unless... the compute happens also in space? Given how dirt cheap solar has become, how cheap shipping stuff to space is becoming and how little there are clouds and nights in the space making solar power production intermittent, it sounds like it might be economically feasible in not so distant future. (no, I haven't done any math on this. If it checks out, feel free to steal the idea)
The issue with large components (we're talking microns instead of 20nm), is the launch weights (coming down), and power (also coming down). Large components also mean larger silicon dies which are much more expensive, and/or fewer components per die, which means now the CPUs are on different chips and need interconnect, which increases latency and interference. Not impossible, just a load of min-max-ing to do.
You would make the stuff in space, too. Give it a gentle shove off the factory loading dock (factory is on an asteroid) and a couple of years later it shows up in earth orbit, if you get your orbital calculations right…
Not literally a gentle push, but very little rocket action is needed. The gravity well of an asteroid is tiny. The rest can be done with the correct slingshot maneuvers, the problem is calculating it. I am sure I have read something or other from NASA about it.
It's not the asteroid gravity that's the issue, it's the solar gravity field, you still have to perform an orbital transfer from the asteroid orbit to Earth orbit unless you want to leave the computer there and do batch jobs with significant latency.
And the Earth is in space, so if we get to the power consumption level where Earth governments need to care about the direct planetary heating effect of the energy source, it's still a win to do the hard thing (dissipation) somewhere else, like the Moon or something.
It is a slow subway train that requires transfers at Jamaica to the airtrain and then an internal airport people mover. Some people need multiple transfers from
WTC/Penn/14th street are centrally connected stations that should have a direct connection to JFK.
Run an express A-C-E train from central-park, 34th, 14th, WTC, Atlantic, Jamaica, JFK. It should not be that hard. While we are at it, run an express downtown manhattan to Newark train/BRT too. LaGuardia is....hopeless.
The upstream comment ambiguously suggested a train as a solution to an 8-10 hour door to door LA to NYC travel time. Either that means connecting the airport to the city by rail or the cities themselves. Sub-8-hour LA to NYC by train is beyond any currently known technology. JFK is already connected to NYC’s subway by rail.
Sorry I meant that you can use rail to get to and from JFK to the parts of New York served by the subway.
If the claim was to build a sub-8-hour LA to NYC train that’s obviously not going to work because of physics. If the claim was we need rail to LAX and JFK that’s silly because both are already served by rail.
I get a slightly uncomfortable feeling with this talk about AI safety. Not in the sense that there is anything wrong with that (may be or may be not), but in the sense I don't understand what people are talking about when they talk about safety in this context. Could someone explain like I have Asperger (ELIA?) whats this about? What are the "bad actors" possibly going to do? Generate (child) porn/ images with violence etc. and sell them? Pollute the training data so that the racist images pops up when someone wants to get an image of a white pussycat? Or produce images that contain vulnerabilities so that when you open that in your browser you get compromised? Or what?
I'm not part of Stability AI but I can take a stab at this:
> explain like I have ~~Asperger (ELIA?)~~ limited understanding of how the world really works.
The AI is being limited so that it cannot produce any "offensive" content which could end up on the news or go viral and bring negative publicity to Stability AI.
Viral posts containing generated content that brings negative publicity to Stability AI are fine as long as they're not "offensive". For example, wrong number of fingers is fine.
There is not a comprehensive, definitive list of things that are "offensive". Many of them we are aware of - e.g. nudity, child porn, depictions of Muhammad. But for many things it cannot be known a priori whether the current zeitgeist will find it offensive or not (e.g. certain depictions of current political figures, like Trump).
Perhaps they will use AI to help decide what might be offensive if it does not explicitly appear on the blocklist. They will definitely keep updating the "AI Safety" to cover additional offensive edge cases.
It's important to note that "AI Safety", as defined above (cannot produce any "offensive" content which could end up on the news or go viral and bring negative publicity to Stability AI) is not just about facially offensive content, but also about offensive uses for milquetoast content. Stability AI won't want news articles detailing how they're used by fraudsters, for example. So there will be some guards on generating things that look like scans of official documents, etc.
Yes*. At least for the purposes of understanding what the implementations of "AI safety" are most likely to entail. I think that's a very good cognitive model which will lead to high fidelity predictions.
*But to be slightly more charitable, I genuinely think Stability AI / OpenAI / Meta / Google / MidJourney believe that there is significant overlap in the set of protections which are safe for the company, safe for users, and safe for society in a broad sense. But I don't think any released/deployed AI product focuses on the latter two, just the first one.
