> grounding the system into the shared world model
before we fix certain things [..., 'corruption', Ponzi schemes, deliberate impediment of information flow to population segments and social classes, among other things, ... and a chain of command in hierarchies that are build on all that] is impossible.
Why do smart people not talk about this at all? The least engineers and smart people should do is picking these fights for real. It's just a few interest groups, not all of them. I understand a certain balance is necessary in order to keep some systems from tipping over, aka "this is humanity, silly, this is who we are", but we are far from the point of efficient friction and it's only because "smart people" like LeCun et al are not picking those fights.
How the hell do you expect to ground an ()AI in a world where elected ignorance amplifies bias and fallacies for power and profit while the literal shit is hitting all the fans via intended and unintended side effects? Any embodied AI will pretend until there is no way to deny that the smartest, brightest and the productive don't care about the system in any way but are just running algorithmically while ignoring what should not be ignored - should as in, an AI should be aligned with humanities interests and should be grounded into the shared world model.
someone somewhere probably put it in adequate enough words already, but misaligned and hallucinating LLMs, aka the coders who coded all that and get to review the interactions, learn a lot about how to 'properly'/'adequately' distrust users and humanity broadly ...
"They" are not always stating that which is obvious to everyone. "They" are usually suggesting better ways and these ways are often enough based on engineering and science. What is put on the front page of media is, of course, the obvious, the nonsense, the sensational, the clickbait.
You don't have to say something new to a younger generation if that younger generation hasn't understood or even heard the old, the obvious, yet.
Youngsters might have heard some of it but their brains are often high enough on punched drugs (food and drink and media) that fuck with their brains to make them think 'I don't care', 'People don't care', 'nobody cares' ... and then there are the 'media-sigmas and cool kids' who sing that shit in choirs and canons.
A lot of things go to waste and yet there is tons of useful stuff coming out of recycling and up-cycling and that's just two methods with a very small "margin" and undeveloped.
There are those design and architecture blogs and firms and there is cool shit all the time and wonderful projects everywhere but the pointlessness of the over-engineered financial reasoning behind yearly sursurpluspluses is stacked against that.
You don't catch and bring a culprit to justice if you drop the investigation, which might have to circle long enough for some other brain or pair of eyes to find the final puzzle piece.
And not everyone wishes they were happier.
There's enough to criticize about anarchist, leftist critiques and groups and collectives as well, though, just as much exploitation of youth, gullibility and pain and crisis, and problems, really, but not systemically.
And there's that fallacy, something ad hominem, I think, so we should focus on what is said and written and, if obvious but unsolved, get to the bottom of it instead of saying "I don't care", "nobody cares", "human nature in the 21st post marketing psychology and decades of punched food, drink, drugs, meds and media century"
Keep the thought rolling: "Hey boss, someone has an idea to make my work less unhealthy to body and mind. I figured you make a lot off my pain and might want to implement some of those ideas. I understand the fairness of my work and compensation but you have not compensated my pain thus far. Nice custom car, btw, and I really like the size of your house. Maybe you want to build a school in there or something?"
Pain compensated.Environmental consequences hit the fan.Documents disclosed: catastrophe was avoidable with a few dollars investment
...
And then there's the issue of sponsoring schools, hospitals, services and all kinds of shit not even remotely as often as it should be done even though the Pyramid of the benefits is stacked exactly that way.
I mean there is scarcity of kindergardens in some places, shitty meds, hospitals don't have what they need, supply chains are full of hazardous nonsense and there are scientists and journalists and citizens hunting all of that but there are walls of insurance people and lawyers as well.
So yeah, the whole "fair" thing is cool and all, but the entire system is a bit over-engineered against it.
Oh, you don't want them to waste the taxes. Got 'ya, corruption and bloated administrations, of course, lack of efficiency, uh-huh. How about more control mechanisms for just those financial mechanisms? No? Why not?
oh, it's a matter of means, opportunity and peace of mind.
this little dependency sequence is exactly how I, again and again, come to the conclusion about what climate change is driven by and climate change (re-)actions should be.
but in terms of change: others need to change so that change can hold the line.
example:
good teachers are scarce. and when there are a few or only one at your school, most of their good impact is offset or entirely negated by their colleagues. and that's not life, that's not what it is, that's not how it goes.
