The simplest way to prove to yourself that LLMs are not intelligent is to give one a "niche" open-ended problem. It's a non-starter with a LLM, it doesn't even know where to begin.
>>109954391>it doesn't even know where to begin.Do you?
>>109954416>i wouldn't know, eitherI know. You and the rest of your AI-using horde are nonsentient.
>>109954391Be that as it may, most problem aren't novel and very few make a living off of solving novel problems. Most work is repeating stuff that's been done millions of times before. If AI takes away even half of that work, it will destroy employment for a huge number of workers.
Where is that niche open-ended problem?
>>109954537Ok, but that's basically my point. Your entire concept of programming is encapsulated by dataset of stackoverflow answers, leetcode problems, webshit github repos and job interview questions. And you think that represents intelligence. No wonder a token guesser is like an omniscient deity to you. From my perspective, none of this hysteria even makes sense. It can't do anything I do or care about.
>>109954574>saaar? i can't even THINK of my own niche and open-ended problemMaybe the bot can invent one for you.
>>109954391It usually does know where to begin, even with new ideas and problem sets. But you're not going to believe that so fuck you.
>>109954391There is a codebase refactor needed to improve performance where I can imagine what the change will look like in my head but I just can't get the new sol to execute it in a tasteful way, but im also demotivated from doing it myself (i probably will have to eventually though). One of the few cases where AI has actually made me slower.
>>109954612>It usually does know where to beginNo, it doesn't. I've tested this myself and together with AItarded colleagues.
>>109954391just feed them link of documentasion to look at. start vibeloping.can you start vibeloping app based on the website documentasion i provide.
>>109954391But most coding tasks are not niche open ended problems. AI can already code better than 80% of SWEs. The only reason SWEs still have a job is because they have other tasks outside of coding that AI cannot or will not be allowed to do yet
>>109954722>most coding tasks are not niche open ended problems.Correct and I accept your concession. No matter how well a slop generator does your slop generator job, it's not a metric of intelligence.
>>109954391Your post would be a lot more convincing if it looked a bit like this>i tried to use an LLM for a niche open-ended problem>specifically, this problem:>i used this prompt:>i used this harness, with this model and these settings>here is a full log of the session>here is a repo of the artefacts it produced>as you can see, the agent was unable to solve the problem and didn't even know where to beginWithout these receipts your post smells like horseshit. The code produced by LLMs (without lots of guardrails) is a bit shit in the sense that it's excessively verbose and repetitive and soul-crushing to read, but they can definitely write it and solve problems.
>>109954598The entire fundamental way you perceive the world is retarded, and that's the point 'AI-using horde' are making.Nobody gives a shit about your work, nobody's ever given a fuck. Nobody cares about artists losing sleep, nobody ever gave a fuck about coal miners losing their job, nobody's ever cared about horse ranchers losing their job when cars came. You're a retarded idealistic child with no concept of how the world works and a very naive understanding of the human condition, it's not my problem that people are losing work, it's making my life easier and therefore I will use it. The worst part is how you retards hide behind a fake altruism, acting as if you do give a fuck, when you're in fact using computers or phones that are only made due to kids in some third world shithole being enslaved to mine the metals needed for it, but AI is the problem?Shut the fuck up nigger, either you care about people, or you don't, stop this moral grandstanding when nobody really gives a fuck and it's so easy to prove that you don't either.
>>109955067>my horde doesn't heckin' care about anything!!!>human activities are le BAD!!!!!!! Notice how my simple observation makes you lose your mind with rage.
>>109955091>noooo I care moreYou seem to not have the grey matter necessary to understand the point, feel free to resume whatever you were doing.
>>109955104You're legit mentally ill if you think your maximally generic talking point, constructed by 100 IQ marketing wagies for mass consumption by 85-90 IQ costumers, is filtering anyone.
