I became interested in learning Go recently after watching the magnificent AlphaGo movie [1] which is free on YouTube. I highly recommend giving it a watch if you haven't already.
Watching the human programmers become dumbfounded as AlphaGo invented novel Go-playing strategies... is what I remember most from having watched this a few years ago (right before GPT3.5/ChatGPT debuted). The algorithm makes [victorious] wildcard moves which no human player would even contemplate [stupid moves become masterplays].
When 9D-master Sudol attributes human qualities to the beauty of his AI opponent's creativity upon formulating certain moves... is definitely eye-opening. Hubris replaced.
I think with the rise of KataGo, its becoming clear that AlphaGo's "dumbfounded" strategies were instead incredibly strong tactical play with hilarious levels of blindness with regards to ladders.
It feels like modern AIs (like KataGo, which are hundreds of times better than AlphaGo) are getting closer to what humans consider appropriate strategic play.
Go players must be humble because if the opponent is stronger then the opponent wins. But Go AI Programmers don't necessarily have to be. Go AI programmers look for the weaknesses, lean upon them and yes, prove that AlphaGo/AlphaZero never learned ladders. Ever. Or other such concepts of strongly forcing moves (ex: loose ladders).
That's one of KataGo's biggest innovations. Explicitly programming a ladder solver so that simple ladders can indeed be factored in by the neural net.
----------
I do wonder how Lee Sodul will react if we told him that the superhuman AI he played could not see ladders... and had other such key weaknesses.
Ladders are a thing so easy that its literally in this beginner series of tutorials. Its a thing that 30kyu beginners learn and master.
Its very strange to me that AlphaGo / AlphaZero was unable to ever learn ladders. It shows that the way humans learn and AIs learn is quite different (or at least, MCTS + Neural Net machines learn differently).
Yes, Go is a strategic game of patterns and perhaps we humans overemphasized the ladder. Nonetheless, its a concept that humans can see and calculate with reasonable speeds that the (earlier) AI was unable to do (and now we've built stronger AIs that can prove this weakness).
---------
This 100% makes the 2016-era discussions about the "strategic brilliance" of AlphaGo come up into question. Now that KataGo is superior (ie: MCTS+NeuralNets + dedicated Ladder Solver), we need to double-check all those "strategically brilliant" moves with the newer AI and see if a ladder messed with them.
Etc. etc.
The things AlphaGo sees aren't necessarily useful to us humans, nor are they useful to modern AI-levels of Go. They're just... that. Trapped in 2016. There should not be any great mystery assigned to the 2016 era game, aside from it being a pivotal moment for AIs.
The game itself is now suspect, now that we know all of AlphaGo's flaws. As Go players, we have better things to study.
DeepBlue was the big chess moment. No one writes or cares about Stockfish vs DeepBlue.
Well, chess players performing opening analysis for their games care. Go players care about KataGo being more readily available and stronger than AlphaGo or AlphaZero. But it's back into the weeds and grind, it's not too interesting a story outside the Chess or Go world.
On the contrary, I would argue that conscious attention is only focused on one of those subroutines at a time. When the ball is in play you focus in it, and everything from your posture to racket handling fades into the background as a subconscious routine. When you make a handling mistake or want to improve something like posture, your focus shifts to that; you attend to it with your attention, and then you focus on something else.
In either case, with working memory for example, conscious contents are limited to at most a basket of 6-7 chunks. This number is very small compared to the incredible parallelism of the unconscious mind.
For all we know, there might be tons of conscious attention processes active in parallel. "You" only get to observe one, but there could be many. You'd never know because the processes do not observably communicate with each other. They do communicate with the same body though, but that is less relevant.
When you are learning a high-performance activity like a sport or musical instrument, the good coaches always get you to focus on only one or at most two things at any time.
The key value of a coach is their ability to assess your skills and the current goals to select what aspect you most need to focus on at that time.
Of course, there can be sequences, like "focus on accurately tossing to a higher point while you serve, then your footwork in the volley", but those really are just one thing at a time.
