The product (marketing) decision to focus on AR passthrough does seem really strange. If I was going to use a VR device for work, then, maybe if the resolution was high enough it could be nice to have much more space, but, I'd really only care that I'm not bumping into things. I'm not exactly sure why this AR route is being pushed before it's ready -- living in the world through cameras seems like a much harder problem to be solved after getting the base VR case to work.
IMO the more interesting business cases would be around simulated environments even as simple as viewing 360 videos for places where workers would be expected to know what's going on when they enter.
I had a thought that this is even beyond the word psychedelia and extends more broadly to all modern health. But then, I think the word is just right, since psychedelia has as much to do with LSD as vitamins have to do with vitamin water. If you disagree with the analogy that’s no issue, it’s simply A Catchy Hook.
Psychedelic simply means expansion of the mind and is independent of any particular practice or substance.
Mindfulness seems to be another, or even, “the,” new culmination word for “safe” psychedelia in the west (as contrast with the “unsafe” roll-your-own a.la burning man). We can tell it’s the current culmination word because it’s polarized, the one word with a significant amount of scientific research, but then at the same time oddly clinical and paradoxically overall too LuLu lemon & vitamin water. It just doesn’t seem to have that same oomph as eating acid and stealing fire in the desert, but it’s receiving medical approvals that pave the way for economics to benefit people at scale. Mindfulness seems to encourage passivity, but then it’s practiced by navy seals. At the core, it’s all wrong by placing too much focus on the mind, just like we’ve always done in Latin-based languages.
Whatever the word is, developing a perspective that we individually believe is worth cultivating is the root of a connection between well-being and performance.
It doesn’t sound like a SASS app because it isn’t. It doesn’t seem like a good business idea because business is a dirty word, like mindfulness, people think there’s a way to do it.
Two quotes stood out to me:
“But what we really need are psychedelic models for business - business that defines new standards for integrity, equity and ethics; business reimagined with a technicolor glow.”
“What would I have done if I had known that this would happen?”
The second seems like exactly the right question. If you have an answer I’d love to chat, even just for fun.
Actually now I’m confused because I don’t understand Latin. Psyche definitely gets at the mind idea, but also soul or spirit. Ick sounds like what I do before I vomit but I guess “delic” is really what we’d be concerned about, but then this seems like it’s just talking about an experience of the first thing, an arrow pointing back to an amorphous blob of language. I’m not sure if it’s better or worse that I brought my own preconceptions to the word at this point, or how my preconceptions are different from the thing itself. I’m seeing other interpretations talking about dissolution. It’s certainly something big happening, that’s for sure! In a way, I’m also confused about mindfulness. Is my mind full or is the fullness of mind different from that? Does it mean something that there’s only one “L?”
Absolutely I think the next interesting discussions could either be about the word, or what to do.
Again, thanks to the author for properly choosing A New Brand, with existing baggage which perfectly typifies where discussions like this tend to go, see comments below.
Wow this is actually scratching an itch of mine now. I love the kindle feature of saving / sharing quotes and have been wanting something for offline. Awesome work Tanner!
I don’t think this is a humblebrag. It’s a success story, and we should celebrate those. There’s nothing wrong with being excited about something turning into more of a success than expected, and there are a lot of general values in the post that may be broadly encouraging to folks here.
Interestingly, in his book outlining the technique he teaches, Andy doesn’t directly recommend people do any sort of guided meditation. Rather, he talks about it like a sort of skill people can learn.
There are probably diminishing returns once you are able to focus through the technique, however, even in monasteries with experienced meditators simple cues like bells are used because everyone gets distracted once in a while!
I thought this was strange too. I was just watching Gary open a few $500 football card boxes today (5 cards in each). Upon opening a box without any rare players he said: "That's what you get sometimes gambling!" So I think people know that's what's going on.
Addicts generally know at some level that they are addicted. The core problem is more that addicts can't break the compulsive behavior without a significant intervention.
Which is one of the reasons why these digital addictions can be such a big problem. Someone can't repeatedly open one pack of baseball cards after another endlessly. A limited number of packs exist, it requires time and energy to acquire them, and you will often have to interact with a human to get them. Those limitations serve as minor interventions. They act as a throttle on the addiction that simply doesn't exist when opening a never ending supply of loot boxes in the privacy of your own home.
In my experience as a recovering behavior addict, the difficulty partially lies in a lot of cultural concepts around addiction, recovery, and learning that run counter to what science tells us about learning. "Instantaneous recovery" is described as a random chance event, as opposed to possibly the outcome of someone who's got a better grip on how to grow because they were missing a single piece of wisdom that, when acquired or acted on, led to cascading network effects in the brain.
I get hackernews think their conscious brains are in total control of their destiny, and I do think the nerdier and smarter you are the more that can be true, but psychology basically has proven this to be a self-delusion.
Your subconscious and instinctual drives have ludicrous control over your conscious thought. Anyone who knows an intelligent drug addict knows the lengths of rationalization that a conscious brain will wrap itself into in order to justify a fix.
> Anyone who knows an intelligent drug addict knows the lengths of rationalization that a conscious brain will wrap itself into in order to justify a fix.
Talking to a clever, articulate addict can be downright scary Their ability to rationalize isn't just limited to fooling themselves and downplaying the problem to others. It's quite a surprise to hear someone justify their habit so persuasively that you have to go back and reevaluate whether they're right.
