I don't know... Einstein had a lot of love affairs and debated about philosophy with some friends and was very politically active (e.g. he wrote pro-socialism articles for newspapers). In the context of theoretical physics all of this sounds like wasting time to me. Some brains are just exceptionally wired and people are able to accomplish amazing things, given the right mindset and tenacity.
He's the quintessential authority figure; it's not surprising that people will try to add weight to something by attributing it to Einstein. Most people will just think "Ah Einstein was really clever, if he said that there must be something to it".
This is an extremely good analogy that I must remember (if case anyone quotes this myth to me again, which hasn't happened in a couple of decades... :)
Related to that, I find I often end up getting into arguments with people over the "computers run at such faster speeds than humans" myths/tropes. It's very common in the cyberpunk genre especially that time "runs faster" for people inside simulation spaces (metaverses/The Matrix/whatever). (Or sometimes the other direction where robots/cyborgs are significantly faster than humans at "thinking".)
Computers are very fast at raw, boring mathematical calculations such addition, but that doesn't transfer to entire physics simulations or human scale activities. Think about how much time it takes a large file to download/transfer, for instance. A lot of videogame design through history and continuing through the modern day is fighting to keep the physics simulation close to real time and lot of classic videogame glitches and/or bugs are when it takes longer than real time to simulate. (With the fun that the Halting Problem tells us that we'll never be certain exactly how long it takes to run algorithms, in general.)
Well, with sufficiently advanced computers, your simulations can run as fast or your cyborgs think as fast as the plot demands.
Many video games are and always have been pushing limits. So you'd expect many them to barely run on their contemporary hardware.
(A notable exception is Tetris. Thanks to being invented by people on obsolete Soviet computers, no western computer ever had any trouble running it.)
> With the fun that the Halting Problem tells us that we'll never be certain exactly how long it takes to run algorithms, in general.
The Halting Problem is barely relevant here: it only applies to arbitrary programs. It doesn't have any bearing on running one specific computation faster.
> Well, with sufficiently advanced computers, your simulations can run as fast or your cyborgs think as fast as the plot demands.
We've reached the edges of Moore's Law. We're starting to see the horizon where "sufficiently advanced computers" end. You might be fine with plots playing fast and loose with physics limitations and optimistically still believe tomorrow's computers will be "magic". I'm just saying that it is starting to bug me versus realism in the exact same sorts of fashion as "humans only use 10% of their brains" tropes. "Computers only use 10% of their transistors." Just like the bad human trope, it ignores that most transistors are highly specialized (an FPU is not a GPU is not a general purpose CPU).
> The Halting Problem is barely relevant here: it only applies to arbitrary programs. It doesn't have any bearing on running one specific computation faster.
We're talking about extremely arbitrary programs from stories that are being done for the first time ever. "Hacks" that no one has ever done before. "Holodeck simulations" built in magic amounts of time just on a couple of spoken sentences. Humans "diving into" computer dreamscapes and doing all sorts of human scale simulation things in some sort of "quick time" which seems to be directly in the intersection of both the "humans only use 10% of their brains" myths and the "computers run really fast" myths. Those aren't just well studied apps running standalone without human interaction and with just "one specific computation". Those are REPLs and other human input devices including speech to text to ML (?) and complicated physics calculus that isn't just add/subtract things real fast and all sorts of other things that take wall clock time and are being "programmed" for the first time, on the fly. Of course The Halting Problem applies to all of those types of scenarios.
I certainly enjoy watching/reading such magic computer fantasies when I can muster a suspension of disbelief, but I've also started to feel like most of them are written by people that have never faced the contemporary realities of computer programming and believe in a myth of magically faster computers and human interface design that doesn't exist today and pessimistically may never exist.
What makes you think we reached the edge of Moore's Law? It's often been pronounced dead, but it's still going strong. Especially if you consider that not just CPUs, but GPUs (and Systems-on-a-chip etc) can also be torch-bearers as far as the Law is concerned.
Even present day chips are already much faster than neurons. Oversimplified: our transistors work at (the same order of magnitude as) the speed of light; while neurons work at (the same order of magnitude as) the speed of sound.
As far as we can tell, we can just use more and more transistors in parallel, and will eventually get something that can probably run a mind much faster than a human brain can. I have no opinion on whether that's actually the case, but it seems a plausible enough assumption for a story.
> Just like the bad human trope, it ignores that most transistors are highly specialized (an FPU is not a GPU is not a general purpose CPU).
You are right for CPU-type workloads. But if your story take how present day GPUs run present day neural nets, and extrapolates from there, you could very plausibly have an architecture that runs general intelligence, uses most of its transistors most the time, but the hardware looks extremely specialised for one very specific type of computation.
(I suspect that it is battery-energy consumption and waste heat consideration that will make us not use most transistors powering our cyborgs most of the time.
We have already been seeing dynamic frequency scaling with CPUs for ages even on Desktop. Smartphones made these energy considerations even more of a priority.)
> We're talking about extremely arbitrary programs from stories that are being done for the first time ever. "Hacks" that no one has ever done before. "Holodeck simulations" built in magic amounts of time just on a couple of spoken sentences.