>>109955203So I called you retarded for acting as if you cared when it's so easy to prove that you don't, and what you took from it is 'omg humans are lebad!!!', and I'm the one regurgitating braindead talking points..? How the fuck do you people even manage to connect your computers to the internet?Why aren't you in a field somewhere putting potatoes in a basket or whatever else could bring you a more fulfilling life than pester people about pointless non-sensical and emotionally charged arguments.
>>109954897Ok. Here's a problem I've been playing around with lately: ever see gameplay footage from games like Outrun? When you go fast in the game, the scenery sprites on the sides of the road turn into a flickering tunnel and it's just unpleasant to look at. It's not that bad on a real arcade machine because the CRT's phosphor persistence gives it a bit of motion blur, but on a modern monitor it looks like shit. It would be nice for such a game to have actual motion blur, IF you found some way to make it look good AND fit in with low-res palette based graphics.Can you get your favorite token guesser find a satisfactory solution on its own, based only on visual feedback and commentary on what you like/dislike about its attempts? You can use whatever harness and prompt engineering magic you want.Protip: it's just not gonna happen.
>>109955271>It would be nice for such a game to have actual motion blur, IF you found some way to make it look good AND fit in with low-res palette based graphics.>Can you get your favorite token guesser find a satisfactory solutionno and that's a good thing. it's a waste of modern technology to do absolutely pointless luddite things like this. i'll keep using SI to generate actual value for customers and you can keep doing... whatever THAT is lmao
>>109954391https://www.youtube.com/watch?v=PoDJZ4Ln9Rwhttps://en.wikipedia.org/wiki/Spike-timing-dependent_plasticityhttps://pith.science/paper/2604.12167SOON...
>>109954391Professional coders were not in the business of inventing anything since well before AI came around. Nobody should care what they think because the corpos they work for destroyed the sector, and they're obsequious to their bosses.However even people who do actually care about making something good are often adopting AI for their projects as well. Even if it can't solve the open-ended problems, it can get the trivial stuff out of the way a lot quicker so you can just focus on those.I don't use any of these tools so I am not steelmanning my own position here. The problem with AI in programming imo, isn't dev side use, it's shipping things with AI based "solutions."Like retards shipping a "translated" version of a game with an active language model changing text in real time and making it run like garbage, instead of just translating the fuckin thing.
>>109954391This is like three years out of date. There isn't anything an AI can't code and if your idea is actually unexplored without any basis on existing code, if you scaffold it, the AI can still continue it with context.
>>109955271You can't even word your problem in a way that a you understand you expect an LLM to read your half-formed thought.Look in the mirror: you're a retarded middle manager.
>>109954391AI can't code because it is artificialSI can code better than you because it is superThank you for your attention to all you need
>>109957175>AI slopper seethes in the most predictable wayWhat part of it didn't you understand?
>>109954391AI is a glorified boilerplate generator nothing else.
>>109957147-t. frontend retard
>>109957281>bro can you make it so sprites emulate CRT motion blur because I have autism and this request is entirely subjective to my neurodivergent nostalgia to the past?
>>109957318>t. mentally ill retardAsk the token guesser to explain it for you, maybe.
>>109957424I don't have problems defining problems to objective facts and solutions. You, however, seem unable to even understand that "CRT blur is like so good so make the sprites look like CRT motion blur" is abstract and subjective with no defined solution. Point back to you being a pointy haired middle manager.
>>109957456>saaar you did not provide exact instruction about how to solve problem SAAAAAAAAAR!!!!!Keep up this mentally ill behavior. I know you can't comprehend this, but you're making my case for me. :^)
>>109957492Explain how I'm wrong with facts and without ad hominems. Given how this is your default response, I can see how you are incapable of having LLMs produce what you want.
>>109957517>Explain how I'm wrong with factsThe fact is that you've become a literal NPC from all the proompting and you literally start foaming at the mouth with rage when confronted with an open-ended, real-world type of problem.
