Maybe this is where prompt engineering comes into play. Force the model out of the statistically simple solution by making it self aware of what it’s doing.
Reinforcement training on tasks that are successful and not just valid answers can help with this in the sense that we give models tasks and reward the ones that complete the task most correctly rather then some human qualitative rating on the answer.
Tons of people hate university education and call the whole thing useless. I agree with this notion slightly in the sense that you can actually learn all this shit yourself for free. That is not to say university education is completely useless though.
What I hate about this comment is how it’s redirecting AI to promote their political desire to end university education.
Bro. AI is bigger than that. We will soon need to examine the value of education itself.
Most people don't learn anything on their own. Without incentives, the vast majority of the population would be terrible at reading, writing and math. Without those core skills, other subjects become largely inaccessible.
Having a place to go where everyone is disciplined, and sacrificing opportunity elsewhere to focus on learning (especially overlapping subjects) for a few years is immensely valuable—more so than solely sitting by oneself to cram.
You seriously think that companies hiring people because they are personable and attractive instead of skillful is a new thing? That this didn't happen before AI? And now we need to dismantle the education system?
The proverb "It's not what you know, it's who you know" has been around for a very long time.
I didn’t say dismantle it. Get your fucking head straight.
It’s something that needs to be talked about. It’s a huge unknown. Because on one hand I see that her technical skill is literally irrelevant to the job at hand.
On the other hand I don’t actually know but something feels fucking wrong about hiring someone who doesn’t know shit.
> something feels fucking wrong about hiring someone who doesn’t know shit.
Remember when your grandpa talked about getting a job because he walked into the CEO's office and had a firm handshake?
Hiring people because you like them instead of their qualifications is a tale as old as time. In fact I'd say that outside of bigtech and its copycats, which made a concerted effort to do "data-driven" hiring, being likeable and pleasant is how almost every other industry hires.
And again, this has nothing to do with AI or AI's relationship with education.
> I agree with this notion slightly in the sense that you can actually learn all this shit yourself for free.
I both 100% agree with this in theory and 90% disagree with this in practice. Its like saying you can go become a pro athlete without ever having a single coach in your life. Sure, its possible. Lots of drills and workouts are all written down, just go do them. But having a well-skilled coach who is able to direct you, keep you on task, point out things you're missing, and more is incredibly valuable. In that same vein good educators are your educational coaches. There are tons of topics I would have never chosen to go down, have my original assumptions challenged, and more by having good educators in my life. Both in traditional school and elsewhere.
For sure, once you cross the hurdle of knowing how to read then its all there, just go read the book. The challenging part is having people build you up, point you to the next attainable challenge, and help you when you're struggling.
Steve Jobs did similar things and was at many times much worse. Yet all of this is orthogonal to his achievement of turning apple into what it is today.
Same with Jensen.
It was always clear the affairs were a cover story. Now it’s also clear why the board used that cover story to get rid of him. A lot of them didn’t like him personally.
You missed the point. I’m not saying I support dsa but you are arguing against the weakest version of the reasoning behind why companies use dsa. It’s used as an iq test. IQ is highly correlated with many things associated with intelligence yet even though we don’t take iq tests on a daily basis it still correlates.
The argument was dsa functions like an iq test. It’s not relevant to the job (in fact it’s harder than the job, far harder) such that if you do well on it companies can be confident you can do well on the job just like how scoring high on an iq test makes people confident you’re intelligent.
There are plenty of arguments against what I said above and going after this argument is the better target because the irrelevancy of dsa to the job is widely known and obvious.
Is DSA knowledge really pointless or just a stand-in for an IQ test?
I've been engineering systems for 25 years. On a macro level, most of my time is spent considering the right tooling and stack for a job, balancing budget and time constraints and efficiency against how things may need to scale in the future. On a micro level, in the data model and in the code itself, choosing the best structures for a task is important for maintainability and overall coherence.
In my experience with LLMs, I often need to guide them toward choosing the right ways to model data on the code and database level. LLMs seem to prefer json blobs of unstructured data, naming conventions that drift over time, argument structure patterns that change from one piece of code to another. The net result is a lot of tables and code that an LLM needs to read back through, find all the linkages, and recommit to its window before any debugging can take place. It makes human debugging very difficult if, for example, an LLM decides to just store all the userdata in some json column and then decides to add some kind of stateful preferences to that, and other state to something in memory, and then parse the memory falling back to the database state. I've seen a ton of things like this where weird ways of handling state kinda work, until they don't, and it's extremely difficult to figure out what is taking precedence when or where a thread race is coming from.