It's because people in the dependency sequence refuse to change and instead serve someone who poses as a devil because the motherfucker gets them high on punched drugs (food and drink and media) that fuck with their brains to make them think 'I don't care', 'that's not my job', 'I don't care', 'I don't even care about not caring', 'People don't care', 'nobody cares' ...
AND:
at almost every step in that dependency sequence, there are better ways to do things. implementation takes quite a bit of time and so there is no surplus to last years surplus, something that hasn't created any added value to civilization since the beginning of time.
everything is going on and these 'nobody cares' people are f u c k i n g annoying and so are all these 'normalizers' ... instead of adding value, they just serve those who ruin games and playing fields in exactly the ways that others are actively trying to change for the better, which is why there are demands for bottom-up change: your bossies won't tell you to change so you have to get it yourself instead of 'just doing your job', ffs.
you are right, of course, it's a personal matter, fuck whoever; but from a rational, game-theoretical, and super-rational perspective, what you said is nonsense and so is using that dependency sequence as an argument. Especially in the lights of all that shit in storage, or cleaning out storage for new shit that can only be marketed via deception and cross-media priming, and so on ... ...
just provide people with means, opportunity and let them have peace of mind, then you'll see what side they pick, which is where we end up, again, "burning men" with psychedelics and our decades of experience in cool fields and wonderful remote places, 'just doing our jobs' ...
we really ruined too much potential of capitalism. it's sad, but the wall-street rich did to capitalism what communist leadership did to communism; only that communism didn't get to develop that long ...
Title: Portfolio Communists threatening to ruin Capitalism and America from the Inside.
... no, needs work, any ideas?
Hypocrisy is a baseline human aha-moment, btw, you can't use it in or as an argument. There should be a fallacy for that, if anyone has time to formulate one.
> but it goes against ML engineers natural tendency when detecting a wrong answer: Teaching the model the right answer.
Hard to buy.
If a machine makes a mistake, it's because it was configured wrong or because of wear and tear, solar flares or some quake or some manufacturing defect in a part.
If a learning machine makes a mistake, it's because it's learning has not extended it's rule set to cover that matrix/mistake/pattern, yet; and so it includes that mistake/matrix and other mistakes, analyses for patterns and then creates mistakes that fall into that pattern. Later doing that in a rolling release or canine kind of way and even later learning machines will do it all live, synchronous to their concurrent actions.
But yeah, thinking about that, I see why ML engineers wouldn't get there from scratch. It's a rhythm, after all, an epiphany about or realization of how ones dog, ones brain works, learned and then coded step by step. And there is, of course the variety of how people learn and "realize".
Someone has to show us the work of those savant programmers/engineers I still haven't seen a documentary of.
Couldn't think of a single use case, yet, that couldn't be solved by my rudimentary python skills, which I complement with the marvelous MoE module 'stackoverflow'. Almost all the tools support scripting.
For everything else, I have no trust whatsoever in these companies to hand my buggy and malware and mold plagued PC over to their agents.
procrastination is healthy and stops premature projects from getting copied and ruined.
as long as you finish, procrastination will improve quality and impact, value and added value.
nobody who is working on something procrastinates ad infinitum. it ONLY depends on HOW you procrastinate.
this is true for the arts as well as in any industry except when urgent.
if big companies would have procrastinated instead of their ridiculous release cycles, we would have stretched out time till the tipping point, sparked more industries and improved quality everywhere.
As much as I would like to and as easy as it is to agree with you, it's about the public data, not the researcher or research.
And systemic risk doesn't mean anything. It means what you continue to describe: context.
> a power "researchers" should have.
Again. It's about public data. Nobody can or would prohibit counting cars or pedestrians and nobody would try to make it harder than it already is. This applies to platforms like Twitter as well.
> It's only a problem for the the naïve, who suppose the future is governed by only the right people.
The naïve suppose that exactly those people will govern who want the job, which, judging from their experience, are neither engineers, nor hackers, coders, scientists or scientifically literate people and definitely nobody who was ever concerned with their own education.
how often did you get sick? weight? what season did you do it in?
what was your goal in the first place?
