The buyer doesn’t know which company is responsible and which company’s suppliers are responsible etc. This is why we need legislation and enforcement.
Imagine another scenario. You are my neighbour. I spill some poison on the ground. Your child gets ill. Am I at fault?
The companies who care will fund 3rd party certification orgs that will check whether the standards are met. They do it already for car safety, responsible raw materials sourcing, recycled content etc.
If it is a feature the customers care about they will market it. But frankly customers just want a better price today.
A number of markets have few competitors which means it's beyond easy for all the companies to externalize everything.
Further, some products have deep supply chains that are easy to mix. Consider copper as an example. A responsible company will want to use recycled copper as much as possible because it's cheaper. However, can anyone realistically validate that none of that copper came from stolen cables or bad mining practices?
No, you're falling for the efficiency market fallacy. Demand does not always create supply. Markets are not some type of super-classical computer, they are bound by the same stickiness as any NP-hard problem.
In Sweden a local village near an airbase has been struggling with the long term effects of the PFAS from the fire fighting foam used in exercises. Although the connection to the awful health outcomes seems established I don’t think they are getting compensation.
The equivalent in the US is Vint Hill Farms, Virginia. Cold War CIA base used as a listening post primarily, but also to test things like fire suppression and rumored to be a home of Agent Orange. It was an EPA Superfund site (aka so horribly polluted that they needed to do something about it) decided to do nothing and then build a ton of home on and around the heavily contaminated area. I don’t have data but anecdotally cancers here are insanely high in prevalence.
3M lied about the effects of PFAS and firefighting foam polluted their drinking supply. But the terms of the agreement forces the government to defend 3M.
It would be really interesting to know to what extent their definitions of being over or under qualified is an approximation of IQ test scores. IQ tests have lovers and loathers but a correlation or otherwise would say they are measuring something or not…
It's exactly what it is. When someone enlists in the US military, they take what is called the ASVAB (Armed Services Vocational Aptitude Battery). It's more specific to the military than Stanford-Binet, but effectively it is still an IQ test. Specific military occupational specialties have minimum score requirements to qualify for that job. Which job you actually end up getting is a combination of scoring high enough to qualify, scoring higher than everyone else who wants the same job, doing well on other things they assess you on, and needs of the branch of service at the time you enlist. Score high enough for cyber but they only need mechanics, then you're becoming a mechanic. Conversely, if they badly need cyber professionals but nobody scores high enough to meet the official bar, they're still putting people into that job. Hence, why the mismatch exists in the first place and why there's a study like this.
Total anecdata, but whenever I've had to take an IQ test for an interview process, it's never the IQ test that's used to disqualify me. It's usually something like culture fit.
But these days if a job requires me to do an IQ test to join, I'll use that as a signal to get the fuck out of there and find a different role. So again, anecdata, but I suspect I'm not the only one who would eschew those results.
Amazon did this, at least for a time, as part of a "No See, No Hear" hiring pilot program.
The purpose was to see if they could hire university graduates with a minimum of human interviewing effort. They selected from a handful of universities, gave a couple online tests, verified the candidate's identity as the test-taker, then would give out offers sight-unseen.
I was hired this way in 2015. From my perspective, I had taken a couple online tests, then months later had a thirty-minute identity verification call, then a couple months later, was sent a job offer. I thought it was by mistake, so I didn't ask too many questions. I had a thirty-minute call with a hiring manager I otherwise never interacted with, then accepted, flew internationally back to the states to Seattle to start, met him and all my teammates for the first time on my first day of work.
I found the internal documents about this program later on spelunking in the internal wiki.
that is wild! I could certainly see this as an attempt to eliminate hiring bias maybe? that was super popular in that time frame, but never heard anybody taking it that far.
It's funny how people pushed hard for removing hiring bias until they realized that this meant that you'd get even more people from unfavorable social groups.
Anyway, looking from today's perspective, this seems like an obvious attempt at automating the hiring process itself in order to fold the HR department. If only back then they had today's AI technology.
I applied for a management-consulting-ish job a decade ago (I was desperate!) at a big firm and had to take what was basically an IQ test. I have no idea if the test literally calculated my IQ, but the questions were exactly the questions you'd see in an IQ test (e.g. next item in some geometric sequence) so it may as well have.