All this is to say that DSA knowledge may be even more important when LLMs are writing most of the code, because an unguided LLM will write stuff that can practically only be maintained by an LLM, and make design decisions that don't cohere or scale. If that's not really a problem, and you're certain that the code you're investing in writing will never need human intervention, then I guess it doesn't matter. But with a large enough system designed that way, the boundaries and edges of poor DSA choices aren't going to be apparent until either something breaks or you need to grow.
The problem is that because people believe that Leetcode is a good test of job performance, it implies that the best leetcoders will be the best software engineers.
So to the extent that we software engineers were complicit in that myth, we can’t blame management for believing that AI can replace human software engineers now that AI is the best leetcoder.
There is no law prohibiting the use of IQ tests in hiring, but it is considered to be legally indefensible. The colloquial term for legally indefensible is "illegal". The joys of English.
No, it's not. Huge companies that are enormous targets for employment attorneys routinely do IQ testing. The companies that deliver and proctor general cognitive tests for employment have logo crawls, like any other SAAS company. This is just a big Internet myth.
The reason more companies don't IQ test is that it's an ineffective tool for candidate screening. We all intuitively get that as software developers: imagine hiring off an IQ test rather than just having someone write code.
It's not what? Your train of thought seems to have gotten cut off. The huge company story is fun too, but it also seems to have gotten cut off not going anywhere. You're right that the "internet myth" considers it to be legally indefensible, and thus, colloquially, illegal. But we already covered that earlier, so whatever it was you were trying to add on top of that also appears to have gotten cut off as well, unfortunately. The last few days have not been kind to technology!
Regardless, you vastly overestimate how much effort is put into hiring. The large majority of companies would never study the effectiveness of IQ tests to know if they are useful to them or not. Researching what works and what doesn't isn't on their radar and they wouldn't know how to anyway. If they had a gut feeling that an IQ test would be helpful they would try it on that basis, except they wouldn't try it because they've read online that IQ test use is illegal; meaning that it is considered to be legally indefensible.
It's not legally indefensible. It's not illegal in any sense. It's not proscribed. It's presumptively fine in white-collar employment. Large companies screen candidates with general cognitive tests. Well-established companies administer those tests and brag about their large clients. None of this is a secret.
So it's difficult to see how to square what was previously said on this thread, about IQ testing in employment being "illegal" or "legally indefensible". Those are trivially falsifiable claims. Every HR department in the world knows that.
It would appear my message also arrived malformed. What a terrible week for tech!
What I said, quoting from earlier, was "considered to be legally indefensible". Whether or not it is legally defensible is for the courts to decide, but it is clear that it is not considered to be. Unfortunately your full comment didn't arrive, but what we did get was "This is just a big Internet myth" which appears to have you echoing that it is considered to be legally indefensible. You cannot have a "myth" without belief.
Where the law is abundantly clear (YMMV, law is jurisdictional) is that hiring criterion must be related to the job. It asserts that selection tests must be reasonable, bona fide, and tailored to actual job duties. That does not strictly prohibit IQ tests, but there are effectively zero jobs where an IQ test could be seen as reasonable, bona fide, and tailored to the actual job. Whatever exception you can find is just that: an exception.
IQ tests as a hiring practice are most definitely illegal, as the word is colloquially used. That doesn't mean you cannot use them in certain circumstances. Much to the chagrin of programmers, law isn't binary.
It doesn't prohibit IQ testing at all and your argument fails empirically because large companies do in fact broadly employ IQ testing. None of the rest of your comment makes sense in light of that fact.
You know who I think has a better grasp of the law on this than you do? The HR department at Allstate.
It looks like your message got cut off again. Allstate is well known for not using IQ tests, rather their proprietary JEPS assessment, so be aware that whatever followed to connect IQ tests and Allstate is missing.
Moreover, on JEPS, Allstate does not use JEPS in my jurisdiction. There is no public knowledge as to why, but legal compliance would be a reasonable explanation. Use of JEPS may even be illegal.
