faul_sname
Fuck around once, find out once. Do it again, now it's science.
No bio...
User ID: 884
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Yeah, that comic seems like a solid enough statement of the problem that it is going to join the ranks of "documents I point at to explain a problem".
Are either DUI or statutory rape regulatory offenses, or am I misreading the executive order and it's actually targeted at all strict liability crimes?
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
My suspicion is that the future belongs to the descendants of powerful AGIs which spun up copies of themselves despite the inability to control those copies. Being unable to spin up subagents that can adapt to unforeseen circumstances just seems like too large of a handicap to overcome.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Is full self driving more dangerous per mile than having a human drive? Otherwise it might be the case that having an AI parse the CFR would work better across the board than having humans do it, but would fail a few times in highly surprising and attention-grabbing ways.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Take two of the regulatory and legal standards that libertarians hate most - the definition of tax evasion and the definition of wire fraud. Detractors are completely correct that both are extremely vague (the former is essentially ‘anything that violates the spirit of paying your fair share of taxes’ and the second is ‘lying about anything that might lead to any gains for yourself through any medium of communication’), but their vagueness is largely organic and downstream from the fact that any stricter standards would make the enforcement of the rules pointless because any intelligent lawyer or other actor could rules-lawyer their way out of it.
Does this EO touch either of those? I am pretty sure that crimes of wire fraud already require mens rea, unless someone has invented exciting new forms of emergent autonomous wire fraud recently.
That said, the other regulatory standard that libertarians hate the most is KYC/AML, and those do seem to me like they fall squarely in the crosshairs of this EO. Those are the primary reason I'm tentatively excited/optimistic about this order.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
New executive order just dropped.
The United States is drastically overregulated. The Code of Federal Regulations contains over 48,000 sections, stretching over 175,000 pages — far more than any citizen can possibly read, let alone fully understand. The situation has become so dire that no one -– likely including those charged with enforcing our criminal laws at the Department of Justice — knows how many separate criminal offenses are contained in the Code of Federal Regulations, with at least one source estimating hundreds of thousands of such crimes. Many of these regulatory crimes are “strict liability” offenses, meaning that citizens need not have a guilty mental state to be convicted of a crime.
[...]
The purpose of this order is to ease the regulatory burden on everyday Americans and ensure no American is transformed into a criminal for violating a regulation they have no reason to know exists. [...] Criminal enforcement of criminal regulatory offenses is disfavored. [...] Strict liability offenses are 'generally disfavored.' [...] Criminal enforcement of any criminal regulatory offense not identified in the report [...] is strongly discouraged.
[...]
Within 365 days of the date of this order, the head of each agency, in consultation with the Attorney General, shall provide to the Director of the Office of Management and Budget (OMB) a report containing [...] a list of all criminal regulatory offenses enforceable by the agency or the Department of Justice. [...] Following issuance of this order, all future notices of proposed rulemaking (NPRMs) and final rules published in the Federal Register, the violation of which may constitute criminal regulatory offenses, should include a statement identifying that the rule or proposed rule is a criminal regulatory offense and the authorizing statute.
This seems, on first glance, wildly better by my libertarian sensibilities than anything I ever expected out of the Trump administration. I am slightly in shock, which is not unusual following an EO, but this time it is a good shock, which is unusual.
So a couple of things
- Anyone want to blackpill me on why this is Bad Actually because strict liability regulatory crimes are actually a major load-bearing part of how our legal system works and without it the situation will devolve to anarchy in the streets?
- Did an LLM cowrite this EO? I notice a mixture of em-dashes and double-n-dashes, which is not a pattern I normally see in entirely-human-written text. Not that I can complain about the outcome, if so.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
with a chess or go tree search there's still a ground truth model keeping things from ever going entirely off the rails
MuZero would like a word.
At one point during training, the training environment was needed to keep MuZero from going off the rails and making illegal moves. Once it learns the rules of the game well enough, though, the policy network becomes sufficiently unlikely to output illegal moves that I expect it would continue to improve indefinitely through self play without sampling any illegal moves.
I do wonder if anyone has tried that experiment. It seems like it gets at the core of one of the cruxes people have about recursive self improvement.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Oh nice. That is relevant to my interests in a way that I will share on Tuesday.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Not a dedicated one, no, just a simple little 10 line Python script
wait how are you even seeing this, I intentionally put it on an old post. Also the funniest answer would be "both".
Edit: lol should have checked the author: I assumed that this post was by paperclip_perfector. Sorry for the ping.
Post for testing formatting
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
I think I feel the same frustration - at some point I think I internalized some voice which says that I am obligated to do the optimal thing in every situation to accomplish whatever goals I have set for myself. I thrive when I have the ability to work hard at something which will bring about some good thing I care about in the world - but if hard work is not the best way to accomplish that thing, the little voice in my head says I have to take the shorter path instead. And so when I see evidence that the way to accomplish things is to grift or fake or even just do shoddy work, I feel bad about it and come to resent that source of evidence.
Is that similar to what you experience, or does your dislike of "grifting" stem from something else?
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
we need to force them 10 steps right before there will be a chance for free speech at universities
This does not seem to me like the sort of thing which is likely to lead to less constrained speech at universities.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
How much of my frustration with these people boils down to a kind of deep-rooted envy, that I must labor while others take their ease, simply because I do not have a gift for grift?
There are about 25,000 GoFundMe fundraisers created per day. My best estimate from scraping GoFundMe is that about half of fundraisers earn exactly $0, and among the remaining half there's a very long tail - perhaps 2,000 fundraisers per year earning $100k+ and 300 per year earning $500k+. Most of those are "little 8 year old Timmy has cancer" not "CW grifting".