(edit, add) Yes, all the other aspects of play are going on in the background of your mind, but you are not working actively on changing them.
One of the most insightful observations one of my coaches made on my path to World-Cup level alpine ski racing was:
"We're training your instincts.".
What he meant by that was we were doing drills and focus to change the default — unthinking — mind-body response to an input. so, when X happened, instead of doing the untrained response then having to think about how to do it better (next time because it's already too late), the mind-body's "instinctive" or instant response is the trained motion. And of course doing that all the way across the skill-sets.
And pretty much the only way to train your instincts like that is to focus on it until the desired response is the one that happens without thinking. And then to focus on it again until it's not only the default, but you are now able to finely modulate in that response.
Yes, how, when, & where you focus your eyes is very key to performance. In your case, it seems like the last-second-focus both let your eye track the ball better to your paddle, then focusing on the target let your reflex aim take over, which was evidently pretty good. Not sure if I'd say the long-focus-on-target compromised the early tracking on the ball, or made your swing more 'artificial'.
A funny finding from a study I read which put top pro athletes through a range of perceptual-motor tests. One of the tests was how rapidly they could change focus from near-far-near-far, which of course all kinds of ball players excelled at. The researchers were initially horrified to find racecar drivers were really bad at it, thinking about having to track the world coming at them at nearly 200mph. It turns out of course, that racecar drivers don't use their eyes that way - they are almost always looking further in the distance at the next braking or turn-in point, bump in the track, or whatever, and even in traffic, the other cars aren't changing relative-distance very rapidly.
In this context, we differentiate between the conscious and unconscious based on observability: the conscious is that which is observed, while the unconscious comprises what is not observed.
Then there is the beautiful issue of memory: maybe you are X consciousnesses but only one leaves a memory trace?
Consciousness and memory are two very different things. Don’t think too much about this when you have to undergo surgery. Maybe you are aware during the process but only memory-formation is blocked.
Or perhaps they all leave traces, but all write to the same log? And when reconstructing memory from the log, each constructed consciousness experiences itself as singular?
Which one controls the body? There is a problem there. You can’t just have a bunch of disembodied consciousnesses. Well, maybe.. but that sounds kind of strange.
It’s a single narrative that controls the body is what I mean. If one consciousness says “I am Peter” then other consciousnesses would know that and be conflicted about, if they don’t call themselves that.
What I mean is that a single narrative “wins”, not a multitude. This has to be explained somehow.
How do you know there aren't several different consciousnesses that all think they are Peter?
How do you know they aren't just constructing whatever narrative they prefer to construct from the common pool of memory, ending up with what looks like a single narrative because the parts of the narrative come from the same pool and get written back to the same pool?
Perhaps each consciousness is just a process, like other bodily processes.
Perhaps a human being is less like a machine with a master control and more like an ecosystem of cooperating processes.
Of course, the consciousnesses like to claim to be in charge, but I don't see why I should take their word for it.
No matter how you twist it at some point two consciousnesses differentiate on some contradictory issue maybe not name, but surely they differ on some issue otherwise they wouldn’t be .. different consciousnesses. Life as a human moves and is narrated as a single story, not the story of a thousand processes.
If that were true I can call my heart a process, my liver, etc. They are in a way part of me but they do not just ex nihilo cohere into a single narrative. That is an active process and whatever does that is the only really interesting one (IMO). So I think there might be a bunch of processes, sub personalities maybe, but there remains the problem of integration. Whatever integrates is the one that really fascinates me.
Anyway, thanks for indulging me. It is hard to go into any depth in this medium. I think you have really interesting ideas. Have a nice weekend.
At some moments there has to be a singular decision taken, such as which of two possible options to take. In such a moment some particular consciousness makes the decision, if it’s a decision made by a consciousness (though consciousness takes credit for more decisions than it actually makes, I think).
But granting that point does not grant that there is a single consciousness that is always (or ever) in charge, and it does not grant that any specific consciousness is associated with any specific singular narrative.