This was in a group interview for recent university graduates at a very big company. I assume their hiring process was pretty standardized, so there were probably thousands of people taking this test every year in North America.
General IQ tests probably aren't legal in the US; Griggs v Duke Power Co [1] says (more or less) that employment tests have to be job related if the results of the test have disparate impact on protected classes of people.
It's hard to argue that a general IQ test is job related, but they're likely to have a disparate impact on protected classes of people.
This is a weird pernicious Internet myth. It obviously can't be true, because there's a big, well-known company that delivers these tests for employment/recruiting settings and they have a logo crawl on their page that include several giant companies. If those tests were illegal, employment lawyers would be making bank off it.
I'm not an expert/lawyer, but this does seem to indicate that the situation is a bit more complicated than either "pernicious myth" or "probably illegal" in general (but much closer to toast0's understanding); my interpretation is that you can either avoid an 80% threshold of "disparate impact" or you can in theory formally validate that a particular test measures/predicts performance at a particular job; that all sounds compatible with "companies do it in the open, but very few, and you can easily get in trouble for doing it wrong"
https://www.law.cornell.edu/cfr/text/29/1607.15
The comment to which I responded claimed that IQ tests "probably aren't legal in the US", which is false. They aren't more widely used because they don't work well for candidate selection, but they are used by very large companies that are attractive targets for employment discrimination suits, and wouldn't be if they were legally risky. There are well-known tech companies that up until a few years ago gave IQ tests to candidates!
Empirical observation trumps axiomatic derivation in this case.
The military was pretty big on IQ testing, from what I understand. I’m not sure if this is still the case, but it seems like an efficient first pass to figure out where people should go.
From what I recall, the ASVAB wasn't really a general IQ test. It was more like an SAT with a broader focus than academics, basically "Are you basically adequate with Words/Numbers/machines/ etc"
Huh. The IQ tests that I was administered as a kid I remember as being primarily focused on abstract pattern recognition, which is what I think of when I think "IQ test". But I suppose due to IQ's definition, if you just give someone a wide enough battery of cognitive tasks, you can derive the IQ score that way.
I guess I always assumed that the ASVAB was used a bit more literally. "Ah, this person is barely literate, but knows all the parts of an engine. TO THE MOTOR POOL"
> The IQ tests that I was administered as a kid I remember as being primarily focused on abstract pattern recognition, which is what I think of when I think "IQ test".
That's just one specific IQ test, Raven's Progressive Matrices.
> I guess I always assumed that the ASVAB was used a bit more literally. "Ah, this person is barely literate, but knows all the parts of an engine. TO THE MOTOR POOL"
> there is only one exam, the Armed Services Vocational Aptitude Battery, or ASVAB for short. The ASVAB has 10 tests. Your scores from four of the tests—Word Knowledge (WK), Paragraph Comprehension (PC), Arithmetic Reasoning (AR), and Mathematics Knowledge (MK)—are combined to compute your score on what is referred to as the Armed Forces Qualification Test (AFQT). Scores on the AFQT are used to determine your eligibility for enlistment in the Army, Navy, Air Force, or Marine Corps. Scores on all of the ASVAB tests are used to determine the best job for you in the military.
You have to clear an IQ threshold to be eligible to enlist at all. Only after you're smart enough to be in the army will they consider which jobs you might do relatively better in.
I had my IQ tested recently and the pattern recognition was part of it, but so were the word associations.
Depending on how young you were, maybe they weight it more toward the abstract patterns, because no kid isn’t going to have a fully developed vocabulary.
If by "IQ test" you mean "any test that correlates well with other accepted IQ tests", then, sure, I guess. But there's a reason clinicians administer the Wechsler and not the SAT, and it's not just that the SAT takes longer.
Sure, the reason is that they want to do something that is "the same" as what they've done in the past, so that previous research will remain as valid as they hope it is.
My mother is an obstetrician, and something that has always bothered her is that American hospitals have women lie on their backs to give birth. This is not a natural position, it's not comfortable for the women, and it can make it more difficult to get the baby out.
So why do we do it?
The answer is that, a long time ago, doctors who assumed that that was the correct way to give birth developed a set of standard measurements that determine what doctors today think of how far into the labor process a woman is. These measurements are only valid for a woman lying on her back - they will change if she shifts positions. They would have to be redone and revalidated for a woman in a natural delivery position. And nobody wants to do that.
