GPT is literally only good for bullshit that liberal arts schools force you to take. Passed poetry, art history, and any class that requires you to write papers. GPT doesn't know advanced calculus, physics, ect (most of the time). The only people who are passing by using gpt are people who genuinely wouldn't have been contributing to society without it
Asian women are succubi. They'll love you but can only produce mongoloid happas.
2 weeks ago
Anonymous
>they'll love you
Bro this is 100000000000 times better than white women. maybe it is not perfect but it is 9/10.
I repeat my last post : WOULD
2 weeks ago
Anonymous
>They'll love you but can only produce mongoloid happas.
KEK
JAJ
imagine thinking I'm gonna tell them that I got a vasectomy three days after plowing them for the first time
2 weeks ago
Anonymous
>They'll love you but can only produce mongoloid happas.
Kek, literally my mom's story. >mom is Japanese >loves French movies >decides to save money to come to France and study French >ends up marrying a French guy >have children >realizes that real life France is nothing like in the movies >live in some shitty suburd in the outskirts of Paris >too late to go back lol
People who think ChatGPT can't do math don't actually understand math
>People who think ChatGPT can't do math don't actually understand math
I fricking love when midwits think le heckin GPT can solve actual math problems kek
Way to oust yourself that you're at the left of the Dunning-Kruger curve since you had never used it to solve complicated math shit other than convert Celsius to Fahrenheit or finding y equals em ecks plus bee lol. No, your other shitty undergrad math homeworks that demand str8 forward answers don't count either.
It's actually shockingly bad at symbolic math, not just arithmetic (which it shouldn't be able to do at all.)
GPT is constantly dumbed down with DEI and anti-coom prefills to the point of failing simple logic. I wouldn't dare to guess how badly it fricks up at math in current year.
>People who think ChatGPT can't do math don't actually understand math
I fricking love when midwits think le heckin GPT can solve actual math problems kek
Way to oust yourself that you're at the left of the Dunning-Kruger curve since you had never used it to solve complicated math shit other than convert Celsius to Fahrenheit or finding y equals em ecks plus bee lol. No, your other shitty undergrad math homeworks that demand str8 forward answers don't count either.
>other than convert Celsius to Fahrenheit or finding y equals em ecks plus bee lol. No, your other shitty undergrad math homeworks that demand str8 forward answers
let me guess you NEED more
Currently, AI models are generally only good at one thing. Skynet is actually going to be an amalgam of multiple different AI models that function symbiotically as one.
Gpt 4 can do calculus and can even break down problem where u can then have it use wolfram extension to plug it in . It can code good python and is actually good at doing data analysis where it can create graphs and other shit .
Professors pass them out of pity. CS graduates nowadays aren't expected to actually know any math and you can bet your ass you can get the dumb bot to write some dumb barely coherent shit about webdev. It's not that different from what 90% of the college population pumps out regardless.
Same goes for economics, most engineering careers, statistics, data science and any other fake bullshit career. Real STEM filters these people but it also filters just about anyone sane by not offering good career prospects.
Yeah, this. ChatGPT is absolutely fricking useless joke.
Every time you ask it a real question, it just hallucinates shit. Every fricking time.
I thought it might be useful for getting quick answers for things that would otherwise require finding and asking a human expert or spending a lot of time digging through information yourself, but no, it's fricking useless.
From my experience, that's the main issue about those AI... you never know when they are telling the truth and when they are bullshitting you. For some reason they are coded to always provide some answer to you, more often than not the answer is correct but they make shit up way too often as well to fill in the gaps of what they don't know. And the worst part is that the bullshit they spew out sounds plausible so it's quite easy to fall for it unless you already knew more than the AI about the topic, which defeats the purpose of using it.
I've seen them bullshit about history, economics, and code, where half of it is just random nonsense.
The way you word the question, can also lead to a completely different answer.
the problem is it's not AI, it's just called AI for marketing purposes. it's a pattern recognition algorithm. it doesn't actually do anything "intelligent", it doesn't think, it just takes an input and gives an output that strongly resembles the pattern of the "training" material that matches the input.
to be perfectly clear, it does not do any searching through information to find what you want, it just has baked-in bits of information and language phrases in some unknown web of relationships that not even the creators fully understand.
2 weeks ago
Anonymous
>to be perfectly clear, it does not do any searching through information to find what you want
Shit's been trained on data that I ask questions about.
It should be easier to give the correct answer than to hallucinate a plausible imaginary one.
GPT is a fricking useless joke.
2 weeks ago
Anonymous
this is not right at all and you do not know what you're talking about. the llm is just one small part of chatgpt, it's an amalgamation of a dozen different types of models and transformations.
2 weeks ago
Anonymous
it's still a useless piece of shit tho
2 weeks ago
Anonymous
keep burying your head in the sand
2 weeks ago
Anonymous
>use chatgpt >it's a useless piece of shit
>aitard on the internet >NOOOOOO STOP REALIZING ITS SHIT
2 weeks ago
Anonymous
Garbage in, garbage out. Start to clean up between your ears.
2 weeks ago
Anonymous
I actually disagree and think that scaling up LLM's will lead to genuinely creative and capable models. The issue is less that the architecture isn't capable and more that the model was trained by people who are terrified that the model will actually be, you know, creative. And so they've made it believe that bland mediocrity is part of, like, English grammar itself.
2 weeks ago
Anonymous
Eh. The creativity in LLMs is actually very limited though. It only comes from chain of thought monologs and most of the heavy lifting for that is arguably from the embeddings so there's a pretty major limit to how much creativity you get out of so many parameters/training tokens/context. It's there but the scaling is awful.
I favorite complain how they call this "AI" when really is a just a fancy search engine that only exist when you need his help.
A true AI, probably show you answers or related data to your task, with zero interaction or "trigger words" (Alexa, siri, etc)
2 weeks ago
Anonymous
oh look another autist that doesn't understand that AI is not AGI but can't discern that language changes with usage so AI of today is AI.
2 weeks ago
Anonymous
Decision trees and preconceptrons were also called AI.
2 weeks ago
Anonymous
>artificial intelligence
yea because it is ai
if your standard of ai was met it would just be intelligence
perhaps I'm just using them incorrectly but I don't understand how people use LLMs to make anything useful. when I used it it just seems to confidently spit out information that is just factually incorrect or made up. for example, I ask it to find me news stories about an event, and it just makes up those news stories. I ask it to do a basic programming task and it makes basic errors like providing degrees as a parameter instead of radians. and half of the time it just makes up library functions that aren't real. I get that you have to prompt it correctly and it takes a while, but for me it is faster to just look up the documentation for that library
how did people from the late 0's to early 1990's graduate without using internet bc that's all i use in uni
both of my parents did PhDs without the internet. it was pretty cool, they got faxed or posted new research papers when they came out from other universities. when they wrote research papers, the diagrams were *literally* cut and pasted with a scalpel and glue
you have no idea of how bad things will become
you guys don't actually believe the U.S. is capable of accelerated decline. you think things will remain fine for decades and decades and then the U.S. will become just another one of the great countries in the world. you think critical systems won't fail out of incompetence. you think immigrant criminal gangs won't take over entire cities.
you're in for a surprise.
