While I think the scale of American decline is overstated, I think there is a degree of Hemingway's law of motion.
A desired task that requires the most skilled makes those skilled people in demand. If a power has significantly more resources then they have more to offer those most skilled people.
It isn't at all disputed that there are a huge number of scientific discoveries that have occurred in The USA by people born outside the USA. That shows the draw of that power, but it is a relative draw. As the ratio becomes smaller the draw is less.
Advances like this are a feedback mechanism, being ahead gives you more resources to stay ahead.
If you consider the average contribution of advances to be a relatively steady force advancing a nation, yet a nation is in decline, it stands to reason that the decline is in another area and is being mitigated by the advances.
If the force propping things up goes somewhere else the change can be quite swift because the force of the decline becomes suddenly much more apparent.
I don't see the world going full Mad Max, but I can certainly see a sudden shift to the USA being considered no different to the UK,Japan, or Germany.
“America is in decline” is the consensus view. America dominating on all fronts is the contrarian view. I expect these views to swing like a pendulum in public discussion until something meaningful happens or until it’s clear in the rearview that America is in fact is more like UK, Japan, or Germany.
Certainly that's the consensus view. However, I have yet to see any evidence that people within the consensus actually have done any analysis to find the truth. They just have sort of feeling which is driven by the press which presents a story.
Whenever i sit down to read research, I remind myself of Lockheed Martin reading the USSR published research[0] on how electromagnetic waves scatter off of surfaces, and using that to fuel the initial stealth technology. The leading theory being that the USSR didn't recognize how brilliant and revolutionary ability these calculations were.
Just because I can't see the immediate brilliance, doesn't mean it is not brilliant in it's own right.
There is a similar story with the discovery of buckey-balls. A researcher at University of Houston had data that demonstrated buckey-balls were created, but he didn't fully understand what he was looking at. Then a researcher at Rice saw the data and recognized c-60 was being created, so he bought the data and the process and then "invented" carbon balls
I have taken almost the opposite opinion, but with an important caveat.
It applies to arguments in general, and increasingly there seems to be fewer and fewer 'pure' technical issues.
I have observed a proliferation of people believing things that are simply not true. Much of this comes from people stating unproven or undecided factors as absolute fact, and then building an argument on those foundations.
The caveat is that I think you have to remain civil, be meticulous at addressing the argument, and to never assume that you know the hidden state of another person's mind.
This isn't about winning arguments, it is about balancing them. This is well established on a court of law. A decision decided after a claim has been robustly challenged is held to be a more objective decision.
I don't feel like my part is to push a narrative forward, but to assist in stemming the tide of absolute ideology. I think the ideas themselves do have the capability to advance on merit, but not if they come under sustained attack.
I think a lot of people have given up on arguing, leading to the voices of only the most motivated becoming dominant, which in-turn, advances the more extreme positions that drive their motivations.
I think, perhaps in such an environment, Andrew Wakefield could have elevated his claims to be a majority opinion, he convinced a remarkable percentage as it was.
If unchallenged ideas becomes majority opinions it becomes very difficult to unseat them. The claim that most people believe a thing is enough to assert it's truth is pervasive.
The insideous thing is how many of these things have gotten through, what falsehoods do we believe that go unchallenged now because everyone believes them. You can't really tell yourself because you as part of the population likely believe it too.
I would characterise myself as being very left. Yet people who I perceive as to the right of me perceive themselves as much further left.
I have always considered compassion the driving force underlying left wing views. I really can't understand the mind of the spiteful left, it seems such a contradiction of values.
I think this is broadly similar to a thing I built as a Proof of Concept a while
( eek youtube video is 11 years old, time does fly) ago. It had a server that only listened on localhost and required a reverse proxy to make it available via https. Numerous one line commands could do that job.
Going to a native, but still browser-ish, client might simplify it somewhat as a ssh rather than https program, electron didn't exist when I started on this thing though.
If you go to the simple tick demo around 7 minutes in
https://youtu.be/7namj7iy16Y?t=433
It shows a minimal node app running and connecting to the socket indicated by process.env.WEBSESSION to open a window on the client and sends it the webpage to handle it's own output.
I have been recently revisiting some of the ideas here using web technologies that have been created since (using promises, web-components for the window). At the moment I'm doing the whole thing client side, which actually makes it a completely different beast. I think both browser hosted backend, or real machine hosted backend have merit, but somewhat incompatible. I'm still pondering how to reconcile this.
The entirely browser side means you can host a command line environment on neocities https://lerc.neocities.org/ (has a bug where you need to reload the page once to get it working, but then it's good) It is also very much just proof-of-concept.
If you try the client thing out on neocities. Some command suggestions that reveal some of the subtleties.
foo
bar
ls -al /bin
view /res/image/slice8.webp
cat terminal.html |hd
> I look forward to your weather report too: "It's always sunny outside until one day it starts raining. Every time."
I once ran across the comment that if you simply predict tomorrow's weather will be the same as today's you'd be correct 80% of the time. Not sure how true that is (can't find the source).
There are momentum indices. Momentum is actually a strange phenomenon from the perspective of the efficient market hypothesis because it does not have an obvious risk to balance its premium compared to other empirical factors.