Doesn't sound magic at all to me. Today Dall-E 2 already produces awesome images from short descriptions, and GPT-3 can produce stories. Building a complete holodeck simulation is a straightforward extrapolation. (Yet again, I don't know whether in reality there will be a smooth road from today's technology. But it's more than plausible enough for fiction.)
The Halting Problem doesn't apply to these kinds of things at all. Dall-E 2 and GTP-3 aren't even Turing complete.
> [...] but I've also started to feel like most of them are written by people that have never faced the contemporary realities of computer programming [...]
I'm sorry I'm about to drop a ton of pessimism here. I know HN loves to be a place of unfettered technical optimism, but this is a thread about myths in fiction and bringing hard reality to them, so it seems a useful exercise to apply some "glass is half empty" to it in this thread.
> What makes you think we reached the edge of Moore's Law? It's often been pronounced dead, but it's still going strong.
Moore's Law was formulated with specifics. It had specific goal posts. We passed those goal posts, so we moved them. You are right, we still believe that we can keep moving those goal posts. I characterized this as "we can see the horizon" (not that we've reached the current edges, but we have a decent idea now where they are), we can see that point in the (near?) future where we will eventually have moved the goal posts so far that we've either lost sight of the original goal line or we've lost sight of why we are still moving the goal posts at all.
> As far as we can tell, we can just use more and more transistors in parallel, and will eventually get something that can probably run a mind much faster than a human brain can.
Transistors still have to obey the laws of physics as we know them. The speed of light is incredibly fast, but it's still a very hard speed limit on everything we think to accomplish. Grace Hopper famously carried around wires to remind people the distance of a nanosecond. It's still shorter than you think it should be even knowing how large a light year is. Nanoseconds add up. You can use a lot of transistors in parallel but you still have to route data to and between them. Even if you are routing the data at light speed (and that's quite a feat), delays add up. There's only so much stuff you can route.
> Today Dall-E 2 already produces awesome images from short descriptions, and GPT-3 can produce stories.
Neural networks are still just Sparkling Statistics Models. They seem amazing to us humans, but in general we humans are terrible at understanding statistics and statistical reasoning. This is in large part why Casinos exist. No one expects AGI to result from Las Vegas running enough Casinos in the same place. But you swap the poker chips for pixels or stories, suddenly people expect a Casino to grow into AGI and that they can extrapolate how AGI will likely work based on Casinos that they have seen.
It's lovely optimism that AGI would look anything like today's neural networks.
But beyond that, do you have a real sense of how much wall clock time it takes for a model like Dall-E 2 or GPT-3? Even ignoring that each one was years and months of wall clock time to "train" in the first place, right now most people's interactions with these models aren't on their own hardware but on corporate-sponsored clouds of servers. Sure, it's easy to imagine a starship might have a whole cloud of servers (just for holodeck simulation alone, though?), but it's still useful to have a wall clock time appreciation for even these "simple Casinos" we've built:
Both Dall-E 2 and GPT-3 models are multiple gigabytes in size each. If you try to run one on a single device you will find out that it is not as "instantaneous" as it feels when borrowing a corporate sponsored cloud. It will take minutes just to load those full models into RAM and/or VRAM, presuming that you have enough free RAM/VRAM to accommodate the full models. (Otherwise you can expect additional wall clock time to page parts of the models in and out of RAM.) CPUs and GPUs are still amazingly fast at spinning a lot of roulette wheels, dealing a ton of poker hands, and rolling a lot of dice, but the staggering numbers of such still add up when running a model like this and it can take minutes to hours of wall clock time on a single device.
> The Halting Problem doesn't apply to these kinds of things at all. Dall-E 2 and GTP-3 aren't even Turing complete...
Somewhat coincidentally, I watched it - the film Lucy - yesterday and can confirm that it does indeed get worse. I would have liked it better if the story arc was more like 'Flowers for Algernon'.
We use only 10% of our roads, cars only need the parts that their tires touches so we could save 90% by just building those parts and skip everything else!
> It's a bit like saying a piano player will only use 10% of the keys at a time.
In a pedantic way, it is close to the truth as most pianists use up to 10 keys at a time (a few more if you count pedals). There are many keys available to the pianist but the pianist remains constrained by their input capacity.
I remember seeing Ben Folds once slam his forearm on the keyboard as a deliberately dissonant sort of percussive stab… that was 15-20 keys with just one arm!
>Studies of brain damage: If 10 percent of the brain is normally used, then damage to other areas should not impair performance
This is unbelievably naive to use as some sort of proof, using 10% does not imply that its 10% of the physical organ. Rather 10% compared to what one could do with better training, improved lifestyle, etc. And this does not have to be purely intellectual, it can be imagination capabilities, communication skills or hundreds of other things which a particular individual did not develop fully.
It looks like it. I am not related to any brain studies so naturally I can not go deep into the topic. The main thing I wonder about here is does this redundancy mechanism come for free or does it come at the cost of some other (possibly untapped) potential?
> Rather 10% compared to what one could do with better training, improved lifestyle, etc.
If you're talking about 10% of some non-physical concept of brain "potential", then you have to define, potential WHAT? Creativity? Intelligence? Speed of thought? Hand/eye coordination? Something else deemed important? All of it? None of it? How could you possibly attach a 10% potential value to any of this unless you operationalize it down to some tiny measurable part that lose the gestalt? If we're not defining the 10% as physical brain area, then the 10% proposition seems not only unlikely true but likely meaningless, a priori.
