Conversation

Replying to
"we have lapsed VR users, but there are no lapsed mobile phone users" most insightful point so far... he's made a lot of good "engineer's engineer" points but this is the big strategy point he's made... if you can't solve the lapsing problem, metaverses are done
1
30
This talk was very infinite-game. Carmack is clearly committed to the future of the technology itself, and dedicated to keeping the tech advancing and protecting it from mere business-interest decisions and strategies, while being pragmatic about it.
1
6
Re jailbreaking, interesting thing in the Carmack talk was his sidebar on allowing root access to Oculus Go now that it's basically been end-of-lifed, and his passing comments on openness.
1
4
All his technical points were very astute as expected. Especially liked the point about how local+cloud rendering would have to be application specific and is not really a general problem. Sounds true, and if so, it suggests initial grain of metaverse will be pretty vertical
1
4
He clearly has a strong bias against horizontally organized stacks, and the swipe at "architecture astronauts" is also a shot across the bows at horizontal efforts, and more generally, federated efforts in general (sdks, toolkits etc). Solid reasoning, but revealing too.
1
5
Though he clearly appreciates the low-fi janky stuff, his sympathies are clearly with the high-performance, hi-fi optimization end of tech evolution. Possibly a mix of engineering tastes and history as a game builder, where squeezing out more performance is key.
1
5
Kinda funny, but Carmack seems like the sort of guy who is capable of coming up with both the ideas for a primitive car and a faster horse, but if he happens to have both ideas at once, will choose to breed a faster horse over building a janky car.
2
10
My VR experiences so far have been nice enough, but yeah I'm definitely in the "lapsed user" category. The mix of affordances, frictions, and encumbrances does not add up to anything I want to truly dive into.
1
4
You don't get any points for disagreeing with anything in Zuck's keynote because that was almost pure PR rather than clear theses. But registering disagreement with Carmack is more serious, since he made nuanced and careful but conservative assertions.
He was too diplomatic to say anything was impossible, but he expressed skepticism or distaste for many things, so a good way to test your metaverse forceasting chops is to place your anti-Carmack bets on those things, since that's where, by Clarke's law, he's likely to be wrong.
1
2
I think he's wrong about the early VR version of the metaverse replacing many other things, including possibly one of your primary existing screens like a laptop or phone. That won't happen until we get to proper all day AR/MR, which by his own analysis is far too hard right now
1
1
He expressed distaste for a "high level" solution to stereo 3d (basically taking 2 povs in-universe and synthesizing) and talked about how he then did it the right way, using low-level system calls. Interesting to see what lessons he drew from that...
1
For him the lesson was there needed to be more low-level access and faster product management. I don't think that will happen. I suspect the hardware will instead get good enough that the crappy high-level solutions will proliferate. Moore's law driving "worse is better."
1
3
The larger point there is, the fraction of programmers that can solve problems in his idea of the "right way" is probably like <100 in the world right now. "Carmacking" hard problems is just not scalable.
1
8
Which means ways of solving basic metaverse problems that use *mediocre* programmers working with *superior* hardware is actually likely the least-effort path evolution will take.
1
10
It's funny because he recognized that in a different context, arguing against feature-rich expensive hardware, and arguing essentially for cheap, least-common denominator hardware that is minimally reliant on fast wifi (and therefore cloud processing and costs). But...
1
2
... you know what's even more expensive than high-end hardware? Carmack-grade programmers. Even though they are highly leveraged by orgs around them, it still requires amortizing.
1
4
I suspect things like this may unfold in the future of VR/metaverse. People at Carmack's level, but with different tastes/philosophies heading elsewhere to do things off the Carmack roadmap. Makes me wonder... who are the Carmack proteges worth keeping an eye on today? 🤔
3
4
Anyhow, of the major tech strands evolving today, this is probably the least personally interesting to me because of purely physiological reasons... my eyesight is getting worse (and more easily strained) faster than the tech is getting better. When it's good enough, I won't be.
1
5