(Of course, another part of this is that a really self-aware addict is especially likely to see the motivators of their behavior. If they're consciously self-medicating, even harmfully, "just quit" is a crappy answer.)
I would love to use Alpha Zero. The article makes it sound like Komodo (which uses similar techniques to Alpha Zero) could also beat Stockfish (since Alpha Zero beat Stockfish). Stockfish still beat Komodo and all other engines according to the computer chess championship which uses equal-and-limited compute requirements.
You can run LeelaZero, which is weaker than AlphaZero but neck and neck with Stockfish. It lost 50.5-49.5 in the recent TCEC championship [0], which sounds like it is what you are referring to. (It historically has used equal hardware, but that has been complicated by the emergence of GPU engines).
What I'd like is a shim to act as a UCI engine but actually relay moves to/from LeelaZero on a remote machine. Chessbase offer something like this, but as a paid service on a proprietary protocol.
Yea the TCEC is one. I also think the chess.com computer chess championship (CCC) is interesting (Stockfish is winning here as well). The CCC uses more advanced hardware. Both are live now and interesting to watch, I mentioned the link for CCC but the TCEC is here: https://tcec.chessdom.com/
A server-based championship would also be interesting but pretty degenerate. I guess that's why computers playing more complex real-time games is becoming the bleeding edge.
This criticism refers to DeepMind's late-2017 results however. In late 2018 DeepMind have published a follow-up paper[1] in Science where they claim to have beaten Stockfish 8 under non-cherrypicked conditions, specifically the conditions under which the latter won TCEC 9 in 2016.
AFAIK there are no AlphaZero results against either the current stable SF release (Stockfish 10) or the SF development version which narrowly beat Leela at TCEC 14.
I would still love to see the Stockfish team go up against the AlphaZero team, with the Stockfish team being allowed to modify their code however they need to in order to get the performance they need for the event. I think they should agree on some hardware equivalences upfront since they dont have the same requirements, and I think that the Stockfish team should have an API to play the AlphaZero algo in preparation, just the same way that AlphaZero already has unfettered access to StockFish.
I'd relish to read about whoever wins that: a fair competition.
Honestly, Leela Chess Zero was built on top of the AlphaZero algorithms, and I'd argue that LC0 is more relevant and interesting today than what happened two years ago.
Two open source implementations with active development is equal footing. The importance is the algorithm and how it improves, not on a specific implementation that has been abandoned. For those, we have TCEC and CCCC continually measuring those and a fair number of other implementations.
Stockfish and Leela, after 100 games, were virtually tied - Stockfish had one extra win. Leela has passed Houdini and Komodo at this point. At this point, it becomes the question if Stockfish can improve at the same speed as Leela.
You might be missing the point. AlphaZero is a research project, not a commercial project.
The most interesting result is AlphaZero's playstyle. It's a lot less robotic and something that humans can actually learn from compared to Stockfish's.
Its not code that's the big differentiator here. Its hardware.
Neural Networks are an embarasingly parallel problem that can be solved with accelerated matrix-multiplication hardware. The entire network can be easily represented as just matrix-multiplication problems.
Stockfish is still written in a classical fashion, and has issues scaling above and beyond 64 cores. But Neural Networks can keep getting bigger, and bigger, taking advantage of extremely parallel hardware like GPUs (10,000+ shaders per GPU) or Google's Tensor Processing Units.
In some ways, the fight will never be fair. GPGPUs and TPUs can simply throw far more math, using far less power, at the problem.
Furthermore, games of perfect information, like Chess and Go, have always been played well using neural nets.
In effect, Neural Nets are being used as a tool to allow most computation to occur on far more parallel and efficient hardware.
-----------
I think the next big stage for Chess AI / Go AIs is for someone to figure out how to take advantage of SIMD-compute, to achieve similar scalability. As it is right now, both MCTS and Alpha-Beta pruning can only really be done on a CPU, so you can't leverage the huge computational power available today.
I think Chess would be the easier game: Bitboards are represented as 64-bit integers and would easily map to just 2x32-bit registers of a GPU shader. But I get stuck whenever I think of an algorithm that would have low-divergence.
Perhaps Go would be easier due to the similarity of pieces?
I dunno, I've got way too many personal projects I wanna work on. So this isn't really a thing I can do. But its definitely an interesting research problem. My best guess is to convert the whole Chess-playing AI into a SAT-solver and then write a GPU-accelerated SAT-solver (because SAT is uniform in theory, so I would expect a GPGPU-based search of the 3-SAT space to be low-divergence)
I'm just shooting from the hip here, no serious suggestions really. But that's my quickie 2-minute thought process on this particular problem. You don't necessarily have to make the whole thing uniform, modern GPUs are either 32-warp size (NVidia) or 64-wavefront size (AMD). So you just need to figure out how to consistently get batches of 32 or 64 to execute the same code without diverging across if-statements or loops...
Heh, easier said than done.
EDIT: Perhaps organizing the code so that "All Bishop Moves" are evaluated in a batch. Lets say 1-million chess boards are to be evaluated on a GPU, how would it be done? Perhaps, first, analyze all Bishop-moves. Then analyze all Rook Moves. Then analyze all Queen moves. The goal of Stockfish was to "process the most boards" at any given time. Changing the architecture for bandwidth-oriented GPGPU compute would be very much in the same spirit as Stockfish's original goals.
IMO the more interesting business cases would be around simulated environments even as simple as viewing 360 videos for places where workers would be expected to know what's going on when they enter.
Thanks for the detailed analysis!