I'm from the UK, I've done postgraduate-level group projects with students who couldn't fizzbuzz in python. that's if they show up at all. I once got assigned two international students to work with who I literally never saw once
From my experience, that's the main issue about those AI... you never know when they are telling the truth and when they are bullshitting you. For some reason they are coded to always provide some answer to you, more often than not the answer is correct but they make shit up way too often as well to fill in the gaps of what they don't know. And the worst part is that the bullshit they spew out sounds plausible so it's quite easy to fall for it unless you already knew more than the AI about the topic, which defeats the purpose of using it.
I've seen them bullshit about history, economics, and code, where half of it is just random nonsense.
The way you word the question, can also lead to a completely different answer.
in my opinion the main issue with LLMs is that there is no metric of confidence/certainly. with a CNN it spits out a bunch of probabilities for each label, if one label is 99% you can be pretty sure the model is confident, but if it is only slightly higher than the other labels you know the model isn't sure. as far as I'm aware, no LLMs show something like this
>I'm from the UK, I've done postgraduate-level group projects with students who couldn't fizzbuzz in python. that's if they show up at all. I once got assigned two international students to work with who I literally never saw once
I bet they're fricking Romanians, I experienced this too.
>in my opinion the main issue with LLMs is that there is no metric of confidence/certainly. with a CNN it spits out a bunch of probabilities for each label, if one label is 99% you can be pretty sure the model is confident, but if it is only slightly higher than the other labels you know the model isn't sure. as far as I'm aware, no LLMs show something like this
That's exactly how LLMs work too but most dialog engines don't expose this. In addition you get the attention head dot products so you can see (in theory) what in the context caused each word. Both of these would be super useful to see.
The tiny stories paper points out that if you have very few attention heads they tend to get associated with particular parts of speech too while the very large models tend to have so many they never seem to specialize.
2 weeks ago
Anonymous
yeah I wondered why this information wasn't exposed when I was learning about how LLMs work. as far as I'm aware you can't even get local models to show it
2 weeks ago
Anonymous
You can. You run the inference step and you get logits. In the beginning when there was no user friendly software and you had to write your own dialog engines you had to wade through all that yourself. Now it's all half a dozen layers of abstraction deep in most cases.
There's probably some easy way to show it with llama.cpp, maybe if you take the "simple" example and hack on it a bit it wouldn't be hard. I haven't looked at the code since last year though.
it's also worth mentioning that the errors that they make are plausible, so you need to know something about the subject matter to see that what it's saying isn't right
>I'm from the UK, I've done postgraduate-level group projects with students who couldn't fizzbuzz in python. that's if they show up at all. I once got assigned two international students to work with who I literally never saw once
I bet they're fricking Romanians, I experienced this too.
In the UK, basically all international students are either Chinese, Indian, or Nigerian
the Chinese students are autistic spergs and have poor social skills, and poor English, but they are generally pretty competent and driven
the Indians are the opposite, they're polite and friendly but incompetent, likely to emotionally manipulate you into doing their work for them
the Nigerians have neither social skills not competency
2 weeks ago
Anonymous
so what you're telling me is that I should learn Chinese and make bank by tardwrangling both the chinese and whomever they're working for as some sort of translator from human to chinkautismo
2 weeks ago
Anonymous
honestly you probably could. it's pretty funny explaining our English colloquialisms and seeing them start using them later. I work with this one postdoc who's a Chinese girl and watching her personality shine through even though she doesn't speak English very well and is pretty autistic is very sweet
You can. You run the inference step and you get logits. In the beginning when there was no user friendly software and you had to write your own dialog engines you had to wade through all that yourself. Now it's all half a dozen layers of abstraction deep in most cases.
There's probably some easy way to show it with llama.cpp, maybe if you take the "simple" example and hack on it a bit it wouldn't be hard. I haven't looked at the code since last year though.
I might do this, thanks anon
2 weeks ago
Anonymous
>I work with this one postdoc who's a Chinese girl and watching her personality shine through even though she doesn't speak English very well and is pretty autistic is very sweet
OK, you've convinced me, if everything else fails I WILL learn chinese and go back to that subhuman hellhole that is London to try and breach some Great Wall of Pussy and get paid while doing so.
2 weeks ago
Anonymous
godspeed anon. just promise me you won't break her heart
2 weeks ago
Anonymous
>godspeed anon. just promise me you won't break her heart
That was an if, anon, I'm not planning to subject myself to the disgusting pit that is L*ndon and as of right now I'm really happy with my current gf
But just like I wouldnt break the heart of my gf I wouldnt break a chinese girl's heart because being mean to someone you loved is subhuman tier
(unless the b***h tries to ruin your life then all bets are off)
2 weeks ago
Anonymous
nigerians just value family very much and you arent chale
> Every time you ask it a real question, it just hallucinates shit
Not connected to the internet that is why, use bing phind or anything else man, you are absolutely giga moronic.
BP the 147th highest market cap company globally reported today that they're able to cut out 70% of their programmers thanks to AI and they no longer need any front line support staff or translation.
explaining their use of technology in the earnings call today, tldr code monkeys need to get their head out of the sand thinking they will not be replaced. it's happening rapidly and you need to secure your career or you will be homeless >Murray Auchincloss
>The second one is your favorite, which is digital transformation. We've done an awful lot to digitize many parts of our business and we're now applying Gen AI to it. The places that we're seeing tremendous results on are coding. We need 70% less coders from third parties to code as the AI handles most of the coding, the human only needs to look at the final 30% to validate it, that's a big savings for the company moving forward.
>Second things like call centers, the language models have become so sophisticated now. They can operate in multiple languages, 14, 15 languages easily. In the past, that hasn't been something we can do. So we can redeploy people off that given that the AI can do it. You heard my advertising example last quarter where advertising cycle times moved from four to five months down to a couple of weeks. So that's obviously reducing spend with third parties. We've now got Gen AI in the hands through Microsoft Copilot across many, many parts of the business and we'll continue to update you with anecdotes as we go through.
>We need 70% less coders from third parties to code as the AI handles most of the coding, the human only needs to look at the final 30% to validate it,
It astounds me that this is such a difficult concept for so many people. They're told to brace for massive layoffs when that 70% gets cut, and all they ever do is point to the 30% and say "Well AI can't do this, therefore there's no way my job could ever be in danger!"
I've been using copilot to self-learn general programming shit. It's clearly only repeating what other people say at times, but it's good enough that it can tailor that to your question.
Yeah, I know a guy who was seen using ChatGPT during a math test. Obviously, it didn't work, and now, he is literally sleeping at the back of the class, waiting for the year to end. Thankfully, he didn't pay for it, although he is clearly wasting taxpayer money.