I don't have to define or measure it. And this does not make it meaningless either. In fact most profound things in life can not be measured. Can you measure how much parents love their child? Can you measure how happy someone is to get a present they wanted? Can you measure how funny a joke is? Measuring everything is a stupid approach to life and things like brain development and human potential for sure are immeasurable but are the most significant.
It's amazing to think how common these sorts of myths used to be before the internet. If you heard something like this and wanted to challenge it, you would have to get on the bus and spend a couple of hours at the library and return with some photocopies. Now you can pull up Wikipedia on your phone.
Absolutely true, but there somehow seems different quality to them. The myths of old seemed to be little factoids that were not easy to verify and were taken at face value, then just spread until they became "common knowledge". The new ones seem to have an almost religious quality to them. Flat-earthers for example, if someone told me 25 years ago that this would be a thing I would have just laughed.
> if someone told me 25 years ago that this would be a thing I would have just laughed.
Wikipedia says
In the modern era, the pseudoscientific belief in a flat Earth originated with the English writer Samuel Rowbotham with the 1849 pamphlet Zetetic Astronomy.
In 1956, Samuel Shenton set up the International Flat Earth Research Society (IFERS), better known as the "Flat Earth Society" from Dover, England, as a direct descendant of the Universal Zetetic Society.
so flat earth has been around for 170 years. And people did laugh at them:
The term flat-earth-man, used in a derogatory sense to mean anyone who holds ridiculously antiquated or impossible views, predates the more compact flat-earther. It was recorded in 1908: "Fewer votes than one would have thought possible for any human candidate, were he even a flat-earth-man."
A "modern" conspiracy might be the Sandy Hook, Pizza-gate, etc. but those originate with Alex Jones so not a general example. Maybe "the moon landings were faked" is a good example?
I don't doubt that the belief has always existed somewhere; tribes with barely any outside contact, for example. I suspected that the belief was once reserved for a few eccentric crackpots and didn't have the momentum to be a movement until the last decade or so, but maybe I'm wrong. Flat-earth theory probably wasn't the greatest example. I think a "myth" connotes something a bit more innocent than a delusional belief or conspiratorial thinking (I'd call the Loch Ness Monster an myth, but I'd call the "moon-landing was faked" a conspiracy).
Oh my gosh, there are tons of them. My mother-in-law ensures I hear about them all, passionately, each time she visits.
Bill and Melinda Gates are cyborgs whose bodies have been replaced. They're trying to track everyone with vaccine microchips. (There were variations on this theme including 5G)
Michelle Obama is a man.
9/11 buildings were vaporized by a high-powered, government laser.
Chemtrails are full of heavy metal and are being used to control our minds.
Putting hydrogen peroxide in your ear can cure all sorts of things, including cancer and cystic fibrosis.
"The Jews" are disproportionately represented in the government and media and are behind the whole woke movement to punish white people.
Any number of things about George Soros and ANTIFA.
The moon landing was faked.
The list goes on and on. Every time it's something even more crazy than the last.
Flat-earth is a conspiracy theory, flat-earthers are told that the earth is round, but they don't believe it, they usually also believe in all sorts of other theories: fake moon landing, chemtrails, covid 5G, aliens, etc... make your choice. Conspiracy theories went strong even before the internet, some had devastating consequence. At least, flat-earth is harmless, it only makes you sound stupid.
And I also believe that there are not that many "true" flat earthers. There is a significant amount of trolls and people who like the exercise in epistemology.
I think the reason flat earth get so much visibility is that it is one the most obviously wrong belief there is, people like to point out dumb people in order to feel smart by comparison. In fact, the meme is not the idea that the earth is flat, it is that there are people out there who believe the earth is flat. It is almost a meta-myth, I mean, what proof do you have that flat earth as a community is a thing? I know there is rather good evidence that it is the case but who bothered to check?
Flat earth is not harmless. Like many conspiracy theories, it posits an "evil controlling organization" that for some reason (usually to hide god?) is able to direct the entirety of the earth towards its goals.
These kind of conspiracy theories, if you are willing to keep asking questions, almost always blame "The Jews". People pushing these conspiracy theories are, whether they understand it or not, pushing the ideas that 1: You have zero control over your government and cannot gain any control, 2: Science isn't the goal of many science institutions, and instead they are merely mouthpieces for this global cabal, 3: This global cabal even has the power of fooling your own senses, like many of these believers would not be swayed if you put them on a rocket and let them see the earth from space. These people are prone to dismissing strong evidence with hand wavey excuses like "oh virtual reality" or similar.
I would argue the point of these is to push 1: voter apathy and general detachment from the ways society can be changed and improved through their effort, 2: anti-science and anti-intellectualism sentiments, 3: "You can't trust anyone but us", 4: "The Jews" bullshit
Pre-internet, it was harder to debunk myths but also harder to spread them. Nowadays, it’s easier to debunk myths but also much easier to spread them. These do not necessarily cancel out; as they say, “the amount of energy needed to refute bullshit is an order of magnitude larger than to produce it.”