>>109957531So the answer is no, you don't know how to prompt an LLM and you resort to foaming at the mouth with ad hominems by failing to engage any sort of discussion. I can only assume you do this because you know I'm right, you get bad results from LLMs *and* people because you have the communication skills of a ill-tempered toddler. Sorry anon, I think your problem is between the computer and the chair.
>>109957556>you don't know how to prompt an LLMOk. Why don't you tell me how to prompt the LLM, mentally ill insect? I told you what the problem is, gave you a general direction for a solution and what the constraints are. Convert it into a prompt such that the mindless machine can solve it autonomously.
>>109954391Sorry kiddo, RLVR solves everything. For what it doesn't know now it can just do problems in a sandbox over and over again until it does. it's precisely this capability that scares ppl.
>>109957575Let's see how your magical electric genie handles >>109955271, then.Prediction: you will seethe impotently.
>>109957567I told you that your request is subjective so if you want a good result you would need to figure out how to objectively describe your end goals. Your request is difficult for a human who has the benefit of contextual world understanding, I know what a CRT is, I know what it looks like, and I can even *interpret* your request. But as a human, I'm unlikely to do what you want. And you want a machine to do it.
>>109957587>your request is subjective so if you want a good result you would need to figure out how to objectively describe your end goalsThat's not a solution to the problem. Try again?
>>109957601Let me guess, this is what you chat log looks like with the LLM. I bet you have hilarious chains where the statistical LLM degrades over time as you tell it it's wrong over and over again and then you start swearing at it. I heard that works well for shovels too when it won't dig through a rock.
>>109957587holy shit, op is a retard but you're even worse. are wagedrones really like this at work? you don't sound like you can even do anything without autistic back and forth with a manager
>>109957666I didn't realize I was talking to a LLM here but the way you keep trying to nudge me towards your "preferences" is a dead giveaway. Don't worry about it. It was an example meant for actual humans who understand what real world problems are like.
>>109957670No, I'm explaining the problem and the failure to recognize LLMs are the autistic nit-picking programmers you accuse me of being. The only difference is LLMs are programmed to say yes to any task no matter how poorly defined and ultimately achieving the result that frustrates our dear OP. Your mistake is you think what I write in a basketweaving forum is how I treat real people when I would hope that my message indicates that I actually have some empathy and understanding of bedside manner.
>>109954391It doesn't matter whether AI is intelligent or not. What matters is whether or not it can complete a given task.
>>109957703>LLMs are the autistic nit-picking retards like meyeah that's why he says they can't solve """open ended""" problems. how can you be this filtered?
>>109957715>What matters is whether or not it can complete a given task.See >>109955271Reflect on why it can't complete this task. Hard mode: without whining that the task is le bad.
>>109957683Look anon if you want to use a paintbrush backwards and complain that you can't make it work right go ahead. I'm not actually impacted by you being unable to achieve what you want. I thought your question is why you can't make a painting and not a struggle session about how brushes are a scam.
>>109957774I accept your full concession of my point.
>>109957742Humans can't solve open ended questions most of the time, so the expectations are laughable.
>>109957791>retreats back to the 2016-era "humans le also..." generic defense templateWhat better way to signal your own defeat.
>>109957787Given you don't actually say anything concrete or what could be defined as an actual rebuttal the concession to this debate was when you first replied with an ad hominem instead of a response addressing what I said. You have never made a point. You never once said "no anon, you're wrong, look at this great prompt where I said exactly what I wanted and it still failed".
>>109957809Notice how you're stuck in a loop of desperate cope over the bot's inability to solve the kind of open-ended problems that normal human devs have to solve constantly.
>>109957817Hold up a mirror to see the loop.
>>109957809>>109957821>b-b-but saaar you did not give me objective prompt!!I didn't have a prompt. There is no prompt. The information available to the programmer in the given situation has been provided. If a LLM can handle the task using a magical good prompt, write that prompt and come back with the result.