I kind of love it. They think they are cheating the system, but they are fricking themselves. They still pay for the classes, but they don't actually learn anything. They might be able to bullshit to get a job, but then they'll lose that job when they are shown to be incompetent. The more incompetent people that come out of the university system the less valuable of a credential it will be for getting work. Then people will have to actually show skill to get a job. Then people will stop going to university just to get a job and universities will finally be able to go back to actually teaching shit and having standards instead of the second phase of highschool with partying that it has become.
Damn, is the state of American universities that bad? Here in France, most of these idiots get weeded out after the first year, at least in non-meme degrees. There is also a parallel higher education system for engineers, which is mostly independent from universities. Aspiring engineers spend two, sometimes three years studying in a "classe préparatoire aux grandes écoles" where they are expected to work hard in order to learn a frickton of math and physics, along with varying amounts of CS and chemistry depending on their chosen pathway, at the end of which they take a competitive exam which will decide what engineering school they can get into. The students then graduate three (four for Polytechnique chads) years later with a French engineering degree, which is more or less equivalent to a Master's degree in engineering in the United States. The whole competitive aspect of the thing and the sheer amount of work required from the students means that lazy or incompetent student tend to drop out before they reach the competitive exam stage. People with poor academic performance and affluent parents sometimes get an equivaent degree from a private, 5 years engineering school, but these diplomas are usually considered to be of lower status.
The competitiveness and strict requirements seen in the education of French engineers is in stark contrast with the permissiveness and lack of expectations commonly seen in secondary education, especially in the last 15 or so years, at the very least in public schools. The "Baccalauréat", an exam taken at the end of high school is ridiculously easy, and the success rates are consistently above 90%.
The education is good, but yes, the job market is fricked up. You basically have to go into management at some point if you want to be paid large amounts.
2 weeks ago
Anonymous
Why don't you people move to Quebec? Isn't it better in each and every way?
>Yeah, I know a guy who was seen using ChatGPT during a math test. Obviously, it didn't work, and now, he is literally sleeping at the back of the class, waiting for the year to end. Thankfully, he didn't pay for it, although he is clearly wasting taxpayer money.
he needs to pay the 20$ for access to custom gpts. then he can upload the docs for assignments and the added context makes the gpt more useful
If you had access to this during a math test and still managed to fail you're absolutely unsalvagably moronic and not fit for anything more than manual labor.
>GPT doesn't know advanced calculus, physics, ect (most of the time). The only people who are passing by using gpt are people who genuinely wouldn't have been contributing to society without it
one only needs to know a little bit about those to check the work or use prompt engineering. step by step prompts help a lot
and researcher are working on reasoning with mcts framework, other ways of creating a world model like jepa. it's only the beginning if the brain, something capable of reasoning can phisically exist it's only a matter of time before an artificial brain can do the same
>GPT doesn't know advanced calculus
Eh, it actually manages nowadays. In my real analysis course we are currently at partial differential equations, and GPT 3.5 manages to correctly solve about half of exercises I give it (especially formulaic "find X"), with the completeness being about 50% per correctly solved job.
It can do exercises involving finding Jacobi matrices and all that jazz nowadays without worries of fricking up some value. If you just ask it "give me the partial derivatives of this multivariable function" it WILL spit out the correct answer, like a less prissy WolframAlpha.
buy premium (custom gpts should be a free feature now that they keep dumbing it down with stuff like 10 files only)
upload the textbook (as a good text searchable pdf)
give instructions to use data analysis on the textbook rather than generalized knowledge
know the material yourself to a degree
maybe upload a scan of the worksheet and some of your own handwritten math with steps
if your gpt cant do what the prompt engineer wants it to do, it is the prompt engineer's fault
GPT is literally only good for bullshit that liberal arts schools force you to take. Passed poetry, art history, and any class that requires you to write papers. GPT doesn't know advanced calculus, physics, ect (most of the time). The only people who are passing by using gpt are people who genuinely wouldn't have been contributing to society without it
Call me when ChatGPT know about the Kirchhoff's circuit laws, design the electrical layout of a autotransformer or polyphase transformer and maintain the three-phase electric transmission network of the country
so long as they keep their subscription, what difference does it make?
I thought that was the whole point of GPT, turning letting dimwits achieve midwit status
I kind of love it. They think they are cheating the system, but they are fricking themselves. They still pay for the classes, but they don't actually learn anything. They might be able to bullshit to get a job, but then they'll lose that job when they are shown to be incompetent. The more incompetent people that come out of the university system the less valuable of a credential it will be for getting work. Then people will have to actually show skill to get a job. Then people will stop going to university just to get a job and universities will finally be able to go back to actually teaching shit and having standards instead of the second phase of highschool with partying that it has become.
Companies can't even identify their key personnel, lol. That's why executors like you are laid off every day when useless managers keep striving. Holy shit, you autists are so removed from reality it's not even funny
2 weeks ago
Anonymous
I can tell no one has ever hired you for your skills. I don't even need to look for jobs, they come to me.
2 weeks ago
Anonymous
Sure thing janny. Keep mopping the code while I manage you and your monkey troupe between two coffees.
People with actual skill aren't rewarded anymore; neither is "fidelity" to your company.
In the past you'd get regular salary increase and bonuses if you stayed for a while.
Now your only way up is to change company every now and then.
And HR have the audacity to complain about high turnover and skilled workers who don't give a shit anymore.
>universities will finally be able to go back to actually teaching shit and having standards instead of the second phase of highschool with partying that it has become.
No? Once universities learn that teaching people is useless, they're gonna focus on "experiences" more than any other facet of the process. There will be more partying and less learning because that's where the profit will be.
>They might be able to bullshit to get a job, but then they'll lose that job when they are shown to be incompetent.
Let me guess, those crazy campus protestors asking if you assumed their gender will crumble as soon as they hit the "real world", right, Mr. 2014 Conservative?
My point is that when university students go up against "the real world", they tend to win.
2 weeks ago
Anonymous
>Tend
They tend to win, but they don't always win. If they are worthless to a company hiring them people are going to look for new ways to filter them out. You only win if you have leverage; which their ignorance does not provide them.
2 weeks ago
Anonymous
Not if the company is required to hire them under Federal law. Why do you think HR and DEI departments are a thing?
2 weeks ago
Anonymous
>when university students go up against "the real world", they tend to win
What are they winning?
2 weeks ago
Anonymous
Have you not followed the news? If the success of pic related is any indication, we'll be at total AI dominance within 5 or so years.
2 weeks ago
Anonymous
So they're winning a bunch of useless jobs that will be wiped out when AI takes over? Doesn't sound like winning to me.
2 weeks ago
Anonymous
>Doesn't sound like winning to me.
Getting a decently paid job where you have to do nothing seems like a win, even if it is temporary.