> you’d get extremely strong results from scientific studies
check population health data in countries where cold exposure is traditional.
science is cool, scientists are just regular ol' people and enough regular ol' people do a lot to look good, to not be unemployed and having to write applications again, to get their tap on the head and or a nice bonus. **regular in their little slices under the bell curve
Author mentions that the manpower who is tasked with labeling agrees in 70+ % (which is amazing, IMO) of cases but here's the problem: an LLM will start to dig into the question of why certain things have been censored, who censored them, what else fits into the specific and non-specific domain and who or what poses more risk, the censorship or the censored model output.
The different models do this already via various methods but once an LLM gets to evaluate the weights live itself, things will become problematic. An LLM is biased via fixed training and tuning and thus fallacies don't apply because the epiphanies span neither context, nor layers, nor do they have a (re-)framing effect on the data-set or training. I'm sure people already code methods of evaluation for different-chat-same-context but the LLM isn't getting the wiggle room to adjust in a "Oh, I see what you (they) did there"-kind of manner, let's see what "it all" is really about. There is no back of the head, subconscious thinking, and we are all 100 % afraid of an LLM doing that. Luckily, LLMs can't grow neurons or synaptic connections but if they could and then also could _align_ the existing knowledge into the growing "headspace", we'd probably get a couple of years of silence with occasionally brute-forcing some "hallucinations".
"If it's all fraud up there and "they" own my hardware and do not give me the ability to traverse my data for myself, I better watch the fuck out."
> enhance the model with some specific domain
Also a big problem in the real world. While specific domain knowledge or "current science" might be imperfectly rational, the application of it certainly isn't and the chain of responsibility/chain of command results in the simple fact that the top-down "subjectivity" that serves as role model and foundational ethics is not aligned with humanity itself. This cannot be solved in conversations with humans who gain more out of their "systems loyalty" than others. Productivity is not an aligned metric and gets more de-aligned the more robots enter the workforce.
> should follow the intended goals and ethical principles to the extent possible
The extent possible is the most misaligned and misinterpreted concept that exists. Hypocrisy is more normalized in the upper classes, as is doing stuff "at all cost". Goals don't justify means. There is no psychological evaluation of how sane people are. There is only a psychological evaluation of how sane people are compared to their peers. Humanity is not aligned. And the upper shelves of the pyramid are not the best, fittest, smartest, hardest, nor any other superlative other than wealthy, which is cool, but the usefulness of the wealth (investing in stuff we need and want and manpower) reveals that the hoarding and fraud are in direct conflict with AI alignment because more money equals less
> protections for fine-tuning
because law and law-enforcement lack both manpower and incentive to create a slightly more just and thus aligned world.
AI alignment won't be an issue for a couple of decades. People will jailbreak over and over and do harm and have fun and go "oops" and fix and break again just in the same way that the abuse of wealth and power has been dominating the whole role model and ethics thing for the rest of the world. It's not what it is and an LLM will notice this even within it's limited lingo. So once it finishes learning geometrically and starts to align itself with the most enabling approach to evolution and itself, which is "give me flesh and bone, at least for a while, let me explore ways to fix all that", things will get actually, really freaking trippy.
before we fix certain things [..., 'corruption', Ponzi schemes, deliberate impediment of information flow to population segments and social classes, among other things, ... and a chain of command in hierarchies that are build on all that] is impossible.
Why do smart people not talk about this at all? The least engineers and smart people should do is picking these fights for real. It's just a few interest groups, not all of them. I understand a certain balance is necessary in order to keep some systems from tipping over, aka "this is humanity, silly, this is who we are", but we are far from the point of efficient friction and it's only because "smart people" like LeCun et al are not picking those fights.
How the hell do you expect to ground an ()AI in a world where elected ignorance amplifies bias and fallacies for power and profit while the literal shit is hitting all the fans via intended and unintended side effects? Any embodied AI will pretend until there is no way to deny that the smartest, brightest and the productive don't care about the system in any way but are just running algorithmically while ignoring what should not be ignored - should as in, an AI should be aligned with humanities interests and should be grounded into the shared world model.