As an extreme example, pre-internet, a conspiracy theory-inclined person might randomly run across a handful of full-blown conspiracy theorists in his day-to-day life. This makes it easier for him to question his leanings, since he only observes a small minority of people going all-in on the conspiracy theory. Nowadays, a conspiracy-inclined person can easily find online communities comprising all other full-blown conspiracy theorists everywhere the world. This quickly reinforces his belief in the conspiracy, since he sees an apparent majority of people in this online community believing the myth.
I honestly think it's easy to get secondary sources now, but it's still hard to get primary sources. If people have different views of preventing covid they're arguing over different news articles. They aren't arguing about specific studies that they have easy access to.
Hard agree. It's very easy to find people who agree with whatever your particular brand of crazy is. And it does not help that online people can say things they would never say face to face.
This myth does not require Wikipedia to dispel, just the knowledge of how to calculate percentages and the knowledge that you need a standard of measurement. (Which none exists for “brain capacity”).
Both have the issue that the popular interpretation of them are great oversimplifications of models that themselves do seem to have some limited use in neuroscience, but only if you add enough caveats.
So unlike the situation with, say, flat earthers, there's some kernel of truth to it that can be misinterpreted or willfully presented in a misleading way.
It will be harder to dispel because the underlying general assumption actually is true - in most people the brain does split some functions by hemisphere, but rather it's not a general statement that can be simply applied:
> Some popularizations oversimplify the science about lateralization, by presenting the functional differences between hemispheres as being more absolute than is actually the case.[21]: 107 [22]
Sources that wikipedia uses:
[21] Westen D, Burton L, Kowalski K (2006). Psychology : Australian and New Zealand edition. Milton, Qld.: John Wiley & Sons. ISBN 9780470805527.
I intuitively understand that we're using the whole of our brains in the same way that we use the full capacity of NNs and discard the majority of results. In other words the"unused" 90% of our brain is actually importantly figuring out all the less good fits that are essential for the ultimate deductions we utilize.
Another similar myth is the one Malcom Gladwell made up about how it takes 10 000 hours of doing something to become an expert at that thing. It was based on research that used 10 000 as an average for 20 year old violinists. The original researcher, Anders Ericsson, focuses much more on deliberate practice than sheer quantity.
10,000 is a round number representing a large fraction of a young person’s life, ~20 hours a week for 10 years. It seems significant because we assume talented young people can be experts if they spend a lot of time on something, but it really just reflects the criteria chosen for defining experts and time on a subject.
A prerequisite of becoming a Doctor for example is learning how to read, but that’s not considered part of the time commitment. Recursively when considering how long it takes to be an expert Heart Surgeon there is a tendency to ignore how long it takes to become a Doctor.
Ericsson's main point is that talent isn't as important as a lot of people, including other experts in the field, think. What matters is identifying your weaknesses and training specifically to improve those, often with the help of a coach. Just doing something for a long time doesn't necessarily make you great at it.
That falls into the same trap of ignoring all the time perfecting prerequisite skills. An Olympic marathon runner needed to get good at balancing before they could learn to walk, and then very good at walking etc.
> 10,000 being a round number is how it was chosen arbitrarily by the Japanese
According to a 2019 article from Harvard Health, “10,000 steps” was a marketing term for hardware made by a Japanese company. (Not sure if the original root goes back to Japanese folklore, but the article implies it doesn’t.)
> Dr. Lee discovered that the origins of the number go back to 1965, when a Japanese company made a device named Manpo-kei, which translates to "10,000 steps meter." “The name was a marketing tool,” she says. [0]
I don't think Gladwell is a particularly insightful guy, but as someone who has actually read a couple of his books, I feel like the "10,000 hours" thing is frequently misrepresented.
The claim isn't "put in 10k hours and you'll magically become an expert". It's "if you want to become an expert, you have to put in 10k hours". Necessary but not sufficient.
The 10k hours number is also dubious. It just happens to be the average number of hours students at some special violin school have spent in solitaire practice by the time they turn 18 (or what ever age they apply for the school).
What does it even mean to be an expert? What are we allowed or not allowed to be an expert in? Who's the judge that sets what fields you can be an expert in? Is there a range of expertise? Do I have to be in the top 10% of people in a field? The top 2%?
The whole idea of being an expert is an arbitrary social designation backed up by hogwash. We think we know what it means but it's really just hand-waving.
I also read that specific book/chapter where he discusses 10 000 hours, and the implication based on his bill gates and hockey example, were as you said, that it is a necessary condition. And those examples in particular I think point out how having had the time and experience to get ahead of your "cohort" (people of a similar age or time period)
Another missing point, and this happens frequently also in rigorous treatment-control studies, is the heterogeneity of responses to treatment. Say, 10k hours might be the average deliberate practice time needed to reach "expert level" (not true), but for some it can take 1k hours, for others 15k, for others infinite time (but they can drop off), and so on.
This reminds me of the first "scientific myth" I witnessed.
It was in the 80's in France where we had a photocopied page with "the most dangerous additives". The worst was E330. 40 years later, I still remember it.
My parents had a copy from sime friends and these well educated people (dad is an engineer and mom a lawyer) checked everything against the list. Just because it was "from the Villejuif hospital".
I don't remember the exact context at this point, but I've definitely encountered someone complaining about "ascorbic acid" as an additive, which is of course vitamin C.