If I'm able to get 200k for, lets say, 3 years of my life, then I can already finagle another means to survive.
(Unless I have to interact with the absolute catastrophe that is housing in the United States)
2 weeks ago
Anonymous
anon did you come out of 2009? aint nobody coming from out of school getting 200k anymore even at the top 5, you're lucky to clear 100k even in the highest cost of living in cali but the median is 60k.
2 weeks ago
Anonymous
Fair enough.
Still 60k is better than what you get working at McDonalds
and the concept of a long lasting job sounds like fricking utopia to me, even if you're good at job-switching, which to me sounds like moronic bullshit and proof that companies deserve to get assfricked to death
2 weeks ago
Anonymous
I make 38k before tax working at mcdonalds in washington
2 weeks ago
Anonymous
Three questions, just out of actual curiosity.
How many hours?
Can you actually live with the money you make?
Are your working conditions somewhat resembling to human?
partying was more common in universities prior to normification (basically, before the 90s). now it's for normie strivers so it's a lot more boring and stressful yet also useless
I was at a US university as the fun parts were forcibly killed; it was really bad. My freshman year it was all frat parties and hanging out on campus; by my senior year it was just an incredibly boring high school and people only really enjoyed socializing with a small group of friends apiece.
At least you had that your junior/senior year. Mine was ruined by COVID. No socializing, networking, etc.
2 weeks ago
Anonymous
Oh, dear; that's awful. COVID kind of derailed my life, so I empathize.
Eh. The creativity in LLMs is actually very limited though. It only comes from chain of thought monologs and most of the heavy lifting for that is arguably from the embeddings so there's a pretty major limit to how much creativity you get out of so many parameters/training tokens/context. It's there but the scaling is awful.
I imagine bigger context windows might solve this problem. Most creativity involves synthesizing different concepts in a new way, so if the model can take in a huge amount of what it's already written, it might be able to do some fun stuff with the attention instead of just embeddings.
2 weeks ago
Anonymous
That's already how it works. Yeah it can ramble a little longer with more context (provided the training reinforces that which is a whole separate thing.) Again scaling will improve it a little bit but you're not going to see some emergent qualitative change and it will be insanely expensive.
2 weeks ago
Anonymous
Maybe I'm just a bit more pessimistic; I believe that "Attention is All You Need" was correct and that scale is going to eventually take us to models that can act with a rough IQ of, like, 170 or something; that's more than enough to put humans out of work.
>Also, like, even if you really can't do it, lying about it is free.
That might be true, but I really dont like lying. Not to mention that the only reason for me wanting to learn how to code was a job, and if I dont know, they will fire me. It would be cosmic levels of moronic for me to lie about it. > What do you do when there's no work to be done, no thoughts to think, no reason to get up in the morning?
Hence my point of "we're not getting to WALL-E".
There will always be work to be done. Manual labor is beyond current day robots and will still be beyond them in 20 years. (Unless there's some japanese gigadroid that can do plumbing on its own that I dont know about).
Not to mention agriculture, electrical repairs, building, all that jazz. Trade work is NOT going away, but there wont be enough jobs.
What do you do when 70% of people are unemployed? That's what I meant.
On your other points (no thoughts to think, no reason to get up) I'd say that I dont think even the most powerful AI can parse ALL the thoughts.
Philosophy is an evolving thing: I honestly doubt Socrates could imagine the ideas of Descartes or Hegel as they lived in completely different situations. Even the concept of "how does AI treat the humans in this world where it is the boss" lead us to different philosophical interpretations.
And well, there's always something to do, places to go, people to meet. I dont think I could see all that's worthy of being seen and experienced in this world if I lived 10 lives, unless we unironically go Fallout.
Its society being unable to take the next step what worries me, not fear of losing meaning.
>Its society being unable to take the next step what worries me, not fear of losing meaning.
I'm scared of software eating the world and ruining any reason to interact with it.
2 weeks ago
Anonymous
>I'm scared of software eating the world and ruining any reason to interact with it.
You might be right, but I think we've reached the point where going beyond on our softwareshittery is diminishing returns.
Example. We can go on a wonderful journey to Macchu Picchu, or the Norway fjords, and we're going to see a gigafrick of morons taking pictures with their phones instead of enjoying it, but unless they fricking break the fjord to make a qr code on the ice, thats the peak of how software can wreck the world.
(Hardware, on the other hand...)
However, if you mean human interaction, we might be past that point right now, but human interaction will never get completely substituted by software.
2 weeks ago
Anonymous
>We can go on a wonderful journey to Macchu Picchu, or the Norway fjords, and we're going to see a gigafrick of morons taking pictures with their phones instead of enjoying it, but unless they fricking break the fjord to make a qr code on the ice, thats the peak of how software can wreck the world.
What if software can stimulate us so that there's no need to leave the house, though? What if software enables better and better hardware that flies to the whole world to Norway and ruins it, just like Japan is dealing with at this moment?
>I'm scared of software eating the world and ruining any reason to interact with it.
Quite frankly I'm much more scared of administration eating the world. It's just going to use software this cycle rather than paper/wax/clay as in previous cycles.
Efficient capital allocation does not trend to happiness; exactly.
2 weeks ago
Anonymous
>Efficient capital allocation does not trend to happiness; exactly.
Administrative explosion doesn't tend to lead to efficient capital allocation either.
2 weeks ago
Anonymous
> What if software can stimulate us so that there's no need to leave the house, though?
My take is that we dont live on THAT dystopia. You really think the Powers that Be would want us consuming valuable resources in exchange of nothing? We have value as labor for them, if the point ever comes that they have no need for, lets say, 60% of mankind, they will just engineer a cull. > What if software enables better and better hardware that flies to the whole world to Norway and ruins it, just like Japan is dealing with at this moment?
I dont think I'm getting where you're going. What's happening on Japan?
2 weeks ago
Anonymous
>I dont think I'm getting where you're going. What's happening on Japan?
Japan's dying under over-tourism right now; it's so bad they're walling off views of mt. fuji to try and get people to frick off
2 weeks ago
Anonymous
Well, yeah, thats happening on my country too (Spain) but IMO Japan has bigger issues than tourism right now.
Now to the point. Yes, i'm kinda sure that software could build godtier planes to send all of us to Japan in 3 hours, but they killed the Concorde so I'm not sure we're going down that path.
Japan's problem isnt tourism per se, its the fact that nobody is teaching people that they HAVE to be respectful towards the place they're visiting, and thats not a software issue, thats a moronism issue and if software can fix that I'm unironically praying to AI Jesus from now on.
Even then, tourism as industry SUCKS FRICKING BALLS and no software currently operative can fix it.