>>109957838Go workshop your prompt with your LLM.
>>109957858>mentally ill brownoid continues loopingConcession accepted. Moving on.
>>109957869Actually I told you to workshop your prompt with an LLM. Best faith is you want me to do your homework for you.You looped with an ad hominem which is your default response whenever you feel intellectually insecure.
>>109957791>the absolute state of shameless wagebots
>>109955271>based only on visual feedback and commentary on what you like/dislike about its attemptsThat would be extremely painful but what makes you think it can't be done?
AI is overwhelmingly "okay">hey Claude how do I do this thing in HaloPSA>yea, you can't>oh? *checks myself*>find it>hey Claude, but the option is right here>you're absolutely right! *posts unrelated source*repeat for basically every IT issueit's great for consolidating data that i don't care to read for the CEO to pretend like he read it, but other than that, the gains have be overwhelmingly disappointing compared to how much money is being invested into it
>>109957939>what makes you think it can't be done?Because if you don't run out of patience, you'll run out of tokens. Or more likely you'll start thinking about the problem yourself and telling it what to do.
>he said for the 6th time
>>109954391Anon, AI solved Navier Stokes. I don't know what else do you need.
It needs concrete direction and context because it can't derive that itself. That is the human part. I'm not mad that it still needs us lol.
>>109958025>AI solved Navier StokesThat's a lie, though.
>>109957999At least you admit that you don't understand the problem, let alone what an acceptable solution is.
>>109958094What are you on about, retard?
>>109958025It takes two years for millenium prize math problem proofs to be verified. If I remember correctly, the only documented prize receiver waited even longer than that.And OpenAI themselves said they don't actually want the prize money, removing all stake in the proof's verification.In short: they didn't solve shit. They loed just like they always do.
>>109954598>>109954537>Q: How do I do X?>A: Don't do X, do Y.>A: Use case for X?
>>109958110You just said you don't even want to figure out the problem, and if you can't define the problem, then how, exactly, is the AI supposed to read your mind?Let us help you anon:Step 1: in two sentences, define the problem.Step 2: in two sentences, what is the end product that would solve this problem?>Ok. Here's a problem I've been playing around with lately: ever see gameplay footage from games like Outrun? When you go fast in the game, the scenery sprites on the sides of the road turn into a flickering tunnel and it's just unpleasant to look at. It's not that bad on a real arcade machine because the CRT's phosphor persistence gives it a bit of motion blur, but on a modern monitor it looks like shit. It would be nice for such a game to have actual motion blur, IF you found some way to make it look good AND fit in with low-res palette based graphics.What exactly do you want anon? A shader that runs on Retro Arch that physically accurately simulates a CRT display? Decompiling the game and rewriting the graphics engine to modern standards with actual motion blur? I don't even think you understand the scope of work you're asking.
>>109958134It "solved" an easier sub-category of the problem, where instead of proving an intrinsic blowup, you prove a blowup by introducing an external force. The idea behind this variant was that if you can at least prove anything that way, you'll surely produce insights into the equations themselves and make significant progress towards solving the full problem. Back when the Clay Institute formulated it, no one could imagine that a bot swarm's LACK of insight and understanding could be weaponized to find a "solution" that adheres to the technical requirements without producing anything interpretable as a fluid mechanics situation. It's literal specification gaming playing out in the wild.Mathematicians have actually already proven that OpenAI's approach is a dead end as far as solving the full problem goes. It really was just worthless AI slop.
>>109958188>You just said you don't even want to figure out the problem, and if you can't define the problemYou're hallucinating.>What exactly do you want anon? Ask the token guesser if it's too difficult for your own brown mind.