Examples include:
Society + Company: Depictions of Muhammad could result in small but historically significant moments of civil strife/discord.
Individual + Company: Accidentally generating NSFW content at work could be harmful to a user. Sometimes your prompt won't seem like it would generate NSFW content, but could be adjacent enough: e.g. "I need some art in the style of a 2000's R&B album cover" (See: Sade - Love Deluxe, Monica - Makings of Me, Rihanna - Unapologetic, Janet Jackson - Damita Jo)
Society + Company: Preventing the product from being used for fraud. e.g. CAPTCHA solving, fraudulent documentation, etc.
Individual + Company: Preventing generation of child porn. In the USA, this would likely be illegal both for the user and for the company.
You sound offended. My apologies. I had no intention whatsoever to offend anyone. Even if I am not diagnosed, I think I am at least borderline somewhere in the spectrum, and thought that would be a good way to ask people explain without assuming I can read between the lines.
I think ELI5 means that you simplify a complex issue so that even a small kid understands it. In this case there is no need to simplify anything, just explain what a term actually means without assuming reader understanding nuances of terms used. And I still do not quite get how ELIA can be considered hostile, but given the feedback, maybe I avoid it in the future.
Saying "explain like I have <specific disability>" is blatantly inappropriate. As a gauge: Would you say this to your coworkers? Giving a presentation? Would you say this in front of (a caretaker for) someone with Autism? Especially since Asperger's hasn't even been used in practice for, what, over a decade?
> In this case there is no need to simplify anything
I don't see how this is a response to anything I've said. They're speaking to other humans and the original use of their modified idiom isn't framed as if one were talking about their own, personal disability.
This. Amartya Sen has claimed that two actual democracies have never been at war between each other. At least I find hard to find significant counterexamples in history. (Not sure about the Falkland war. And I think Finland was technically at war with UK in the second world war)
I think the democratic and developed countries need to change their game plan pretty soon. The countries that are willing to join the club should be offered actual help to develop. By actual help I mean trade treaties that are designed to benefit those countries, not developed countries. Includes IP vaiwers, duties that protect local industries etc.
The countries that do not want to join, (including China and Orban's Hungary) then again, should be punished in all ways possible. Massive duties to commodities and other products imported from those countries, as a starter.
Open democracies do not need to be nice guys if they are threatened. Must not be, to be more precise. See Popper and paradox of tolerance.
Do you mean the war between Argentina and England?
Argentina was not only under a military dictatorship then, but the war was definitely triggered by the military junta.
> This. Amartya Sen has claimed that two actual democracies have never been at war between each other. At least I find hard to find significant counterexamples in history.
Off the top of my head -- a war of 1812 between Great Britain and United States. Both countries were democracies at that time.
> […] a war of 1812 between Great Britain and United States. Both countries were democracies at that time.
By what definition the British Empire, a constitutional monarchy, had been a democracy in 1812 if all (and not just the fourty-shilling freeholders) men aged 21+ with some women aged 30+ only became allowed to vote in 1918, 106 years later, and all women aged 21+ were finally allowed to vote in 1928?
In 1812, both Britain and the US lacked universal suffrage - not only were all women excluded from voting, but in both countries, so were many men - so if lack of universal suffrage made 1812 Britain not a democracy, the United States wasn’t one either.
But, historically, democracy was not considered to require universal suffrage. Athens is often cited as one of the world’s first democracies - and it is from Ancient Greece that we get the word - yet in ancient Athens, most adults couldn’t vote (either due to being female, due to being slaves, or due to being non-citizen resident aliens)
In 1812, Britain was a constitutional monarchy with a parliamentary system dominated by landed aristocracy and gentry – 3% circa of the population, and it was exclusively males.
Members of Parliament (MP's) were often elected in rotten and pocket boroughs where a single patron could (and unashamedly did) buy up the votes in one or multiple boroughs and essentially dictate the outcome of elections. Charles Howard, 11th Duke of Norfolk controlled 11 boroughs, as an example.
The rotten boroughs were only disenfranchised by Reform Act 1832 which granted seats in the House of Commons to large cities instead.
In the pocket boroughs, the landowner could evict electors who did not vote for the two men he wanted, the practice that did not cease to exist until secret ballots were introduced in 1872.
There were also «open boroughs» where the vote was more competitive but still limited to male property owners only.
Since «demos» in «democracy» stand for «people», calling Britain a democracy in 1812 is not even a stretch.