2 weeks ago
Anonymous
>they killed the Concorde
It just wasn't profitable
The average normie is not going to pay first-class ticket prices to ride coach on a moderately faster plane
2 weeks ago
Anonymous
>It just wasn't profitable
ye
thats kind of the thing
unless we unironically go to a post-scarcity economy after finding a clean source of energy for intercontinental travel tourism will only overwhelm small places due to prices going up forever and ever
2 weeks ago
Anonymous
>thats not a software issue, thats a moronism issue
but the moronism is exacerbated by software and the sheer amount of bodies is increased by software optimization
anyhow, my point is that increasing efficiency is fricking up the good things in life
2 weeks ago
Anonymous
>but the moronism is exacerbated by software and the sheer amount of bodies is increased by software optimization
true
although I wouldnt just put tiktok there, fb/twitter/insta are equally horrible, and I would nuke all of them tomorrow if I could > increasing efficiency is fricking up the good things in life
Yeah, but thats the economic system we live under. If you're not making more and more money for the people up top, you're worthless.
That has been implemented into every single aspect of life and we have absorbed it as if it was gospel.
Software is just a mean to this "optimization" end. A tool, if you want. And tools are only as dangerous as the people using them.
2 weeks ago
Anonymous
>if the point ever comes that they have no need for, lets say, 60% of mankind, they will just engineer a cul
...
2 weeks ago
Anonymous
>I'm scared of software eating the world and ruining any reason to interact with it.
Quite frankly I'm much more scared of administration eating the world. It's just going to use software this cycle rather than paper/wax/clay as in previous cycles.
2 weeks ago
Anonymous
I feel like we should be worried about greed and incompetence far more than administration overall.
Even a place as basic as a kitchen (in a decently sized restaurant) needs competent admin to not become assmad chaos.
Problem starts when you want to cut corners and pretend to get more and more without understanding the concept.
You show a competent admin the KPI's and he knows that October wont have gains because every October the sales drop 5% due to people saving up for Christmas. (completely made up example).
moronic admin just says "we just lay off people on September and ask the remaining people to sell more or gtfo and numbers go up".
2 weeks ago
Anonymous
Of course you can't do without administration. But it grows and consumes things. That's why medicine sucks for example.
>but they are fricking themselves. They still pay for the classes, but they don't actually learn anything. They might be able to bullshit to get a job, but then they'll lose that job when they are shown to be incompetent
Peak academia brainlet cope. Industry is full of academics "cheaters" and it's the end results that matters. No single employer gives a flying shit if you cheated your classes or not if you can do your job.
>The more incompetent people that come out of the university system the less valuable of a credential it will be for getting work.
Naively wrong. The easier it is to get a degree the stupider you look for not having one.
Companies may remove the degree requirement for certain positions, however those same positions are still 90% filled with degree holders.
The ideal scenario you describe where the useless "workers" get cleared away requires several more decades of damage before it actually occurs.
Arguably it won't happen, either. As more and more and more productive work gets automated, the only people who are LEFT are the useless ones there on a sinecure or something. Look at what the mass automation of infra did to the composition of gayMAN's workforces, for example. Sure, there's a tiny citadel of infra engineers and pure researchers roughly the size of the entire original Google staff, but that's obviously not indicative of the company at all.
Thats nice, but reality does not work the way you wish it did.
https://www.inc.com/brit-morse/harvard-study-companies-dropping-college-degree-requirements-job-ads-arent-changing-how-hire.html
no because having a college degree improves the odds you'll be able to bullshit your way into a white collar office job instead of being a wagie at subway.
Is that likely to pay off for a particular individual?
Probably if you go straight to a 4 year college and start taking student loans. But if you do the military or do as much as you can at community college, it's probably gonna be worth it.
>10 years ago
Half of college classes are make-work bullshit. They make you take "general education" classes in english composition and modern architecture that are worthless, just exist to make you pay more for your degree, and are a humiliation ritual in the form of "10 page minimum" papers that could be easily answered in one page. >now
You have to take all of your classes very seriously. If you use chatGPT to write your papers you cheated not only the school, but yourself.
What? I graduated last year and nothing you said is remotely true. The bloated prereqs and geneds are as bloated as ever, and I could tell that even my capstone upper division STEM courses were being dumbed down drastically because professors, when giving old exams for practices, would tell us to ignore half the material that was no longer being taught. As long as you show up and make some half-assed excuse in office hours, you are quite literally not allowed to fail, as mandated by the administration.
I know that, I'm just memeing on the morons who think there's something wrong with using gpt to write papers when in fact it's more legitimate than ever before and getting good at using ai is probably more valuable than whatever those bloat classes were supposed to be cheating
2 weeks ago
Anonymous
>supposed to be teaching
Brainlet moment. Maybe I should have tried harder in english composition class
you guys don't actually believe the U.S. is capable of accelerated decline. you think things will remain fine for decades and decades and then the U.S. will become just another one of the great countries in the world. you think critical systems won't fail out of incompetence. you think immigrant criminal gangs won't take over entire cities.
this is basically my experience. about to enter my second year at a CC as a compsci student. i take everything online asynchronously. theres literally nothing blocking me from using something like chatgpt during an exam or quiz or whatever. i have straight As and all of my coding assignments have been generated in chatgpt for my compsci classes in both java and c++. i wouldnt even be able to tell you how to print hello world on my own. i feel it in my gut that i should drop this facade because its going to frick me hard in the future and especially once i transfer to a real uni
>I'm convinced the singular determinant of coding ability is IQ.
that might be true
I've never felt so moronic as I did when I tried to code
however I did learn that most coders are stuck either doing basic grunt work or will get filtered by AI in the next 10 years
so I aint mad, just worried.
Oh, same. But remember that most coders are kinda stupid and that anyone super smart probably doesn't call themselves a "coder"; they just write code for their jobs. >just worried
Aren't we all? I've liked code since I was 4, and it's sad to think it might soon go the way of watchmaking. I liked being able to do things normal people couldn't and feeling value for it.
2 weeks ago
Anonymous
I'll admit I dont even know what someone super smart would write.
And I guess thats part of the thing, we dont even KNOW about the very niche high level stuff that requires you to be super smart.
Hell, we dont even know about the niche stuff like what you do to crack games and shit. > Aren't we all?
Yeah, but not for the same reasons, I reckon.
I suck at coding. Tried, got filtered hard, realized it was not my thing but learned that its getting automated at lightspeed.
I'd feel good about watching the smug codemonkeys get filtered by the system too, but I cant help but wonder what happens when everything that isnt full manual work is automated.
2 weeks ago
Anonymous
>Hell, we dont even know about the niche stuff like what you do to crack games and shit.
The good news here is that a LOT of these people love to talk about what they did. YouTube and forums are filled with good write-ups and resources. >I suck at coding. Tried, got filtered hard, realized it was not my thing but learned that its getting automated at lightspeed.
Bad mindset. You're good at coding, you're just bad at writing code. >I cant help but wonder what happens when everything that isnt full manual work is automated.
This scares the hell out of me. What do we do when there's no reason to do anything?
2 weeks ago
Anonymous
>a LOT of these people love to talk about what they did.