Not at all, if someone tells me that "This stock is historically likely to regress to and beyond the mean," it's information I can use to evaluate my risk tolerance. Just because a piece of information doesn't let you time the market like a psychic doesn't make it worthless, it's just not what you were looking for.
Productive workers in society need to learn new things all the time. I can't think of any career that hasn't changed in my life. I recall a garbage man (sexism probably wasn't required even then, but I never recall females) hanging off the back of the truck while the driver drove to the next house - the driver today needs to know how to operate the arm on the truck that lifts my can. Fast food used to be cooked within 10 minutes of when it was thrown, now they obviously are keeping things warm for a lot longer.
As a counterpoint to that, I encountered a conversation where people were lamenting the toxic nature of communication and someone described being told to kill themselves for expressing what they felt was a compassionate statement.
Someone asked where that happened and they said "On X" and the response was "Holy shit, That's the kind of thing you expect to see on Bluesky, not X"
The thing is, The comments were terrible, and the average user of either platform would probably wholeheartedly agree that they were terrible.
If you exist in your own little community on these platforms then you don't see those bits. Those hideous extreme elements are there though. I don't know how representative they are of their respective populations, or even how much of it is automated stirring. I'm not sure anyone does. it seems quite difficult to find an analysis that is not pushing an agenda. The nature of agenda driven research over truth driven research makes it much easier to find the agenda driven stuff, because it's only reason to exist is to be found. The hard working people who try and find the nuance are too busy doing that to run a PR operation for their work.
There's a dark irony that with the decline of platforms like Twitter and Reddit descending into places of astroturf and brigading, there are fewer places to find conversations where informed people are discussing things publicly. A person searching for what an informed individual would say on the matter cannot find it. There's not really even any bots pretending to be those informed individuals. The bot game is more basic. Throw so much obviously fake crap around that nobody trusts anything.
This is older than the public internet. My parents got a CB radio circa 1967 and I watched my Mom make the first transmission on the new radio. She promptly received a reply, which told her to take a long walk on a short pier.
>My point is, deterministic logic matters in certain circumstances 100% of the time. Forcing the LLM to make something unlikely is not good enough because a series of mistakes could very quickly bankrupt the company.
If your argument is that the danger of equal values being selected inconsistently breaks determinism, that's a trivial problem to solve.
Any non-infinite precision numbering system by definition is at the limits of it's precision when equal values occur. If you need to order such values you can extend the precision and add on a deterministically unique tiny value (position, order encountered, etc.) . Your original value stays in the same precision range but they are now unique.
It's usually more likely that you want to sacrifice a little precision for determinism so you can quantise to allocate the range where you apply the unique ID
For example if you had an array of 256 fp32 values but you required them to be unique, you can lop off 8 bits of mantissa and replace it with its index in the array, Every value is then unique.
Granted token dictionaries make for some fairly hefty indexes now, but the principle applies in general, it's easily solvable if you are prepared to spend some precision or do some extra calculation.
> Wouldn't a one-on-one interview be a better way to establish a broad evaluation of competency.
1) Possibly, depending on the material, but most classes aren't looking for a "broad evaluation of competency", they're looking for the specific material taught. It'd make more sense as a graduation requirement: the equivalent of a dissertation but for undergraduate work.
2) Even in small classes (e.g. 20-30 students) that wouldn't scale, let alone the massive courses earlier in a curriculum or that are shared by many degrees (e.g. hundreds of students).
Maybe not everything needs to "scale"? You're asking each student to spend well over 100 hours in lectures and study for each class. Surely you can find 20 or 30 minutes per student for evaluation.
Please by all means direct more funding at universities, hire several times as many professors, and have them each teach half as many classes with a fraction of the students. The result would absolutely be an improvement.
(That said, any kind of subjective assessment has its own pile of hazards compared to objective assessments.)
A problem with not scaling here is that just adding multiple examiners means that now there will be differences in how students are being graded depending on which examiner they get.
There are so many of these projects to wrangle AIs I think we might need an AI to go through, analysing each and amalgamating the good bits.
It makes me think of MakeFiles.
Make is sufficiently bad that everyone who has used it has considered writing a better way to do it. A good percentage of those people have done so.
On the other hand, make is also not so sufficiently bad that it cannot do its job. The choice becomes picking the thing that everyone has or one of the many many alternatives that proclaim their strengths and leave their weaknesses lurking to bite when they are least expected.
No single replacement to make dominates, and make lives on. I wonder if AI management is on a similar path.
A desired task that requires the most skilled makes those skilled people in demand. If a power has significantly more resources then they have more to offer those most skilled people.
It isn't at all disputed that there are a huge number of scientific discoveries that have occurred in The USA by people born outside the USA. That shows the draw of that power, but it is a relative draw. As the ratio becomes smaller the draw is less.
Advances like this are a feedback mechanism, being ahead gives you more resources to stay ahead.
If you consider the average contribution of advances to be a relatively steady force advancing a nation, yet a nation is in decline, it stands to reason that the decline is in another area and is being mitigated by the advances.
If the force propping things up goes somewhere else the change can be quite swift because the force of the decline becomes suddenly much more apparent.
I don't see the world going full Mad Max, but I can certainly see a sudden shift to the USA being considered no different to the UK,Japan, or Germany.
reply