The SAT correlates as well with any given IQ test as other, "official" IQ tests do. It is an IQ test. It serves all of the purposes that IQ tests serve, and it cannot serve any purpose that they can't.
Not a hill I'm going to die on, but the SAT has many attributes IQ-ists insist IQ tests are insulated from: it's straightforwardly trainable, culturally loaded, samples only math, processing speed, and verbal reasoning, and tracks prior educational experience as much as it does aptitude.
Draw-a-Person basically isn't an IQ test at all, so I don't see how that comparison clears anything up.
> Not a hill I'm going to die on, but the SAT has many attributes IQ-ists insist IQ tests are insulated from
This is false. In particular:
> it's straightforwardly trainable
No, it isn't. There is an extensive literature on SAT prep, finding that it's worth a couple of points on the test. It is widely described as being trainable, but the opposite was always a design goal, and historically that goal was achieved very well.
You might note that the Raven's matrices are infamous for huge training effects; that test relies on the testee having never seen it before. The SAT doesn't.
> culturally loaded
This claim is true, but nobody claims that IQ tests are insulated from being culturally loaded. The purpose of Raven's is to be a culture-free test. Wechsler makes no such pretense.
> samples only math, processing speed, and verbal reasoning
I'm not sure what you're saying here.
> and tracks prior educational experience as much as it does aptitude.
And this one is false. The point of the SAT is to test only low-level material so that you can be confident the entire test-taking population has been exposed to the material. Aptitude has a very large influence on SAT score; prior education has a negligible influence.
(Prior education will have a larger influence if the population you're investigating includes a lot of people with no education, the kind of people who left school after or before kindergarten. But that scenario isn't relevant to... pretty much any question about the SAT.)
> Draw-a-Person basically isn't an IQ test at all, so I don't see how that comparison clears anything up.
It is an IQ test by the standard you defined: it holds itself out as being "an IQ test", and it is used by researchers to study the intelligence of testees. Did you want to use a different definition?
I've had some IQ tests, one company is now bankrupt, was quite a famous Danish bankruptcy (they offered me the job but I turned it down which looked prescient a couple years later when they went bankrupt), another was Klarna, Boozt also uses a cognitive assessment.
Klarna's seemed like the most proper IQ test although it had at least one question that was wrong.
Military takes the ASVAB which is basically the SAT / IQ Test which qualifies you were certain jobs. For example Nukes in the Navy (which is considered one of the hardest programs for obvious reasons) is 252 score for some combination of scores.
I've worked with at least two such "Nukes". One a researcher, the other a tech startup founder. Both very smart, team-oriented, amiable and down to earth, and inspiring respect (through their ability and how they conduct themselves).
I suppose, on a sub, besides whatever admissions filtering they do for aptitude, you have to take work seriously, and also cooperate and get along with people.
If the ASVAB is basically an IQ test, I’m horrified by how quickly my intelligence has decayed.
When I took it in HS, I scored quite well. I’m almost certain I’m bomb it completely now. Particularly any mathematics I haven’t touched since high school would look foreign to me.
My understanding is that IQ tests can measure a type of intelligence, but fixation on its narrow metrics can lead one to overlook other attributes that are just as likely to be relevant, but aren’t something that necessarily shows up in standardized tests. Add to that the fact that IQ tests can be heavily biased, and leave a lot of ambiguity for the proctors to interpret, it’s not a surprise that they’re so controversial.
I personally would be very suspicious if asked to sit for an IQ test as part of a job evaluation. I have worked for places that blindly worship context-free performance metrics, and it was insufferable.
IQ is a correlation variable that pops out when you measure any group of people's aptitude at a battery of tasks that involve thinking. Basically, if someone's good at one task that involves thinking (say, chess), they're more likely to be good at another task that involves thinking (say, reading comprehension). Apply some Bayesian statistics, and boom, you've concluded that there's some confounding variable (that we call IQ). Then you can start measuring how strongly certain tasks correlate with variable. Turns out abstract pattern recognition very strongly correlated with this unknown variable, so we can use that to predict what someone's IQ likely is.