We know, scientifically speaking, some things that call the idea of a single consciousness with a single narrative into question. We know, for example from psychology of testimony that the same person’s memory of the same events differs at different times, and that the act of remembering rewrites memories. We have reason to suspect that the brain attributes to conscious choice decisions that are made too quickly for sensory data to reach the brain (and which may therefore be made elsewhere in the nervous system, even though the brain claims to have made the choice after the fact).
And I know from personal experience that some phenomena that normally appear to be singular conscious experiences can devolve into something else under some circumstances. For example, I have experienced blindsight, in which I cannot see something but can nevertheless collect accurate information from it by pointing my eyes at it. I have also experienced being asleep and awake at the same time.
Experiences like these are hard to account for if I assume that my consciousness is singular and continuous and in charge, but not so hard to account for if I assume that it’s a useful illusion cobbled together by a network of cooperating processes that usually (but not always) work well together. For example, many people might claim that it’s nonsense to say that a person can be asleep and awake at the same time, but it’s nonsense only if asleep and awake are mutually exclusive states of a singular consciousness. If, on the other hand, they are two neurological processes that are normally coordinated so that they don’t occur at the same time (because it’s less than useful for them to do so), then it’s not nonsense to observe that under unusual circumstances that coordination might be disrupted. Similarly, if seeing something is one process and consciously experiencing seeing it is a different process—normally, but not necessarily coordinated—then blindsight is not so hard to account for.
Not to mention that it’s trivially easy to find examples of consciousness not being in charge of our behavior, although it likes to think that it is.
I suggest that the supposed singular consciousness, supposedly in charge, may be an illusion constructed by a system of mostly, but not perfectly, coordinated cooperating processes.
There is a long tradition in India, which started with oral transmission of the Vedas, of parallel cognition. It is almost an art form or a mental sport - https://en.wikipedia.org/wiki/Avadhanam
It is the exploration and enumeration of the possible rhythms that led to the discovery of Fibonacci sequence and binary representation in around 200 BC.
Sounds very much sequential, even if very difficult:
> The performer's first reply is not an entire poem. Rather, the poem is created one line at a time. The first questioner speaks and the performer replies with one line. The second questioner then speaks and the performer replies with the previous first line and then a new line. The third questioner then speaks and performer gives his previous first and second lines and a new line and so on. That is, each questioner demands a new task or restriction, the previous tasks, the previous lines of the poem, and a new line.
My point is that what we call conscious and subconscious is limited by our ability to express it in language: since we can't verbalize what's going on quickly enough, we separate those out. Could we learn to verbalize two things at the same time (we all do that as well with say different words and different body language, even consciously, but can we take it a step further? eg. imagine saying nice things to someone and raising the middle finger for someone else behind your back :))
As the whole article is really about the full brain, and it seems you agree our "unconscious mind" producing actions in parallel, I think the focus is wrongly put on brain size, when we lack the expressiveness for what the brain can already do.
Edit: And don't get me wrong, I personally suck at multi-tasking :)
What you consider a single thought is a bit ill defined. A multitude of thoughts together can be formed as a packet, which then can be processed sequentially.
Intelligence is the ability to capture, and predicts events in space and time, and as such it must have the capability to model both things occurring in simultaneity and sequentially.
Sticking to your example, a routine for making a decision in tennis would look something like at a higher level "Run to the left and backstroke the ball", which broken down would be something like "Turn hip and shoulder to the left, extend left leg, extend right, left, right, turn hip/shoulder to the right, swing arm." and so on.
This is great! I love The Devil's Plan, I wish I could play all of the games. Some sort of in-person escape room-style experience re-enacting the games would be the best, but I have also thought about making multiplayer video game versions. Wall Go is a perfect candidate! Well done!
The AI seemed easy to me, I know the rules from watching the show and I won by a large margin first try.
I recently listened to this Diary of a CEO podcast episode [1] discussing the potential link between blood glucose from carbohydrate-heavy diets and cancer. It is an alternate metabolic theory of cancer, and the podcast guest claimed that fasting followed by a keto diet was showing early success as an intervention for cancer patients (as an addition, not a replacement, to existing standard of care). I have not seen the data nor do I know the sample size, but the discussion convinced me to rethink my carb intake.
