong direction in recent years because of all of this stuff. 

I wont say RT is a full on scam, but I think its impact on the market was always overstated. This is what happens when you leave the entire industry in the hands of a handful of these tech billionaires with a god complex trying to sell their weird visions for what life should be. And then because they hold disproportionate control over the entire industry, they just push it in that direction. And it just isn't all it was cracked up to be. 

Like with AI, maybe tech will get there....in a few decades. But not now, not with current tech. They pushed this way too early, thinking gamers would just accept playing games at some low resolution at 30 FPS again when in reality, no, we like the higher frame rates and higher resolutions more. It's like rather than just give us more raw processing power, they keep trying to sell us technologies instead that are one step forward and 2 steps back. DLSS, frame generation, ray tracing. And then when we tell these guys what we want they look at us like we're luddites. You can think that about me here, but I really do like to see myself as pro technology. But we have to actually think about whether the technology makes sense. 

Like, to go back to election models, yeah, I did GREATLY improve my election forecast abilities by switching to google sheets. And yet....I abandoned ever increasing complexity to my models, and trying to build a version of my simulator that would spit out hundreds, if not thousands of simulations at once. 

Why? Because it didn't make sense. It didn't actually make things better. More complex models have more that can go wrong with them, and I very quickly could tell what went wrong when I started using them. And Im still critical of these big names like 538 who create models so complex that they end up predicting nothing because they're so mired in variables that you can never tell what's going on because things can change. Likewise, I didn't go with the mass simulation model because the thing fricking broke whenever I had to change the data. I have my own technical debt and spaghetti code in my models because quite frankly, I'm not that good with excel. If I take that model and then ramp it up 1000x, the thing bogs down my entire computer when I use it. And if i change the variables or the data, even for routine changes, the whole thing just...breaks down and starts spitting out errors. 

If I were one of those tech bros, I'd basically be like THIS IS THE FUTURE AND I WILL FORCE IT TO WORK ANY WAY BECAUSE IM A SUPER SPECIAL VISIONARY OUT OF TOUCH WITH REALITY. But being the more reasonable human centered capitalist who tends to operate closer to bill gates "lazy person" approach to things, beyond some point I'm like "this is more work than it's worth, this is close enough."

Ya know? Technology is a TOOL. It's supposed to be useful. If it actually CREATES work and makes things harder and creates massive externalities, or arguably saves only a tiny bit of work at the expense of those externalities, maybe the tool isn't worth using. This doesn't even mean scrap it. I mean, i STILL mess around and experiment with my election models trying to find ways to make them better. I just find that when I do my efforts don't go anywhere, and I just revert to the older, simpler model. Simple works. Complex breaks down and has problems. 

The same can be said with computer graphics. Graphics are a good thing. Making games better is a good thing. BUT, if it comes at the expense of higher costs or the need for weird workarounds like DLSS or frame generation JUST TO BE BORDERLINE VIABLE, it's like...NO, I DON'T WANT THIS! And that's why ray tracing is in the state that it's in. it was released the second it was just barely viable, it barely functions, and the industry hasnt evolved it enough to make it worth using for most people. In the long long term, yeah, maybe it's the future. But we're talking, what, 20-30 years from now? By then we'll be measuring processors in picometers instead of nanometers (if we dont just shift to a different standard altogether) and we'll have tens to hundreds of times more processing power in theory. There's no need to force this crap before it's ready. Which is what these guys are doing. Again.