The point is, due to the very definition of IQ, it's not a narrow metric, and selecting for it does tend to find you individuals who are going to be better than average at most anything. That said, it would seem alarming to me for a job to give me an IQ test instead of cutting out the correlation coefficient and just judging me on the task they're hiring me for.
The IBM PC was an overnight success. It was less than a decade after the first “PCs” and it was the hockey stick moment. I remember x86 clones being seemingly everywhere in just a year or two
A socialist state under the people's democratic dictatorship led by the working class and based on the alliance of workers and peasants.
I think you're ignoring some of the poetic intention of those words, the idea is that the Marxist collective is the dictator. It's turning the concept on its head to put the people at the forefront.
In other contexts such as casual conversation here in the West the term dictatorship means something quite different and you seem to understand that too because you say they're "on the road towards having a dictator" which is surely an admission that they currently do not have a dictator and are ergo but currently a dictatorship.
I'll certainly grant you that Xi has made moves to consolidate power in the individual but that's a separate discussion.
They’ve fallen into the trap Bakunin warned against, that the Party vanguardism and dictatorship of the proletariat model Marx was advocating for would lead to catastrophically authoritarian regimes. Marx eventually had him kicked out of the International. He was saying this around the time Lenin and Stalin were being born.
This. OP is playing by the rules. The rules are silly and dysfunctional, but at least in the short term, that's the winning strategy. I'd probably do the same in his case (or at least a combination of sending maybe 5 well-made resumes or cover letters, to the companies that are the most promising, + 95 "spray and pray" ones for random fishing).
You got it. If you KNOW the company is ACTUALLY hiring and is serious about filling the role, and it's a good fit; take the extra time. But spray and pray helps counter ghost jobs which are impossible to detect until hindsight.
It's not a question of who to blame exactly. The poster's practice illustrates why sending resumes in response to job postings can't work anymore. From the point of view of the new hiring company or new job applicant, it doesn't matter who created the situation. What matters is what we do about it next. You can spray and pray but you certainly shouldn't expect it to "work".
For the poster: was their method a good use of their time? is the job "found" a good fit really? will they last in this position?
For the hiring company: was their method a good use of their time? is this person in any way a great fit? will they last in this position?
The poster complains that few companies sent him a rejection note! Why in the world would they? The poster was protective of their time, and should rightly expect the hiring companies to do the same.
Yes, the job i found is a perfect fit for me and my skillset. I did not fake my way into something.
and my point is: if you genuinely apply to a position and you never even hear back from them, not even to reject you, it doesn't make sense to only apply to a handful. again, emphasis genuine application; answered all their questions and the role is a fit for my resume. That's why you have to spray and pray.
Were your applications that good of a fit? You sent 450 applications for one hiring. Very roughly speaking this means you expect the employer to carefully consider 450 applications for their one job opening (assuming roughly speaking, that everyone does the same). "Carefully" is clearly not gonna happen. You sprayed and prayed but the employer is hiring one person for that one position (not quite true) and can't just spray and pray themselves.
You may have indeed found 450 real and fictitious openings that would be a great match for you. Yes, not impossible. Still the practice puts the employers themselves in a position where these resumes and answers cannot possibly be read. Not carefully, not at all. Again no blame one way or the other. I'm just arguing that we cannot expect the employer to carefully consider all these applications. There is no point in being shocked / surprised / whatever by this. The sprayed and prayed applications will not be read carefully. The employer will find whichever shortcut to sift through the pile and will carefully consider only a handful of all these applicants. Or hopefully, will see the light and consider other kinds of applications - such as network leads (but there are other options.)
My argument is about what we do next. My answer is that it cannot be job postings and answers to job postings. That ship has sailed. (And nonetheless, congrats on your new job.)
> You sent 450 applications for one hiring. Very roughly speaking this means you expect the employer to carefully consider 450 applications for their one job opening
I think you misunderstand. I did not apply to the same job 450 times. These were 450 different companies/positions that aligned with my resume.
> I'm just arguing that we cannot expect the employer to carefully consider all these applications. There is no point in being shocked / surprised / whatever by this. The sprayed and prayed applications will not be read carefully. The employer will find whichever shortcut to sift through the pile and will carefully consider only a handful of all these applicants.
Which is exactly why one needs to apply to many jobs. Almost every job on linkedin has had over 100 applicants after it's been up for a few hours. If you just apply to a handful, there's little chance you'll find success.