And in 2024, it still is a constitutional monarchy - how is that relevant to the question of whether it is or was a democracy?
> Since «demos» in «democracy» stand for «people»,
The meaning of words isn’t determined by sum of parts like that. The people who coined the word “democracy” used it to refer to systems in which only a small percentage of “the people” counted. Herodotus used the word in the 5th century BCE, whereas universal adult suffrage didn’t exist in most places until the 20th century, well over 2000 years later. If you want to argue that everyone was using a word incorrectly for over 2000 years, well, you can define words however you like, but other people aren’t obliged to accept your novel definition
I am not following it. On the one hand, valid points have been made that I have a wholehearted agreement with, yet the next moment there seems to be the gish galloping going on. Or that is how I perceive it, anyway.
> In 1812, Britain was a constitutional monarchy […]
[…] a parliamentary system dominated by landed aristocracy and gentry – 3% circa of the population […]
That was the key point that was conveniently or otherwise omitted. 3% of the entire population being landed aristocracy and gentry is not the rule of people by any measurable account, and, no, it does not constitute the British Empire a democracy in 1812. There was also a reason why I originally emphasied «Empire» – empires do not place a focal point on the power of people as there are no people in an empire, there is only the empire.
> The meaning of words isn’t determined by sum of parts like that.
Yes, the modern defintion of «democracy» is multifaceted and elastic, it has evolved to mean much more since its inception, including universal suffrage as a fundamental property of a functional democracy today. Yet, democracy in Ancient Greece is not what most would call a democracy today. I am accutely aware of that.
> […] other people aren’t obliged to accept your novel definition
This is a hard one since the guy managed to do his trick once Hungary already joined all clubs. On the other hand, there's little hope anybody can manage to overthrow him anytime soon, and in any case restoring democracy in Hungary will take decades.
I offer one more alternative. Double the tax every year, and once the ip holder decides not to pay the tax, IP is released to public domain. No compnay has money to keep IP indefinitely in such a scheme.
I've proposed this alternative on message forums like this for probably 20 years now, but specifically for copyright. I think it'd be a great way to let Disney and co. have the long copyright terms they want for extremely valuable properties, but get other stuff into the public domain much, much faster.
Instead of doubling every year, my proposal is 5-year terms: the first is free, then $1000 for the next 5, $10k for the next 5, $100k for the next 5, etc. Feel free to adjust the actual dollar amounts, but you get the idea. Most stuff would be in the public domain after 10 years, if not 5.
I would suggest only for corporations above a certain size. I would appreciate if some old music composer somewhere gets to continue being old and stuff without worrying about this sort of thing.
It would have to be managed by tracking the number of active patents. You get 100 active patents tax free. Over that, and you have to pay an annual fee. This allows for independent inventors to operate as the system intended while clamping down on NPEs.
What do you think will happen if IP becomes impossible to afford, as it surely will under such a policy? Do you think companies that value IP will bother investing further in R&D, let alone even stay in your country?
Congratulations on the massive net loss in taxable income in your country.
It'll only be impossible to hold for long periods of time. We can start the tax at $1, it hits $1mil after 20 years - and for 99.9999% of IP by year 20 it's either clearly worth zero or clearly worth >$1mil so it'll be an easy choice. It'll force things into the public domain faster and make it expensive to hold a big bullshit patent catalog, but for actively used properties it'll be fine.
Companies are going to want to sell in one of the richest markets in the world; they can either pay for IP protection or not be granted it.
edit: I'm not suggesting these exact $ values as clearly being the correct ones, it's just an example.
You do not seem to understand how patents work because of other comments thinking that you can patent something without publicly disclosing the invention. In and of itself, this comment is silly because the answer to your question "What do you think will happen if IP becomes impossible to afford, as it surely will under such a policy?" is "exactly what happens when patents expire currently." It appears you are unaware that companies still invest in R&D knowing they at most get a few decades of exclusive rights.
So, for example, the secret sauce that makes the CPLEX and Gurobi solvers tens or hundreds of times faster than open source equivalents should simply be released to the public, leading to the immediate loss of 90% of those products' competitive advantage?
You don't see how such a policy would spur terror among large, profitable companies with trade secrets, leading to them moving overseas?
In these threads it always end up being: "Sure Foo would be nice, but, I don't know if you are aware, we are actually held at gunpoint by the status quo; its all really a non starter when you consider this fact."