>>109958212Ask it what anon? Did you even read what I wrote? Do you want something that simulates a CRT monitor as a screen effect perfectly tuned to your emulated Outrun game? Do you want it to find the closest working CRT monitor for you to buy? Do you want it decompile the game and rewrite it to modern 2D game standards?I don't know why you react with such hostility but as mentioned, it's likely because you're frustrated because you realize you can't answer what the problem and solution is, which demonstrates a fundamental scope understanding problem and why you really should go back to the drawing board. I'm helping you critically think and hopefully you'll stop doing this kneejerk anger reaction in the future.
>>109957939For you
>>109954391>@grok, is P=NP?
>>109958255>Ask it what anon?Ask it what the request means. You're too dumb for basic reading comprehension and I bet the token slop can do better than you.
>>109958192Frankly if thats the case, it's further than I expected them to take it. I would have assumed they didn't out out anything coherent and just appealed to the integrity of the verification process until we all just forgot. They have already done exactly this with coding challenges and even benchmarks in the past.
Any time i consult the oracle i am unimpressed, but of course i'm like op. Not in the sense that i'm a gigantic faggot, but that i write firmwares (shit's too custom to abstract with LLMs. They have tries selling code generators and HALs on us for literal decades, but anyone worth a damn writes code bare metal), create new products and figure out new systems that don't exist yet.My colleague cursed with writing the apps (but hey, he likes it) uses the tools and we keep having the same conversation: the shit keeps changing stuff it's not supposed to, or writes incredibly flacky functions.Useful for prototyping, but there is no substitute for skill, and you get skilled by doing actual work. I WANT to use it for all the menial bullshit, like generating docs, or is there a plugin for thunderbird so i don't have to waste time to type answers to your LLM generated walls of texts?But i want local LLMs and my PCs have too little RAM to be useful, fucking shortages.
>>109954391>AI cant code>Personally vibe coded about 7 different services for my home server>AI cant code Hmmmmm
>>109954537I suspect the exact opposite to be true
>>109954537This is true. It's also very easy to ask the AI "can you take these 4-5 solved problems and blend them to help me create a new novel solution" and it will work. Not writing every line of code anymore is depressing but I would be lying if I said I wasn't making more stuff faster.
>>109958308But you clearly didn't, that's why you're mad, and I explained why your request is extreme scope and given you still refuse to respond what your solution is despite me giving you three tells me you don't know what the solution is, and I mean this in the high level "as a user I would be happy if it did this". It's not hard to put an idea as two sentences anon, instead of reenacting forth grade where the teacher made you scream after she asked what 10 times 10 is, maybe you take some deep breaths.
>>109958269>Penis = No PenisLiterally just tue tranny agenda.
>>109955271Respectfully post a video of what you're talking about. Something "looking like shit" is not a descriptive explanation. And Claude probably can solve it if you can articulate it.
>>109958381Naw he's right. Companies would rather pay for the occasional mistake and use AI then to pay some faggot Cs MaJoR +$100k a year with benefits and paid leave who will constantly be calling out sick, shitting out broken unoptimized code full of more bugs than your moms fuckbox.
>>109958188his problem is temporal aliasing in an outrun-style game. he wants to solve it using some kind of motion blur that complements the game's art style. he doesn't tell what kind of motion blur to implement because figuring it out is the problem. that wasn't hard, you're just a massive brainlet
>>109958454Okay, and how is that motion blur added to the game? Modding the ROM? Decompiling the game and rewriting the renderer?