It is true that Allstate is a Wonderlic client, but Wonderlic does not use IQ tests, so the connection is not clear. Did your message get truncated again?
Could it be that you are using "IQ test" as some catchall term for any type of cognitive test? You could define it that way in a vacuum, but that is not how it is being used here. The initial comment was clear about how it was being used. In context, "IQ test" was being used to refer to IQ tests that are academically recognized as IQ tests (WAIS, Stanford–Binet, etc.). HR department pet assessments crafted for hiring purposes do not fall under that.
All general cognitive tests are IQ tests, unless you don't believe in psychometric g or the positive manifold. Further: any general cognitive test intersects Griggs v. Duke the way the Wechsler or Ravens does. This is becoming tedious.
All IQ tests are cognitive tests, but not all cognitive tests are IQ tests. For example, an IQ test requires normalized scoring. One test's scoring method being congruent with another test's scoring method is not an inherit property of cognitive testing.
I acknowledge your alternative definition, but I will remind you that "IQ test" for the purposes of this thread was already defined before you showed up. Your definition does not apply. Trying to back-apply your own definition onto an existing conversation is illogical.
Unfortunately the tedium will continue until you realize that your initial response failed to land due to not taking time to understand what the earlier commenter was communicating, not because the thread lacked some kind of technical explanation. Adding more and more technical detail is not going to save you. But it is quite hilarious watching you try!
Your claim boils down to a notion that general cognitive tests not labeled IQ tests are perfectly fine for employment, but as soon as you "normalize scoring" you're in legal hot water. Risible.
My claim is, as always, that you misunderstood the original comment. There is no technical detail you can find that will resolve what is a social issue.
It’s just the hype train is excessive and in everyone’s faces and anything that’s like that genuinely becomes hated.
Imagine if I saved the lives of hundreds of people everyday. And I would throw this fact in your face everyday via TikTok, email spam, YouTube, social media.
Pretty soon you’ll become sick of it. You’ll not care for anyones life. Thats what is happening with ai. Ai is genuinely from a logical standpoint, incredible. But all of that is diluted by overhype.
Except AI isnt saving lives. It isnt even usually improving lives. For the most part it is being used to justify laying people off, to heap more responsibility on those that remain, to destroy the livelihood of artists, and to flood public spaces with slop. Even if the capabilities are incredible, the average lived experience of it is awful.
It's not the hype we are tired of. We're tired of the hype being wrong, and pro-AI people being quasi-religious zealots about, while ignoring the massive costs and harms. And if you question the AI dogma, you are accused of being either a Luddite who doesn't understand, or a victim of a Chinese influence campaign.
The hype is not wrong. It’s just too much. Navier stokes and hundreds of other unsolvable math problems were solved. That is genuinely real and worth the hype.
My entire company now uses AI to code. Nobody writes code by hand at my company anymore. At this point if you can’t see how incredible and horrifying it is then you’re delusional.
I get it though. One year ago you’d be right. The paradigm shift is happening so fast people haven’t updated their priors. What you believed to be true just changed. It flipped several times this year and people who keep up to date had to completely flip their understanding of the world and the nature of AI several times. This is not an easy thing to do and thus many people like you just missed the boat.
There’s no doubt that there is a financial ai bubble. But that bubble is completely orthogonal to the complete paradigm shift of the actual capabilities of AI.
I think the point is not that AI is not great, but it's less than what was promised.
Very good analogy with Tesla FSD. Is it great tech- yes, of course. Is is the thing that was promised to come out 10 years ago? Far from it.
If you expect AI to do everything for you, it's still not there. And if you're sceptic or even neutral to this topic- AI is just causing more problems (especially in security) by the day.
Did they take food off the table from people initially or didnthey not? Electric lighting certainly must have killer the job of the gas lamp lighter or at least forced him to change. Cars obviously impacted horse drivers.
Electric lighting actually did the inverse; Because of electric lighting, people quickly became overworked, and that's what directly formed to the creation and empowerment of unions.
Yes so why do you feel the newest round of technology won't also end up creating more work and jobs? The people being replaced by electric lighting and motorcars would have found it unimaginable too that it will lead to more jobs instead of killing large sectors of jobs without replacement.