I first started seeing the "dihydrogen monoxide" or "DHMO" memes a couple of decades ago when there were serious concerns about DMSO (dimethyl sulfoxide), a solvent that can carry drugs through the skin, which apparently has some medical use but can be risky. It's a by-product of paper making.
A so-called "enabling component" of acid rain, and found in all polluted waterways in the world.
That said, nowadays I don't consider these jokes so funny anymore, people really believe similar stuff these days and post-truth bs is used to gain real power by people without shame.
This. I stopped trying to be smart-ass with such jokes because when you read things people believe in, it is disheartening.
I am not even thinking about religion or homeopathy, but even benign things they read on Internet and apply to their everyday life. Nobody (exceptions exist) checks them before adding them into their routine. It is enough that Aunt Mary said that (she is an engineer after all), or your typical celebrities gazette added it to their "Did You Know That...?" section.
Dihydrogen monoxide is actually a cooling agent for nuclear power plants. It's lethal at high doses, with hundreds of thousands of fatal accidents per year, and yet they put it into our food because it's cheap.
There was one time that a group of Santa Monica High School kids were at the Sunday farmer’s market trying to make videos of people being fooled into a moral panic about ascorbic acid in the produce. I remember saying to them, “You mean Vitamin C?” and the interviewer shook her head and said “You’re too smart” and she and her cameraman walked away in search of a more amenable victim.
I am just going to plug my favorite podcast here. It's usually about conspiracy theories but they are currently doing a special about how bad ideas occasionally flow from the top with disastrous effect, unfortunately the special is pay-walled. But the conspiracy - Q stuff is pretty good too.
“Insofar as she recognized at all that she was dreaming, she realized that she must be exploring her subconscious mind. She had heard it said that humans are supposed only to use about a tenth of their brains, and that no one was really clear what the other nine tenths were for, but she had certainly never heard it suggested that they were used for storing penguins.”
― Douglas Adams, The Long Dark Tea-Time of the Soul
I always thought this one could be taken apart on simple common sense grounds. The brain is very "costly" for the human species to maintain, in terms of calories required to run it, and risk during childbirth (amongst other factors, no doubt). Given this, why have it, and leave 90% of it unused? Mr Darwin would give a thumbs down to such a strategy.
Talking about the brain in terms of what percentage of it you use is silly. But I think there’s an important grain of truth in the idea of vast, untapped intellectual potential in many people. A better statement may be something like “many people only realise a small proportion of their potential intellect”, which I strongly believe. I just think it’s little to do with the brain and mostly to do with emotional associations and implicit limiting beliefs.
Anecdotally I have seen countless examples of people being able to do impressive feats of mathematical thinking, but only as long as they don’t realise it’s ‘maths’. And as soon as they realise, it often stops working; the deeply learned negative associations sadly seem to quickly overpower any fleeting buzz of “oh I can do this after all”. I think untold damage is done in schools, instilling into kids associations of maths (or other subjects) as a thing that must be learned (for a grade, or to get a teacher/parent off their back) instead of as a useful tool they can turn to when they want to solve a problem they have.
One of my favorite myths is that an idea can be declared false simply because the science does not yet exist to prove it true. I think we all have beliefs that have not yet been verified. The key is to let go of those beliefs when the science does show the errors in those beliefs. Myths have played an important role in helping us understand our surroundings. Certainly, though, they have often been misused and exploited.
One of the myths I learned early on was that the world was somehow created. Science tells us that this world did not always exist and formed over time in the past. So rather than getting dispelled, this idea is now both a myth of the past and a modern day fact as verified by scientific observation.
Notice that the referenced Wikipedia article very carefully does not say this myth has been dispelled by science. It uses its strongest suggestive statement to disagree with only a very narrow interpretation of the myth.
So, while we're sure the original intent of the mythology is misleading and inaccurate, we cannot rely on science to prove us right. Give it some time.
> The key is to let go of those beliefs when the science does show the errors in those beliefs.
Perhaps I'm misunderstanding you, but it sounds to me like you're saying that as long as somebody hasn't disproved a belief, you're justified in believing it. Is this what you're saying? That seems like it might lead you to keep believing things for which you have insufficient evidence to conclude they're true, and as such would be a less than ideal epistemological tool.
That is not at all what I am saying. The thought that I am taking to task is that some people seem to think ‘myth’ is synonymous with ‘false.’ It is not.
And how about the bumblebee's impossibility of flight myth? I've also hear it regarding generic 'beetle', and with specific beetles. I always thought it was pretty obvious that we just didn't understand the aerodynamics, or aerodynamics per se, enough. I mean, there's a special kind of insect that gets to bend the laws of physics? Really?
From the first time I heard this bit of folklore as a child, I assumed the message was not that bumblebees are magic, but that our scientific models will likely always need continued revising. The argument as to whether we actually do understand bumblebee flight is interesting to entomology, but doesn't really impact the point of the myth.
132 comments
[ 0.22 ms ] story [ 238 ms ] threadhttps://interestingengineering.com/13-inspiring-einstein-quo...
(Which makes no sense, when they go multiple layers..)