>>109958418>But you clearly didn'tYou know what? I just did it for you. I simply asked gemini: "what do you think they're actually after?" followed by a verbatim quote from my post. It replied:>They are looking for a retro-compatible temporal smoothing technique—a way to eliminate strobing and phantom array effects on modern displays without destroying sharp pixel art or breaking a restricted color palette.>The Problem: Sample-and-Hold vs. Phosphor Decay>In pseudo-3D rasterizers (OutRun, Super Hang-On, Space Harrier), roadside sprites scale and shift across massive pixel jumps per frame as you speed past them (often moving 30–60+ pixels in a single step).>On a CRT: Phosphors don't turn off instantly. As the electron beam moves on, phosphors decay exponentially over several milliseconds, acting as a natural hardware temporal low-pass filter (motion blur).>On Modern LCD/OLEDs: Sample-and-hold panels hold sharp pixel states instantly until the next refresh. When high-contrast pixel sprites jump large distances frame-to-frame, your eye tracks the smooth motion across the screen while the display delivers discrete steps. This creates temporal aliasing—stroboscopic flickering, double-imaging (phantom arrays), and visual fatigue.>Why Standard Motion Blur Fails>If you apply conventional post-process screen-space motion blur (linear alpha blending or velocity vector blurs) to low-res pixel art:>Palette Pollution: Linear interpolation generates hundreds of intermediate RGBA colors outside the game's strict color palette, making sprites look like dirty, semi-transparent grease smears.>Loss of Pixel Definition: Sharp edges turn into muddy gradients that clash with the crispness of surrounding art assets.And then it lists a bunch of "common sense" solutions that don't really work. I can see why you're so impressed with it. It can actually simulate reading comprehension, which is an impossible godlike skill for you.
>>109954537That's true of, for example, CRUD apps. The problem is that we've largely solved that problem with technology like RAD, ORM, and other labor saving tools
>>109958511>guys I used the shitty cheap Gemini model for a highly technical problem so the conclusion is all LLMs including Astra and Fable are bad modelsyou gave me a laugh anon, also you conflated a description of the problem from the dictionary with a solution, which it isn't
>>109958611The conclusion is that shitty cheap Gemini is more intelligent than your race."How did you figure out they're interested in the technique itself and aren't asking about modding the ROM, decompiling the game and rewriting the renderer or something else like that? ">It comes down to three specific cues in the phrasing and the technical realities of that era's hardware:>1. "Games like OutRun" (Genre-level, not binary-level)>They didn't name a specific release, platform, or file (e.g., "the Saturn port" or outrun.zip). Framing it as "games like OutRun" means they're pointing at a broad architectural artifact common to the entire era of pseudo-3D rasterizers (Super Hang-On, Space Harrier, Lotus Turbo Challenge). They are thinking about a class of visual problems, not a single software target.>2. The "IF" clause is an aesthetic paradox, not an engineering bug report>When they say:> "IF you found some way to make it look good AND fit in with low-res palette based graphics.">They are highlighting a theoretical rendering dilemma: How do you introduce temporal smoothing without violating the rules of retro pixel art? Standard motion blur breaks low-res pixel art by introducing dirty alpha-blended color pollution. That’s a game design and shader rendering question, not a reverse-engineering request.>3. Original 16-bit ROMs physically can't do it>In authentic arcade hardware (like Sega System 16 or 32), sprite rendering was handled by dedicated hardware scanline fill units. The hardware had hard limits on sprite multiplexing, fixed color lookup tables, and zero support for arbitrary per-pixel transparency or alpha blending during sprite scaling.>You couldn't mod a 1986 OutRun ROM to do motion blur even if you wanted to—the Motorola 68000 CPU and video hardware would choke immediately. Therefore, any viable solution naturally lives in modern renderers, shader pipelines, or source ports (like Cannonball, the C++ decompilation of OutRun).
>>109958447You've understood half of it, I think that just like when steam locomotives became more efficient, we didn't see demand for coal become lower, I think this will have a multiplicative effect on demand in the market
>>109958192>It "solved" an easier sub-category of the problemIt literally solved the official C and D statements. This qualifies as "solving Navier Stokes millennium problem". If they wanted, assuming the proof is correct, they could even get the 1M dollars.>>109958134>And OpenAI themselves said they don't actually want the prize moneyIf the proof is correct, they could absolutely get the 1M dollars. They don't do it to look like philanthropists or whatever.
>>109959018>It literally """solved""" the official C and D statementsThat's literally what I said minus your intentionally misleading marketing babble.