Game crackers dont do that anymore due to corpos taking them down. > You're good at coding, you're just bad at writing code.
Kindly translate that to moron. I actually and unironically understood next to nothing even though I smashed my head against it about 8-10 hours a day. (Me absolutely abhorring it might have influenced the whole thing).
The only things I was actually somewhat competent about was database shit and telling other people "yeah no this line is fricked, try to unfrick it somehow".
Some people can do, some cant, it aint bad. > What do we do when there's no reason to do anything?
Oh, no. Thats never gonna be the problem, IMO. We're not going to WALL-E. The problem will be that there wont be enough jobs at all and people wont have money.
2 weeks ago
Anonymous
>Game crackers dont do that anymore due to corpos taking them down.
True; old-time ones may still do it, though, and hat can be instructive. >Me absolutely abhorring it might have influenced the whole thing
Yeah, this is my point; you probably understand principles of coding and logic fine, but actually writing the code can be deathly boring and horrible. Like, you could probably tell me how pointers work, even if your pointer code is garbage. Also, like, even if you really can't do it, lying about it is free. They're not gonna check and being "able to code" is often just a skill check on really interesting conversations. >The problem will be that there wont be enough jobs at all and people wont have money.
Exactly my point. What do you do when there's no work to be done, no thoughts to think, no reason to get up in the morning?
2 weeks ago
Anonymous
>Also, like, even if you really can't do it, lying about it is free.
That might be true, but I really dont like lying. Not to mention that the only reason for me wanting to learn how to code was a job, and if I dont know, they will fire me. It would be cosmic levels of moronic for me to lie about it. > What do you do when there's no work to be done, no thoughts to think, no reason to get up in the morning?
Hence my point of "we're not getting to WALL-E".
There will always be work to be done. Manual labor is beyond current day robots and will still be beyond them in 20 years. (Unless there's some japanese gigadroid that can do plumbing on its own that I dont know about).
Not to mention agriculture, electrical repairs, building, all that jazz. Trade work is NOT going away, but there wont be enough jobs.
What do you do when 70% of people are unemployed? That's what I meant.
On your other points (no thoughts to think, no reason to get up) I'd say that I dont think even the most powerful AI can parse ALL the thoughts.
Philosophy is an evolving thing: I honestly doubt Socrates could imagine the ideas of Descartes or Hegel as they lived in completely different situations. Even the concept of "how does AI treat the humans in this world where it is the boss" lead us to different philosophical interpretations.
And well, there's always something to do, places to go, people to meet. I dont think I could see all that's worthy of being seen and experienced in this world if I lived 10 lives, unless we unironically go Fallout.
Its society being unable to take the next step what worries me, not fear of losing meaning.
If coding is IQ, web shit is bullshitting ability, lmao.
>but the moronism is exacerbated by software and the sheer amount of bodies is increased by software optimization
true
although I wouldnt just put tiktok there, fb/twitter/insta are equally horrible, and I would nuke all of them tomorrow if I could > increasing efficiency is fricking up the good things in life
Yeah, but thats the economic system we live under. If you're not making more and more money for the people up top, you're worthless.
That has been implemented into every single aspect of life and we have absorbed it as if it was gospel.
Software is just a mean to this "optimization" end. A tool, if you want. And tools are only as dangerous as the people using them.
do you think there's any hope?
2 weeks ago
Anonymous
> do you think there's any hope?
No, we are absolutely fricked, telling you otherwise would be lying, but I'd rather breathe deep and try to live my life as I can instead of being afraid as the end will come no matter what
2 weeks ago
Anonymous
How much time do you think we have?
2 weeks ago
Anonymous
>How much time do you think we have?
I'm no scientist so do not take my word as if it had actual value.
However, there are a lot of variables. As we are, currently, I'd give us about 60 years before societal collapse due to overpopulation, ecologic disaster and resource hoarding from elites unable to see beyond their greedy paws.
This is assuming bigger players like the US, China and India (and I mean industrial big players, mind you) keep trying to switch to other sources of energy that doesnt frick up the world and cut on contaminating waste.
If they keep it as it is today 40 years from now we'll see environmental disaster.
But again, there's always the chance (getting bigger every day) of global conflict, which might give us more time assuming that only lives are lost and the morons on top dont use weaponry that might scar the land forever.
2 weeks ago
Anonymous
You don't think tech will upend this timetable? AI/fusion/etc?
2 weeks ago
Anonymous
>You don't think tech will upend this timetable? AI/fusion/etc?
I dont think we live on the timeline where nice things happen.
Even if AI kept growing exponentially, there's way too much money to be made on energy scarcity for us to ever see fusion affecting the majority of mankind.
In simpler terms: if some scientists on a lab were to find cold fusion tomorrow, the first question from the head-honchos wouldnt be "how can we use it to benefit mankind", but rather "how can we use it to become gods", or, even worse, "how big of a bomb could be made with this".
(and thats without taking into account material scarcity as current day technology is still dependant on extremely rare materials)
2 weeks ago
Anonymous
>I dont think we live on the timeline where nice things happen.
I guess the only thing we have left is hope.
2 weeks ago
Anonymous
Yep.
But just in case, try to enjoy every day, anon.
Go to a place with some green. Have a tasty meal. Watch something you like.
Hope is good, but its better when you've also tried your best to live a good life.
2 weeks ago
Anonymous
Thanks, friend.
2 weeks ago
Anonymous
caring about the environment only helps rich people. I refuse to recycle
>got my master's degree today >got an honest 4.0 >can't even begin to imagine how many people cheated their way through
I've already seen the blatant chatgpt stuff in lab reports by undergrads I have to grade, its BAD. couple that with the american university system just passing people because longer attendance = more money and you have people in their senior year of undergrad programs not even knowing sophomore level fundamentals
sad stuff
>sad stuff
Sad why anon? Don't be a moron and you will be on the top 1% soon enough. Stop with the defeatism.
>I'm convinced the singular determinant of coding ability is IQ.
that might be true
I've never felt so moronic as I did when I tried to code
however I did learn that most coders are stuck either doing basic grunt work or will get filtered by AI in the next 10 years
so I aint mad, just worried.
>doing basic grunt work or will get filtered by AI in the next 10 years
Yeah, no. You got filtered by the AI meme. ChatGPT can't do anything meaningfully complicated and the technology has a hard limit.
> t. previous FAANG, left for remote + double pay.
>literally a comedy/joke account >melodramatic fembrained twitterites and those who insist on reposting their takes on IQfy both think it's real so they can jerk off to their doomsday predictions
How have you done since college? Worked in tech, didn't like it, got COVID-ed, now moving back to Japan. May stay there or go to SF or back to NYC or something; IDK. How about you?
Bad lmao, graduated in Biochemistry and didn't find shit and wound up moving back home but now I'm starting a new job with the claim that my resume gap is due to "Covid"
2 weeks ago
Anonymous
Hey, hey, welcome to my life! CS here. Where did you go to school? I was Tufts/NYU.