Computers are very fast at raw, boring mathematical calculations such addition, but that doesn't transfer to entire physics simulations or human scale activities. Think about how much time it takes a large file to download/transfer, for instance. A lot of videogame design through history and continuing through the modern day is fighting to keep the physics simulation close to real time and lot of classic videogame glitches and/or bugs are when it takes longer than real time to simulate. (With the fun that the Halting Problem tells us that we'll never be certain exactly how long it takes to run algorithms, in general.)
Many video games are and always have been pushing limits. So you'd expect many them to barely run on their contemporary hardware.
(A notable exception is Tetris. Thanks to being invented by people on obsolete Soviet computers, no western computer ever had any trouble running it.)
> With the fun that the Halting Problem tells us that we'll never be certain exactly how long it takes to run algorithms, in general.
The Halting Problem is barely relevant here: it only applies to arbitrary programs. It doesn't have any bearing on running one specific computation faster.
We've reached the edges of Moore's Law. We're starting to see the horizon where "sufficiently advanced computers" end. You might be fine with plots playing fast and loose with physics limitations and optimistically still believe tomorrow's computers will be "magic". I'm just saying that it is starting to bug me versus realism in the exact same sorts of fashion as "humans only use 10% of their brains" tropes. "Computers only use 10% of their transistors." Just like the bad human trope, it ignores that most transistors are highly specialized (an FPU is not a GPU is not a general purpose CPU).
> The Halting Problem is barely relevant here: it only applies to arbitrary programs. It doesn't have any bearing on running one specific computation faster.
We're talking about extremely arbitrary programs from stories that are being done for the first time ever. "Hacks" that no one has ever done before. "Holodeck simulations" built in magic amounts of time just on a couple of spoken sentences. Humans "diving into" computer dreamscapes and doing all sorts of human scale simulation things in some sort of "quick time" which seems to be directly in the intersection of both the "humans only use 10% of their brains" myths and the "computers run really fast" myths. Those aren't just well studied apps running standalone without human interaction and with just "one specific computation". Those are REPLs and other human input devices including speech to text to ML (?) and complicated physics calculus that isn't just add/subtract things real fast and all sorts of other things that take wall clock time and are being "programmed" for the first time, on the fly. Of course The Halting Problem applies to all of those types of scenarios.
I certainly enjoy watching/reading such magic computer fantasies when I can muster a suspension of disbelief, but I've also started to feel like most of them are written by people that have never faced the contemporary realities of computer programming and believe in a myth of magically faster computers and human interface design that doesn't exist today and pessimistically may never exist.
Even present day chips are already much faster than neurons. Oversimplified: our transistors work at (the same order of magnitude as) the speed of light; while neurons work at (the same order of magnitude as) the speed of sound.
As far as we can tell, we can just use more and more transistors in parallel, and will eventually get something that can probably run a mind much faster than a human brain can. I have no opinion on whether that's actually the case, but it seems a plausible enough assumption for a story.
> Just like the bad human trope, it ignores that most transistors are highly specialized (an FPU is not a GPU is not a general purpose CPU).
You are right for CPU-type workloads. But if your story take how present day GPUs run present day neural nets, and extrapolates from there, you could very plausibly have an architecture that runs general intelligence, uses most of its transistors most the time, but the hardware looks extremely specialised for one very specific type of computation.
(I suspect that it is battery-energy consumption and waste heat consideration that will make us not use most transistors powering our cyborgs most of the time.
We have already been seeing dynamic frequency scaling with CPUs for ages even on Desktop. Smartphones made these energy considerations even more of a priority.)
> We're talking about extremely arbitrary programs from stories that are being done for the first time ever. "Hacks" that no one has ever done before. "Holodeck simulations" built in magic amounts of time just on a couple of spoken sentences.
Doesn't sound magic at all to me. Today Dall-E 2 already produces awesome images from short descriptions, and GPT-3 can produce stories. Building a complete holodeck simulation is a straightforward extrapolation. (Yet again, I don't know whether in reality there will be a smooth road from today's technology. But it's more than plausible enough for fiction.)
The Halting Problem doesn't apply to these kinds of things at all. Dall-E 2 and GTP-3 aren't even Turing complete.
> [...] but I've also started to feel like most of them are written by people that have never faced the contemporary realities of computer programming [...]
That's most likely true for most authors, yes.
> What makes you think we reached the edge of Moore's Law? It's often been pronounced dead, but it's still going strong.
Moore's Law was formulated with specifics. It had specific goal posts. We passed those goal posts, so we moved them. You are right, we still believe that we can keep moving those goal posts. I characterized this as "we can see the horizon" (not that we've reached the current edges, but we have a decent idea now where they are), we can see that point in the (near?) future where we will eventually have moved the goal posts so far that we've either lost sight of the original goal line or we've lost sight of why we are still moving the goal posts at all.
> As far as we can tell, we can just use more and more transistors in parallel, and will eventually get something that can probably run a mind much faster than a human brain can.
Transistors still have to obey the laws of physics as we know them. The speed of light is incredibly fast, but it's still a very hard speed limit on everything we think to accomplish. Grace Hopper famously carried around wires to remind people the distance of a nanosecond. It's still shorter than you think it should be even knowing how large a light year is. Nanoseconds add up. You can use a lot of transistors in parallel but you still have to route data to and between them. Even if you are routing the data at light speed (and that's quite a feat), delays add up. There's only so much stuff you can route.