If this is the only real problem, than why not just let them go overseas? Let the market play it out? The boon of progress and freedom X country would get from becoming even little more rational about IP would pay for itself and be better for actual people.
> The boon of progress and freedom X country would get from becoming even little more rational about IP would pay for itself and be better for actual people.
That's what China has been doing for decades, and "gongkai" [1] is just one tiny part of it. While life for the average Chinese citizen has gone up - the CCP managed to lift hundreds of millions of people out of poverty - the life of most Western populations has gone down the drain as entire industries, entire towns were unable to cope with unfair competition.
The societal consequences of that will haunt us all for many years to come.
I don't think that the point was that there are no for loops in SQL. I think the point was that almost always using for loops is wrong and super inefficient and there is a much more efficient way to just use joins instead.
What does it matter? Yeah, at the very bottom of it all there will be a loop. I wouldn't attribute that to C but to the von Neumann architecture.
The point is that by getting rid of loops you remove one way of telling the computer "How" to do it. Start here, do it in this order, combine the data this way.
When "How" is a solved problem, it is a waste of your brain to think about that again. "What" is a better use of your brain cycles.
Exactly, SQL is like a very specific natural language used to tell a database what data you need. The database contains AI (the query optimizer) that will use statistics to determine the fastest way to retrieve the data requested. Depending on your query and the data you have the fastest way to get your data will change. It’s hard to manually write this yourself.
> What kind of applications would this be useful for? What can you build with an AI data science intern that's right 75% of the time?
I have written a bunch of more or less complicated SQL during my career. And I am pretty sure that if I need to write a SQL statement that's anything but select * from table, my output won't work 75% of time.
I may be special case, but typically if I work on a hard problem, it is not a single hard problem but a sh*tload of connected simple problems. If I can get someone to solve the simple problems 75% of the time correctly so that I can spend my time figuring out how those simple problems are to be connected, I'm ore than happy. And that's exactly how I use chatgpt. I have learned not to ask too complex questions from it. But the simple ones, it mostly aces and when it does not , they are easy to spot, as it is not that I could not have solved them myself, I just did not want to spend time for that. Now, if only the chatgpt was not almost as lazy as me to produce long simple stuff, that would be awesome.
I think of it with this kind of analogy: the original image is stored with 32 bit color scheme. You can reduce the color scheme to 16 bit accuracy and still figure out pretty well what the image is about. 2 bit is stretching this to a bit far, basically either pixel is white or it is black, but even if you lose lots of nuances in the image, in many images even that gives you some idea whats going on in the image.
I have never quite understood why there can't be stable stallites in dimensions above 3.
I mean, I know the argument that gravity inverse square law becomes inverse cube law in 4d, but what I do not understand is that what/why enforces that. Why in a hypothetical 4d world there just can not be a gravity-like force that is inverse square? Would that cause some kind of contradiction?
If you make a uniform flash of light in 3d space it spreads around you in a shape of a sphere of increasing radius. Energy gets distributed evenly across the sphere's surface which is growing with time proportional to a square of the radius. So energy density (think intensity) decreases as inverse square. In 4d space the sphere's surface grows with the cube of the radius.
This sort of intuition. Applies to electromagnetic waves, sound and gravity all alike.
But there are forces even in our 3d world that do not follow inverse square law (strong/weak nuclear forces). That kind of proves that all forces do not need to follow this intuition?
In particle physics, fundamental forces are generated by the exchange of virtual particles (I have been reading a quantum field theory textbook for the last few months to try to understand precisely what this means, among other questions, but this is accepted fact in the field). The Coulomb force comes from the exchange of photons. So this 1/r^2 argument for intensity leads to the Coulomb force falling off like 1/r^2.
This argument doesn't obviously apply to gravity (though presumably it would for a quantum theory of gravity), but the equations for gravity (general relativity) give the same result.
At a higher level, it turns out that when you try to combine quantum mechanics with special relativity, the resulting theories are highly constrained. It's not like classical mechanics, where you can just say 'suppose there's a 1/r^12 force.' You get mathematical inconsistencies if you stay too far. Weird stuff
All long-distance forces seem to follow it though. I think it's related to the energy conservation law.
It proves nothing of course. When we speak of these N+T universes, we try to imagine a system that follow the same "fundamental laws" but with different N and T. What exactly is fundamental is up to debate. You can even imagine a system that has different math, but it will be very hard to reason about it.
Those forces are also mediated by particles (called gluons, W, and Z bosons). But these particles are massive, charged and interacting, which caused them to behave quite differently and only act on short length scales.