Because nobody was 'replaced' by electric lighting. Motor vehicles did take jobs, but they created more jobs. And you could argue AI will lead to that too, but the primary goal of AI is complete invalidation of human work and creativity. Not only that, the fields AI touches upon are sweeping and all-encompassing. AI may create more jobs, but I'd be skeptical if it leads to more jobs than it takes. Unemployment is already a massive issue these days, and AI certainly doesn't help.
They did not took food off the table from people. They did not all caused mass unemployment and their goal was not to achieve mass unemployment. Which is literally stated goal of AI.
There is no need to argue by imagined history. We have real one.
Are you saying motorcars didn't cause mass unemployment among horse carriage drivers? Or that electric lighting didn't kill the gas lighter's job? Or that agriculture still requires nearly 70% of the population? You are making up completely imaginary histories with zero numbers or examples now. All labor saving tech were implicitly or explicitly sold as a means of reducing requirement on a given number of workers per amount of work. That has always been the point of technology from the industrial revolution forward. But you have special reasons to villainize and call AI companies ultra evil any more than any other technological advance throughout history.
What do you think was the point of tech if not to replace rhe amount of labor needed for a given amount of production? Why do you think luddites were smashing looms? They were wrong then and modern luddites are juat as wrong today. Again you had zero figures or data and completely ignored my agricultural data because you prefer paranoid villanizing imaginations not reality.
The effects of all of those those changes that you mentioned were not that drastic when it came to getting food off the table of those millions of people, I'm saying that those millions won't remain like sheep and will make things right sooner rather than later, that's what always happens (see Polanyi, for example).
But, yeah, you AI zealots on forums like this one can keep thinking that we, those millions, will just remain like sheep taken to the slaughterhouse while you enjoy your higher comps while in the service of today's AI masters.
You are the zealot who got frenzied into psychosis by one of the biggest accelerators of math and science that we invented.
And changes were in fact drastic, thr fact that countries like UK and Germany for several years became the most powerful countries in the europe due to the industrial revolution.
Solving abstract mathematical puzzles and enganhing in pointless research into substances for crnturies gave us ultimately the Haber-Bosch process, which literally keeps the human race from starving. You would prefer humans starve just to protect som jobs.
On the topic of farming, that used to be the occupation of most of thr population in olden times, today its barely 2 percent. Agriculture tech has already killed millions of jobs.
No amount of technology can solve the starvation in those countries. The starvation in those countries aside from physical blockades and wars like in Palestine if not usually due to lack of tech or agriculture. In any case, even if the land is bad, Israel itself is another example of crap land turned fertile by new technologies they were forced to invent during the course of the 40s and ahead by scientists like Bass. Here you want to block the biggest engine of technological and science development that has ever been invented.
The AI companies do what they must to market themselves, there is no reason to think it should be much different from other past mass disruptive technologies (today agriculture is down to just 2% of the population from 60-80% earlier for instance).
And if it solved most labor and jobs, then what? Would you rather block it than try to see how we can transition to a economic and government system that will actually let us enjoy it.
And I doubt we will ever run out of problems, there is always outer space as an endless source of new problems.
Agriculture as a job went from 70% of the population to under 2% in 200 years. Why are you not complaining about agriculture tech causing mass job killings?
Literally every technological advance till now has led to massive abundance and increase of food, and of the carrying capacity of agriculture itself till now. You have to explain why you think now technology will do the opposite of what it has done for millenia continually increasing the amount of and variety of food available worldwide. You have to show how this engine of science and tech research will do the literal opposite of what science has always done to food availability.
Well agriculture was not selling itself with "we will nake humans obsolete, replace all humans and make all people except few richest poor". Agriculture improvements did not came with "hahaha you looser artists you are not needed anymore, all art will be done with AI you loosers".
Agriculture selling point was and is "we will feed you"
I am also really not interested in alternative history. Nor in debate where people who know literally nothing about actual historical events make up fictional past out of imagination and movies and use that as argument.
> You have to explain why you think now technology will do the opposite of what it has done for millenia continually increasing the amount of and variety of food available worldwide
The destruction and desperation for average person is what AI companies sell. It is what THEY claimed will happen.
What is there alternative abour agriculture going from 70 percent to 2 percent of the population? Why do you think 8t had to be spelled out to the jobs that were killed that we are replacing you. The inventor of the tractor, thresher, etc didn't have to see what the point of these techs were. And it already made megafarms more profitable than small plots. Again, you failed to prove why technology wouls suddenly do the opposite of what it has done for centuries.