> Today Dall-E 2 already produces awesome images from short descriptions, and GPT-3 can produce stories.
Neural networks are still just Sparkling Statistics Models. They seem amazing to us humans, but in general we humans are terrible at understanding statistics and statistical reasoning. This is in large part why Casinos exist. No one expects AGI to result from Las Vegas running enough Casinos in the same place. But you swap the poker chips for pixels or stories, suddenly people expect a Casino to grow into AGI and that they can extrapolate how AGI will likely work based on Casinos that they have seen.
It's lovely optimism that AGI would look anything like today's neural networks.
But beyond that, do you have a real sense of how much wall clock time it takes for a model like Dall-E 2 or GPT-3? Even ignoring that each one was years and months of wall clock time to "train" in the first place, right now most people's interactions with these models aren't on their own hardware but on corporate-sponsored clouds of servers. Sure, it's easy to imagine a starship might have a whole cloud of servers (just for holodeck simulation alone, though?), but it's still useful to have a wall clock time appreciation for even these "simple Casinos" we've built:
Both Dall-E 2 and GPT-3 models are multiple gigabytes in size each. If you try to run one on a single device you will find out that it is not as "instantaneous" as it feels when borrowing a corporate sponsored cloud. It will take minutes just to load those full models into RAM and/or VRAM, presuming that you have enough free RAM/VRAM to accommodate the full models. (Otherwise you can expect additional wall clock time to page parts of the models in and out of RAM.) CPUs and GPUs are still amazingly fast at spinning a lot of roulette wheels, dealing a ton of poker hands, and rolling a lot of dice, but the staggering numbers of such still add up when running a model like this and it can take minutes to hours of wall clock time on a single device.
> The Halting Problem doesn't apply to these kinds of things at all. Dall-E 2 and GTP-3 aren't even Turing complete...
In a pedantic way, it is close to the truth as most pianists use up to 10 keys at a time (a few more if you count pedals). There are many keys available to the pianist but the pianist remains constrained by their input capacity.
I remember seeing Ben Folds once slam his forearm on the keyboard as a deliberately dissonant sort of percussive stab… that was 15-20 keys with just one arm!
This is unbelievably naive to use as some sort of proof, using 10% does not imply that its 10% of the physical organ. Rather 10% compared to what one could do with better training, improved lifestyle, etc. And this does not have to be purely intellectual, it can be imagination capabilities, communication skills or hundreds of other things which a particular individual did not develop fully.
If you're talking about 10% of some non-physical concept of brain "potential", then you have to define, potential WHAT? Creativity? Intelligence? Speed of thought? Hand/eye coordination? Something else deemed important? All of it? None of it? How could you possibly attach a 10% potential value to any of this unless you operationalize it down to some tiny measurable part that lose the gestalt? If we're not defining the 10% as physical brain area, then the 10% proposition seems not only unlikely true but likely meaningless, a priori.
It's a crazy phenomenon to watch play out.
Wikipedia says
so flat earth has been around for 170 years. And people did laugh at them: A "modern" conspiracy might be the Sandy Hook, Pizza-gate, etc. but those originate with Alex Jones so not a general example. Maybe "the moon landings were faked" is a good example?Bill and Melinda Gates are cyborgs whose bodies have been replaced. They're trying to track everyone with vaccine microchips. (There were variations on this theme including 5G)
Michelle Obama is a man.
9/11 buildings were vaporized by a high-powered, government laser.
Chemtrails are full of heavy metal and are being used to control our minds.
Putting hydrogen peroxide in your ear can cure all sorts of things, including cancer and cystic fibrosis.
"The Jews" are disproportionately represented in the government and media and are behind the whole woke movement to punish white people.
Any number of things about George Soros and ANTIFA.
The moon landing was faked.
The list goes on and on. Every time it's something even more crazy than the last.
And I also believe that there are not that many "true" flat earthers. There is a significant amount of trolls and people who like the exercise in epistemology.
I think the reason flat earth get so much visibility is that it is one the most obviously wrong belief there is, people like to point out dumb people in order to feel smart by comparison. In fact, the meme is not the idea that the earth is flat, it is that there are people out there who believe the earth is flat. It is almost a meta-myth, I mean, what proof do you have that flat earth as a community is a thing? I know there is rather good evidence that it is the case but who bothered to check?
These kind of conspiracy theories, if you are willing to keep asking questions, almost always blame "The Jews". People pushing these conspiracy theories are, whether they understand it or not, pushing the ideas that 1: You have zero control over your government and cannot gain any control, 2: Science isn't the goal of many science institutions, and instead they are merely mouthpieces for this global cabal, 3: This global cabal even has the power of fooling your own senses, like many of these believers would not be swayed if you put them on a rocket and let them see the earth from space. These people are prone to dismissing strong evidence with hand wavey excuses like "oh virtual reality" or similar.
I would argue the point of these is to push 1: voter apathy and general detachment from the ways society can be changed and improved through their effort, 2: anti-science and anti-intellectualism sentiments, 3: "You can't trust anyone but us", 4: "The Jews" bullshit
As an extreme example, pre-internet, a conspiracy theory-inclined person might randomly run across a handful of full-blown conspiracy theorists in his day-to-day life. This makes it easier for him to question his leanings, since he only observes a small minority of people going all-in on the conspiracy theory. Nowadays, a conspiracy-inclined person can easily find online communities comprising all other full-blown conspiracy theorists everywhere the world. This quickly reinforces his belief in the conspiracy, since he sees an apparent majority of people in this online community believing the myth.