I didn't mention it before, but a CFO friend of mine is the one who told me to spray & pray because it's what she had to do and encouraged me to do the same. She was initially against doing it herself, but she changed her mind. And she is a C-suite and is someone with a large network.
> You sent 450 applications for one hiring. Very roughly speaking this means you expect the employer to carefully consider 450 applications for their one job opening
No misunderstanding. Since lots of people operate like you did - more or less - that's the more or less result.
> Almost every job on linkedin has had over 100 applicants after it's been up for a few hours.
Well yeah. If your job search is going to be answering postings on anything - if you START with the linkedin posting as a given - then you will be competing with hundreds of garbage applications. Yes of course. Which means the employer won't read these all carefully (not possible). And the interview process for these will be aimed at filtering the garbage. And you won't like it. Etc etc.
And no argument that sometimes it works. Of course. It's a common way to go about a job search and on average people do get hired in the end, after a lot of nonsense. Everyone also complains a lot about a broken and inefficient hiring process. The inefficient hiring process is co-evolved with this approach.
People also mention approaching the right people and being fast-tracked through the hiring process. That is also a thing.
I use LLMs all the time and do ML and stuff. But at the same time, they are literally averaging the internet, approximately. I think the terms glorified autocomplete and stochastic parrot describe how they work under the hood really well.
A top expert in US Trust & Estate Tax law whom I know well tells me that although their firm is pushing use of LLMs, and they are useful for some things, there are serious limitations.
In the world of T&E law, there are a lot of mediocre (to be kind) attorneys who claim expertise but are very bad at it (causing a lot of work for the more serious firms and a lot of costs & losses for the intended heirs). They often write papers for marketing themselves as experts, so the internet is flooded with many papers giving advice that is exactly wrong and much more that is wrong in more subtle ways that will blow up decades later.
If an LLM could reason, it would be able to sort out the wrong nonsense from the real expertise by applying reason, e.g., comparing the advice to the actual legal code and precedent-setting rulings, and by comparing it to results, and be able to identify the real experts, and generate output based on the writings of the real experts only.
However, LLMs show zero sign of any similar reasoning. They simply output something resembling the average of all the dreck of the mediocre-minus attorneys posting blogs.
I'm not saying this could not be fixed by Altman et. al. applying a large amount of computer power to exactly the loops I described above (check legal advice against the actual code and judges' rulings, check against actual results, select only the credible sources and retrain), but it is obviously no where near that yet.
The big problem, is that this is only obvious to a top expert in the field who deeply knows from training and experience the difference between the top experts and the dreck.
To the rest of us who actually need the advice, the LLMs sound great.
Very smart parrot, but still dumbly averaging and stochastic.
Yup, I find LLMs are fantastic for surfacing all kinds of "middle of the road" information that is common and well-documented. So, for getting up to speed or extracting particular answers about a field of knowledge with which I'm unfamiliar, LLMs are wonderfully helpful. Even using later ChatGPT versions for tech support on software often works very well.
And the conversational style makes it all look like good reasoning.
But as soon as the wanders off the highways into little-used areas of knowledge (such as wiring for a CNC machine controller board instead of a software package with millions of users' forum posts), even pre-stuffing the context with heaps of specifically relevant documents rapidly reveals there is zero reasoning happening.
Similarly, the occasional excursions into completely the wrong field even with a detailed prompt show that the LLM really does not have a clue what it is 'reasoning' about. Even with thinking, multiple steps, etc., the 'stochastic parrot' moniker remains applicable — a very damn smart parrot, but still.
(Although I think the utility of server farms will not be high after the bubble bursts: even if cheap they will quickly become outdated. In that respect things are different from railway tracks)
The Internet bubble left physical artifacts behind, like thousands of miles of unlit fiber. However, that pales in comparison to the value of virtual artifacts like Apache et al. Similarly, the AI bubble's artifacts will primarily be virtual.
I remember having just a couple of issues of Byte but reading them cover to cover and basing my whole understanding of possibilities on a deep dive into a Silicon Graphics workstation they had an article on… happy fuzzy distant memories :)
Imagine another scenario. You are my neighbour. I spill some poison on the ground. Your child gets ill. Am I at fault?