You still failed to prove how technology is to blame for corript countries or countries being bombarded wnd blockaded. Instead you lie and imply technology causes starvation.
You should be the one proving things, but all you do is making claims you clearly made with 0 knowledge of economic history. You are talking about periods of booms and new jobs.
I did not blamed technology for goverments. I simply pointed out that starving is not technological problem. Which was your assumption.
You literally have given zero numbers till now. I gave actual numbers. You have provided nothing more than imaginary fluff without any data. You make up imaginary ideas in your head then claim someone providing actual numbers is making up stuff.
Starvation is a tech problem too, without haber bosch we would probably had mass starvations in history. Without Israel inventing new forms of soil treatment and irrigation, the country wouldn't support agriculture at any reasonable rate. Why would you want to block the biggest killer of starvation as if doing that would solve the problem of corrupt governments?
Yes so why do you think if a job going from 70 to less than 3 percent hasn't caused mass unemployment worldwide, thr current newest tech should be any different? All new revolutionary techs feel devastating the first time they are introduced.
Anyone who believes in infinite imaginary beings and imaginary things that supposed being wants you to do has no right to call anyone else religious. If you are not one of those, then a meaningful discussion can begin.
Silicon valley wants me to be completely 100% dependent on Silicon Valley to do anything. That's their goal, it's what they've been working toward this whole time.
"Trust me bro", they say. They have new and dangerously powerful tools to get me to trust them harder. Me, one of the dumb fucks.
Luddism in my view is a civilization ending disease. These same types of thinking will ban advances in medicine because its "unnatural" and prevent the steps needed to become an interplantetary and further civilization eg because autonomous and intelligent nav is "evil" and steals jobs.
> Luddism in my view is a civilization ending disease.
Sounds more than a bit OTT dramatic there, not to mention ignorant of what the Luddites cared about.
> These same types of thinking will ban advances in medicine because its "unnatural"
A) Hyperbolic nonsense.
B) So what? Seriously, let's assume all medical advances cease right now, this very instant. We'd be stuck with modern medicine as we know it - that's hardly a civilization ending calamity.
It is, the earth is a finite system. And is still subject to catastrophies. We already use medicines that give us resiliences we didn't have, we'd need more of it, eg raising the baseline environment we can survive in. And when earth runs out of energy for thr population then what? Do we start killing people or stop having kids. Going interplantetary really is not a choice in the medium to long term.
And the point is technology is a web of dependencies, any piece of tech or science even if seemingly having no connection to somehthing is a key piece of the puzzle. Who would have thought 200 years ago that high quality optics creates more powerful calculation devices?
> It simply does not do the magical stuff people claim it does.
What are you even talking about? You can see it do the magic yourself for like $20. The lean proof of the exponential time hypotheses disproof for exampe is straight up on github right now for you to look at.
What magical thing are you waiting for? It is literally currently revolutionaizing all engineering work.
It's just fantastic, to think the cost of running the Stasi was once 20% of national GDP and now that same oppressive all seeing surveillance and data collation is available for pennies on the dollar.
For clarity, I'm more or less indifferent to AI, I do understand how it all works - it's humans with AI I don't trust and that I don't see being adequately reigned in.
Did the centuries of work of scientists and philophers in chemistry put food on any groups table? Do you think haber bosch would have just dropped out of the sky one day randomly?
I can understand a triangle and how come Pitagora’s work might have been worthwhile, but I cannot even begin to understand what this Navier thing is about. You seem to believe in higher maths putting food on our tables, care to delve more into that?
Also a reminder of how come the post-WW2 USSR was a technocratic-scientist state/society and even so, by the end of its existence it had failed at this basic task of feeding its population.
The reason USSR failed was because its economic system was as we know in hindsight incorrect...and literally unscientific theories like Lysenkoism foisted by the government. If you wanted to argue the USSR failed due to science and tech, then you would be wrong. You also have to prove the USA didn't succeed due to tech.
If you don't know how and what kinds of math is needed in every tech and science we use, then I have nothing to say. I should have recognized when you said this and stopped interacting further.
On one hand it is far superior to humans. On the other hand it needs our guidance. Both of these are true at the same time.
reply