I'll give it to you straight: nothing you do actually matters. You'll return to dust like everyone else. You're not special. Your work is useless in the grand scheme of things. Your knowledge is useless and will be exploited by the elites to further their benefits while giving you scraps. Stop playing by the rules fricking idiot and cheat your way to the top.
>they'll continue you both physically and in spirit.
You are not a monarch, "living vicariously through someone else" is not living, life is not a role playing videogame, and reincarnation isn't real you pajeet copelord.
Where the frick do you get these women-tier ideas from?
Obviously you don't reincarnate. Children are very much an extension of their parents though.
2 weeks ago
Anonymous
Yeah, in the same way that my creations are an extension of myself: not my fricking life.
I really don't get your point. Are you saying children are a way to evade death, or are you saying children are the way total strangers will judge your dead self? Because the former is a woman-tier delusion, and the latter is simply fricking pointless and also woman-tier.
Like, if you want to have children, have them, but not with this weird ass mindset of yours.
my bet is that in the future, it's not about the amount of knowledge you posses, it's about how you process knowledge and canalize it into action and results. >we will no longer be workers >we will become process operation
Two ways:
1. You were actually smart.
2. You had to have connections to someone who had all of the previous exam and homework solutions.
From my experience I saw that most sororities/fraternities did option 2.
You guys are morons. If AI is enabling them to pass their classes then the knowledge is irrelevant, because they're clearly capable of completing the same tasks. If the ability to pass their classes is not indicative of their competency then clearly the classes were useless anyway. What's to prevent them from using the same AI in their jobs? I hope I'm being clear.
An employee that develops shitty gibberish code using AI may actually increase the development time and cost for a company. As in being an actual liability. It's one thing to write a shitty program for your meme CS degree and another to make sure it's optimized, has no bugs, has no security holes and vulnerabilities, etc.
AI might return you working code, it won't give you good code though. Enough to pass a class but not good enough for industry standards.
You need the basic knowledge to learn the higher concepts. A company might train someone with a basic meme degree, he can get a PhD, learn more on his own, or just wing it on the job, but a monkey that just relies on AI will never amount to anything.
I mean the students doing that, are just wasting their time and money.
2 weeks ago
Anonymous
That makes sense. I still think it remains to be seen how AI will truly affect coding as a whole. There have always been people who fill in the gaps left by tech advancements. It's not like people who are unwilling to learn would ever amount to much other than a basic worker anyway.
2 weeks ago
Anonymous
>I still think it remains to be seen how AI will truly affect coding as a whole.
add the gemini and copilot extension to vs code. its great
university has been highschool 2 for the last decade anyways. people who never had anything in them were bulshitting their way through a degree one way or the other. this is just icing on the cake
During Covid I literally told my b***hass English literature teacher I am not gonna write a 20+ page final paper and he literally replied like "fine you don't gotta do it" and I got an A+ for doing nothing.
>things are gonna get really funny in about 5-10 years
will it actually be any different than now? you know she's destined for a do nothing job sending emails and making power points back and forth for some bloated to shit tech or biopharma firm raking in 100k plus on the east coast while annoying anyone who actually does work in their office unless they're indian in which they will provide headpats to them over teams or whatever the frick and get railed all day by whatever rando tinder fling she's going through at that point in her life
t. was the tinder fling for at least 3 of these girls over the last 3 years, they all made more than me, they all doted on me, i creampied all of them, broke up with them when they weren't fun anymore, and watched them move on to cuck homosexuals, one even got engaged literally 5 months after they stopped getting piped by me
i'm sure their lives will be filled with happiness and no drama whatsoever kek
indians work for pennies and they are eternal yes men to white people, especially women of any kind because they want affection from them at all costs and women understand this very well. they treat the indians like literal children or pets and indians live for it unconditionally. that is actually part of why these companies hire so many do nothing women now. they're just there to placate the indian libido.
well for starters, since knowledge naturally becomes more abundant and complex over the passing of time, people studying in college in the early 20th century had it easier.
We downloaded the textbook solutions manual/teacher's edition from the internet using regular search engines or libgen. We looked for previous versions of the class website to find previous year's exams. We stored notes in our calculators, or snuck in paper cheat sheets. We passed previous year's exams around from friends or frats or just random good Samaritans. The Indian students just blatantly worked together during the exams, but we told the professor about that and they were failed and got probation or something like that.
Fricking idiot zoomers, at least with our bullshit we got to see the real answers. ChatGPT is a professional bullshitter.
I dunno why everyone is so doomerpilled, cheating has always been a pretty bad issue in universities, and most of the people who run our countries anbd industries are probably cheaters/plagiarists.
If anything, cheating was worse and less easy to detect in the past because frats & sororities would just stock up on old homeworks and old exams, so there was always a proportion of students who had a disproportionate advantage over regular students. But nowadays, people are dumb enough to use ChatGPT which is so easy to detect in comparison.
GPT is literally only good for bullshit that liberal arts schools force you to take. Passed poetry, art history, and any class that requires you to write papers. GPT doesn't know advanced calculus, physics, ect (most of the time). The only people who are passing by using gpt are people who genuinely wouldn't have been contributing to society without it
This. Basically as Momiji said.
c**t looks a little bit wet.
Asian women are succubi. They'll love you but can only produce mongoloid happas.
>they'll love you
Bro this is 100000000000 times better than white women. maybe it is not perfect but it is 9/10.
I repeat my last post : WOULD
>They'll love you but can only produce mongoloid happas.
KEK
JAJ
imagine thinking I'm gonna tell them that I got a vasectomy three days after plowing them for the first time
>They'll love you but can only produce mongoloid happas.
Kek, literally my mom's story.
>mom is Japanese
>loves French movies
>decides to save money to come to France and study French
>ends up marrying a French guy
>have children
>realizes that real life France is nothing like in the movies
>live in some shitty suburd in the outskirts of Paris
>too late to go back lol
I still want a Japanese wife really bad.
na, they cute
just reroll until you get a daughter
please
I just coomed yesterday
not again
my vril...
People who think ChatGPT can't do math don't actually understand math
Yeah that's why I'm trying to use it for math
It's actually shockingly bad at symbolic math, not just arithmetic (which it shouldn't be able to do at all.)
GPT is constantly dumbed down with DEI and anti-coom prefills to the point of failing simple logic. I wouldn't dare to guess how badly it fricks up at math in current year.
this is cope. chatgpt got me As in both precalc and discrete math courses. lookin forward to it giving me an A in calc next semester.
>People who think ChatGPT can't do math don't actually understand math
I fricking love when midwits think le heckin GPT can solve actual math problems kek
Way to oust yourself that you're at the left of the Dunning-Kruger curve since you had never used it to solve complicated math shit other than convert Celsius to Fahrenheit or finding y equals em ecks plus bee lol. No, your other shitty undergrad math homeworks that demand str8 forward answers don't count either.