RealityKit (iOS & visionOS AR/VR/3D library) makes use of USD. Native apps and games for the Vision Pro will also use USD, alongside their new implementation of Entity Component System (ECS).
Yes, but using .reality files is better for stuff like games in RealityKit. For example it would contain a GPU-ready texture instead of a jpeg (and a lot of other differences), so it's more in-line with how assets are typically loaded in games.
One of my favourite board games is called Burn Rate. The premise is a parody of silicon valley during the boom. You compete with the other players for talent (each employee is a card, with varying skill levels).
You could slow down other players by attacking them with “bad idea” cards that would tie up one or more of their engineers for an amount of time. To defend against bad ideas, you need a good manager (one type of employee card).
I could be misremembering some of the details, but the wisdom of the above game mechanic has stuck with me :)
I have some unique insights that I wanted to put into practice. The past month has been the most intellectually stimulating, challenging, and rewarding period my life. Some interesting utility has already appeared. I won’t discuss the details but I will say it does not involve statistical learning.
In scenario #1, there is no financial incentive (read: livable wages for researchers, manufacturers, etc.) to create medicine because your profit is $0, and we have 1000 deaths because nobody makes the medicine.
In theory if the amortized cost of a dose (including R&D salaries, equipment, drug trials, and manufacturing) is truly $1 per dose for such a tiny addressable market, competitors will pop up all over the place and undercut you on price.
From a game theory perspective, a firm choosing option #3 would lose ALL of their market share to another firm choosing option #2, and would have profit of -$1000 because they are unable to sell any of their doses.
>In scenario #1, there is no financial incentive...
FALSE, In scenario 1 there would be a profit but not the maximum possible profit. This means some dudes will have to use fancy cars and not giant yachts.
>In theory if the amortized cost of a dose (including R&D salaries, equipment, drug trials, and manufacturing) is truly $1 per dose for such a tiny addressable market, competitors will pop up all over the place and undercut you on price.
FALSE, competitors will not be allowed to make this for 10 or 20 years.
>From a game theory perspective, a firm choosing option #3 would lose ALL of their market share to another firm choosing option #2, and would have profit of -$1000 because they are unable to sell any of their doses.
FALSE, if there are 2 companies that sell same products there is an equilibrium where you split the market in half and have both more profits then one dominating the market but with small prices. This is the prisoner dilemma, if you cooperate you get better results for both.
Your profit here is zero, in your own hypothetical scenario. Help me understand why the Math is different.
> competitors will not be allowed to make this for 10 or 20 years
You didn't mention your stance on patents and/or intellectual property. If a medicine is patented, would you be in favor of stealing the patented ideas from the inventor in order to lower costs for life-saving medicines through increased competition? This sounds nice, akin to stealing a loaf of bread when you are hungry, but I can't help but think a society that condones stealing inventions from inventors would breed many inventors. Then we are left with the "nobody made the medicine" problem.
EDIT: And by "inventors", realistically I mean "firms willing to sink hundreds of millions of dollars into R&D".
I don't think I have a solution, one idea I have but it probably won't work since billionaires would make sure it won't happen, have a more fair taxes, if you made an obscene profit we would tax it and invest in public research. The issue is big corporation hide their profits, are you wondering why is Google, Apple Store taxing me on the sales and not on my profit? they don't want you to hide your profits as they do.
The only analogue that comes to mind is in financial fraud detection: moving money slowly or in a predictable pattern (monthly rent payments etc.) triggers no alarms, but large or unexpected transfers raise alarms.
I remember when I left my last job that my manager cautioned me against making any large file transfers since it would trigger IT alarms about employees trying to steal the company's IP.
Clearly, he didn't think I was a threat, or if I was, that I would have been smart enough to do it long ago, and slowly :-)
[1] https://www.youtube.com/watch?v=WXuK6gekU1Y