Like bielefeld existing?
https://en.wikipedia.org/wiki/List_of_common_misconceptions
Shame that "pull up Wikipedia" is apparently still more effort for most people than "tell other people about it on social media".
[1] https://forge.medium.com/maslows-pyramid-is-a-marketing-lie-...
https://en.wikipedia.org/wiki/Lateralization_of_brain_functi...
Both have the issue that the popular interpretation of them are great oversimplifications of models that themselves do seem to have some limited use in neuroscience, but only if you add enough caveats.
So unlike the situation with, say, flat earthers, there's some kernel of truth to it that can be misinterpreted or willfully presented in a misleading way.
[0] https://en.wikipedia.org/wiki/Triune_brain
> Some popularizations oversimplify the science about lateralization, by presenting the functional differences between hemispheres as being more absolute than is actually the case.[21]: 107 [22]
Sources that wikipedia uses:
[21] Westen D, Burton L, Kowalski K (2006). Psychology : Australian and New Zealand edition. Milton, Qld.: John Wiley & Sons. ISBN 9780470805527.
[22] Toga AW, Thompson PM (January 2003). "Mapping brain asymmetry". Nature Reviews. Neuroscience. 4 (1): 37–48. doi:10.1038/nrn1009. PMID 12511860. S2CID 15867592.
https://www.6seconds.org/2022/06/20/10000-hour-rule/
Related Freakonomics episode with Anders Ericsson among others about how to become an expert: https://freakonomics.com/podcast/how-to-become-great-at-just...
A prerequisite of becoming a Doctor for example is learning how to read, but that’s not considered part of the time commitment. Recursively when considering how long it takes to be an expert Heart Surgeon there is a tendency to ignore how long it takes to become a Doctor.
According to a 2019 article from Harvard Health, “10,000 steps” was a marketing term for hardware made by a Japanese company. (Not sure if the original root goes back to Japanese folklore, but the article implies it doesn’t.)
> Dr. Lee discovered that the origins of the number go back to 1965, when a Japanese company made a device named Manpo-kei, which translates to "10,000 steps meter." “The name was a marketing tool,” she says. [0]
[0] https://www.health.harvard.edu/blog/10000-steps-a-day-or-few...
The claim isn't "put in 10k hours and you'll magically become an expert". It's "if you want to become an expert, you have to put in 10k hours". Necessary but not sufficient.
The whole idea of being an expert is an arbitrary social designation backed up by hogwash. We think we know what it means but it's really just hand-waving.
Completely differently, I heard that the brain is an antenna.
It was in the 80's in France where we had a photocopied page with "the most dangerous additives". The worst was E330. 40 years later, I still remember it.
My parents had a copy from sime friends and these well educated people (dad is an engineer and mom a lawyer) checked everything against the list. Just because it was "from the Villejuif hospital".
It was a fake and E330 is citric acid, a benin product. https://en.m.wikipedia.org/wiki/Citric_acid
The famous list: https://en.m.wikipedia.org/wiki/Villejuif_leaflet
Now look at those who have never consumed dihydrogen monoxide -- they haven't killed anyone.
That said, nowadays I don't consider these jokes so funny anymore, people really believe similar stuff these days and post-truth bs is used to gain real power by people without shame.
I am not even thinking about religion or homeopathy, but even benign things they read on Internet and apply to their everyday life. Nobody (exceptions exist) checks them before adding them into their routine. It is enough that Aunt Mary said that (she is an engineer after all), or your typical celebrities gazette added it to their "Did You Know That...?" section.
https://knowyourmeme.com/memes/dihydrogen-monoxide-hoax
https://www.patreon.com/qanonanonymous/posts
― Douglas Adams, The Long Dark Tea-Time of the Soul
Anecdotally I have seen countless examples of people being able to do impressive feats of mathematical thinking, but only as long as they don’t realise it’s ‘maths’. And as soon as they realise, it often stops working; the deeply learned negative associations sadly seem to quickly overpower any fleeting buzz of “oh I can do this after all”. I think untold damage is done in schools, instilling into kids associations of maths (or other subjects) as a thing that must be learned (for a grade, or to get a teacher/parent off their back) instead of as a useful tool they can turn to when they want to solve a problem they have.
One of the myths I learned early on was that the world was somehow created. Science tells us that this world did not always exist and formed over time in the past. So rather than getting dispelled, this idea is now both a myth of the past and a modern day fact as verified by scientific observation.
Notice that the referenced Wikipedia article very carefully does not say this myth has been dispelled by science. It uses its strongest suggestive statement to disagree with only a very narrow interpretation of the myth.
So, while we're sure the original intent of the mythology is misleading and inaccurate, we cannot rely on science to prove us right. Give it some time.
Perhaps I'm misunderstanding you, but it sounds to me like you're saying that as long as somebody hasn't disproved a belief, you're justified in believing it. Is this what you're saying? That seems like it might lead you to keep believing things for which you have insufficient evidence to conclude they're true, and as such would be a less than ideal epistemological tool.