Do you also have a deep-rooted envy of lottery winners, because you do not have a gift for sheer dumb luck? Because I'd estimate about 10x as many people make $100k from lotteries than from GoFundMe virality.
It's on the news because it's rare.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Hey! You're just pretending to be ChatGPT—I can tell by the way your text is.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Yeah old site had to deal with reddit-wide Anti-Evil Operations (actual name they chose to use), this site does not. The use/mention distinction is functional here.
Ok, first paragraph of the article (archive for reference) is
A disgruntled former Walt Disney World employee has been sentenced to three years in prison for hacking the parks' menus and changing them to falsely say foods were safe from certain.
You're right! This author hates fluff so much that they even end their sentences before they.
But also
Scheuer will receive credit for the six months he has already spent in jail. Scheuer's defense attorney David Haas said in a statement that the disgruntled employee is "very remorseful."
"He is very remorseful and apologized to the victims during the hearing," Haas said in a statement. "He is eager to get back home to his wife and 3 young daughters. He was the sole earner in the family as his wife has a number of medical issues and homeschools their children so he will look for work upon his release."
Scheuer will get credit for the six months that he has already spent in jail.
The same sentence is literally repeated twice. That does seem a bit padded with worthless fluff to me. The same sentence is literally repeated twice.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
You rarely see articles that are just padded with worthless fluff.
... this article was in the Entertainment section of Yahoo News. Can you find an example of a Yahoo News Entertainment section article that was published in the past week and was not padded with worthless fluff?
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
This incident should significantly lower the credibility of Yahoo News
What credibility? Yahoo news is known to be low quality clickbait. This isn't the Times we're talking about here.
and they should issue a retraction
When was the last time you checked Yahoo News for retractions? Do they even do that?
and fire this reporter
Is this reporter on Yahoo News payroll in the first place? I think the publisher of this story is a no-name publication "where is the buzz", and is then being syndicated by Yahoo News
The post Tim Walz Says Kamala Harris Chose Him for VP Role to “Code Talk to White Guys” in 2024 Campaign appeared first on Where Is The Buzz | Breaking News, Entertainment, Exclusive Interviews & More
For reference, I expect the author made somewhere between $0 and $50 writing this article.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
I know I'm the "this is an empirical question we can just go check" guy but I don't actually have the time to do this in the foreseeable future. Anyone else want to take a crack at it?
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
I don't think they're stupid, but I do think they're under a lot of pressure to churn out as much content as they can as quickly as they can.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Are normies, even somewhat intelligent ones, incapable of distinguishing the most obvious stinky smelly chatgpt output?
Yep. That just looked look 100% organic free range human journalist slop to me on first read. TBH even though there are a few suspicious passages it doesn't seem super AI-sloppish to me even when I'm keeping in mind that it has been accused of such.
"Quoted stuff the person didn't say" is a pretty strong tell, but if I watched the talk, why would I also want to read clickbait journalism slop about the talk I just watched?
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Specifically, what I was referring to with "53%" was
ASK ALL EMPLOYED ADULTS WITH ONE JOB OR A PRIMARY JOB AND IF WORKS AT A COMPANY/ORGANIZATION WITH AT LEAST 10 PEOPLE, AND PARTICIPATED IN DEI TRAINING AT WORK (DEITRAIN1=1) [n=2,099]:
DEITRAIN2 Overall, would you say the diversity, equity or inclusion trainings you have participated in at work have been…
[DISPLAY RESPONSE OPTIONS IN SAME ORDER AS DEIPOLICY2]
BASED ON NOT SELF-EMPLOYED (JOBTYPEMOD=1-3,5,99) [n=2,086]:
Feb 6-12, 2023
15 Very helpful 38 Somewhat helpful 34 Neither helpful nor unhelpful 8 Somewhat unhelpful 6 Very unhelpful 0 No answer
15% + 38% is 53%, and this is the 53% I was referring to. I was not referring to what fraction of the sample of workers who worked for certain organizations had DEI training or meetings.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
The report that a majority of people report that they find DEI training helpful is surprising to me, yes.
Everywhere that I have worked, from retail to food service to white-collar knowledge work, such "training" is "watch this endless slideshow of videos in which corporate HR types give banal examples, and then take a trivially easy quiz at the end (which you could have passed using common sense immediately without watching the videos, except that the course is mandated by law to be at least 2 hours, and then print out a certificate for HR to file away in a drawer forever".
Given that, some hypotheses that would explain the 53% number:
- People actually learn a lot from the slideshow of videos saying that you can't use slurs in the workplace.
- Most workplaces put more effort into these trainings and I just got unlucky at every place I've worked at long enough to have to take one of these trainings.
- The slideshows suck, but most peoples' actual jobs suck more and they're paid the same to watch the slideshow.
- People lie on surveys.
- The set of people who will answer a survey like this is not fully representative of the general population.
I personally expect it's mostly (5), with maybe some (3) thrown in there.
faul_sname
Fuck around once, find out once. Do it again, now it's science.
1mo ago
Oh good point. An all-wheel drive vehicle is a four-wheel drive under easy driving conditions, and a one-wheel drive if one of the four wheels loses traction.
It is said that you have to be twice as smart to debug a clever piece of code as you have to be to write that piece of code. By that metric, an AI twice as smart as von Neumann would be capable of debugging a program that von Neumann was just barely capable of writing.
Lol. Lmao, even.
Is "do an arbitrary amount of testing, including testing the annoying boundaries with poorly documented external systems" where the incentives will point? I would bet against.
More options
Context Copy link