>>109959037You were implying that the proof is useless because it's an easier problem and that somehow it's "invalid" because It lacks insight. Assuming the proof is correct, if a human presented that exact proof, they would get the millennium prize.
>>109959018>They don't do it to look like philanthropists or whatever.No, they're not filanthropic in the slightest. They are declining to keep the demand for proof validation relatively low among the community.
>>109959184>No, they're not filanthropic in the slightest.Never said they were.>They are declining to keep the demand for proof validation relatively low among the community.This could be a reason too.
>>109959071>You were implying that the proof is useless because it's an easier problemYou're just lying blatantly now, Schmuel.
>>109959276Why did you call it "worthless AI slop" then? Just because it had no muh insight? I bet you would view this proof as worthwhile if a human mathematician proved it. Why the double standard?By the way, buckbrokenmaster and the anthropic nigger (the two mathematicians who accused openai of "stealing" their solution) apparently had a similar approach in mind, so your criticism about the fact that the swarm had no insight whatsoever is wrong, because two human mathematicians tried to use that exact same approach (unless you want to claim they didn't have no insight whatsoever into the problem).
>>109959389>Why did you call it "worthless AI slop" then?The post explains it in detail.
>>109959400>"the proof is worthless ai slop">gets called out>noooooo you are just lying, I never said it's useless>gets called out again>just read my self contradictions again bro
>>109954391>work for fang >use this shit daily>works>provide it an architectural plan >fuck off >finishes what I want Your copium is nuts.
>>109959473Anyone who cares to read this exchange will see what a lying, diseased subhuman you are and correctly associated it with your AI marketing agenda.
>>109959490>it was real in my mind
>>109958874When is that going to happen? Because right now it's harder to get a software engineering job than at any other point in history and it keeps getting harder and harder every year.
>>109959389>By the way, buckbrokenmaster and the anthropic nigger (the two mathematicians who accused openai of "stealing" their solution) apparently had a similar approach in mind,Manufacture cobtroversy to distract from the fact that the proof is never beinf verified.The fact that one of them worked for anthropic is proof of that.
>>109957586You're nostalgic for a time that has long passed, and not in need of any computer programs. What you actually want is to bond with someone who will try to recreate that feeling you remember while you're both poking at the problem and of course no AI can substitute such a meaningful interaction with someone else.
>>109959862>realize LLM can't do the thing>start psychoanalyzing me insteadThis is the most desperate cope ITT so far.
>>109959871Looks like I was right
>>109959901Looks like you've mindraped by your cult.
>>109959705There are layers of confounding variables that interact with how easy it is to get a job, how much shit people want built is just one factor in the equation
>>109960556Not really. It really is as simple as supply/demand. In 2021 any retard who could solve fizzbuzz was guaranteed a job at FAANG because there was that much more demand than supply. Now it's completely flipped and you have to have been in the top 10% of your class in a top 10 CS school just to be considered for an interview.
>>109960577The short lived post pandemic hiring boom is hardly representative of the job market at large, and headcounts today are still much larger than they were in 2019
>>109960623The pandemic hiring boom wasn't that different from how things were in the 2000s, 90s, and 80s. In 2007 if you could do fizzbuzz you were pretty much immediately hired on the spot. In the 90s if you knew html you were immediately hired on the spot. In the 80s if you knew C or assembly you were immediately hired on the spot. Nothing like that exists now and nobody back then had to go through anything even 1% as difficult as what the hiring process is like today.>https://blog.codinghorror.com/why-cant-programmers-program/
>>109959490>fangIt's still fagmanAnd so are you.
>>109960658The complexity of the hiring process reflects the increase in interested applicants, you're just complaining about a buyer being selective in a buyer's market, demand for solutions is entirely separate to that
>>109959490>work for fangIs this why windows and macos keeps getting exponentially buggier with every new release?
>>109954391>be me>*gives a niche open-ended problem to AI*>I'm now in the frontier of research, science at my command, 10000k starting