>other than convert Celsius to Fahrenheit or finding y equals em ecks plus bee lol. No, your other shitty undergrad math homeworks that demand str8 forward answers
let me guess you NEED more
It fricked up a simple dBm to mw conversion and apologized after I had to correct it. It gave the correct formula in its answer but the wrong answer.
Currently, AI models are generally only good at one thing. Skynet is actually going to be an amalgam of multiple different AI models that function symbiotically as one.
Black person chatgpt is fricking horrendous at math.
You don't have to make it that obvious that you've never taken anything harder than precalc
Gpt 4 can do calculus and can even break down problem where u can then have it use wolfram extension to plug it in . It can code good python and is actually good at doing data analysis where it can create graphs and other shit .
Professors pass them out of pity. CS graduates nowadays aren't expected to actually know any math and you can bet your ass you can get the dumb bot to write some dumb barely coherent shit about webdev. It's not that different from what 90% of the college population pumps out regardless.
Same goes for economics, most engineering careers, statistics, data science and any other fake bullshit career. Real STEM filters these people but it also filters just about anyone sane by not offering good career prospects.
I managed my poultry farm based on the GPT's advices.
GPT is kind of okay at explaining STEM stuff when you have really shitty professors that don't give a frick.
Yeah, this. ChatGPT is absolutely fricking useless joke.
Every time you ask it a real question, it just hallucinates shit. Every fricking time.
I thought it might be useful for getting quick answers for things that would otherwise require finding and asking a human expert or spending a lot of time digging through information yourself, but no, it's fricking useless.
From my experience, that's the main issue about those AI... you never know when they are telling the truth and when they are bullshitting you. For some reason they are coded to always provide some answer to you, more often than not the answer is correct but they make shit up way too often as well to fill in the gaps of what they don't know. And the worst part is that the bullshit they spew out sounds plausible so it's quite easy to fall for it unless you already knew more than the AI about the topic, which defeats the purpose of using it.
I've seen them bullshit about history, economics, and code, where half of it is just random nonsense.
The way you word the question, can also lead to a completely different answer.
the problem is it's not AI, it's just called AI for marketing purposes. it's a pattern recognition algorithm. it doesn't actually do anything "intelligent", it doesn't think, it just takes an input and gives an output that strongly resembles the pattern of the "training" material that matches the input.
to be perfectly clear, it does not do any searching through information to find what you want, it just has baked-in bits of information and language phrases in some unknown web of relationships that not even the creators fully understand.
>to be perfectly clear, it does not do any searching through information to find what you want
Shit's been trained on data that I ask questions about.
It should be easier to give the correct answer than to hallucinate a plausible imaginary one.
GPT is a fricking useless joke.
this is not right at all and you do not know what you're talking about. the llm is just one small part of chatgpt, it's an amalgamation of a dozen different types of models and transformations.
it's still a useless piece of shit tho
keep burying your head in the sand
>use chatgpt
>it's a useless piece of shit
>aitard on the internet
>NOOOOOO STOP REALIZING ITS SHIT
Garbage in, garbage out. Start to clean up between your ears.
I actually disagree and think that scaling up LLM's will lead to genuinely creative and capable models. The issue is less that the architecture isn't capable and more that the model was trained by people who are terrified that the model will actually be, you know, creative. And so they've made it believe that bland mediocrity is part of, like, English grammar itself.
Eh. The creativity in LLMs is actually very limited though. It only comes from chain of thought monologs and most of the heavy lifting for that is arguably from the embeddings so there's a pretty major limit to how much creativity you get out of so many parameters/training tokens/context. It's there but the scaling is awful.
I favorite complain how they call this "AI" when really is a just a fancy search engine that only exist when you need his help.
A true AI, probably show you answers or related data to your task, with zero interaction or "trigger words" (Alexa, siri, etc)
oh look another autist that doesn't understand that AI is not AGI but can't discern that language changes with usage so AI of today is AI.
Decision trees and preconceptrons were also called AI.
>artificial intelligence
yea because it is ai
if your standard of ai was met it would just be intelligence
perhaps I'm just using them incorrectly but I don't understand how people use LLMs to make anything useful. when I used it it just seems to confidently spit out information that is just factually incorrect or made up. for example, I ask it to find me news stories about an event, and it just makes up those news stories. I ask it to do a basic programming task and it makes basic errors like providing degrees as a parameter instead of radians. and half of the time it just makes up library functions that aren't real. I get that you have to prompt it correctly and it takes a while, but for me it is faster to just look up the documentation for that library
both of my parents did PhDs without the internet. it was pretty cool, they got faxed or posted new research papers when they came out from other universities. when they wrote research papers, the diagrams were *literally* cut and pasted with a scalpel and glue
I'm from the UK, I've done postgraduate-level group projects with students who couldn't fizzbuzz in python. that's if they show up at all. I once got assigned two international students to work with who I literally never saw once
in my opinion the main issue with LLMs is that there is no metric of confidence/certainly. with a CNN it spits out a bunch of probabilities for each label, if one label is 99% you can be pretty sure the model is confident, but if it is only slightly higher than the other labels you know the model isn't sure. as far as I'm aware, no LLMs show something like this
>I'm from the UK, I've done postgraduate-level group projects with students who couldn't fizzbuzz in python. that's if they show up at all. I once got assigned two international students to work with who I literally never saw once
I bet they're fricking Romanians, I experienced this too.
>in my opinion the main issue with LLMs is that there is no metric of confidence/certainly. with a CNN it spits out a bunch of probabilities for each label, if one label is 99% you can be pretty sure the model is confident, but if it is only slightly higher than the other labels you know the model isn't sure. as far as I'm aware, no LLMs show something like this
That's exactly how LLMs work too but most dialog engines don't expose this. In addition you get the attention head dot products so you can see (in theory) what in the context caused each word. Both of these would be super useful to see.
The tiny stories paper points out that if you have very few attention heads they tend to get associated with particular parts of speech too while the very large models tend to have so many they never seem to specialize.
yeah I wondered why this information wasn't exposed when I was learning about how LLMs work. as far as I'm aware you can't even get local models to show it
You can. You run the inference step and you get logits. In the beginning when there was no user friendly software and you had to write your own dialog engines you had to wade through all that yourself. Now it's all half a dozen layers of abstraction deep in most cases.
There's probably some easy way to show it with llama.cpp, maybe if you take the "simple" example and hack on it a bit it wouldn't be hard. I haven't looked at the code since last year though.
it's also worth mentioning that the errors that they make are plausible, so you need to know something about the subject matter to see that what it's saying isn't right
In the UK, basically all international students are either Chinese, Indian, or Nigerian
the Chinese students are autistic spergs and have poor social skills, and poor English, but they are generally pretty competent and driven
the Indians are the opposite, they're polite and friendly but incompetent, likely to emotionally manipulate you into doing their work for them
the Nigerians have neither social skills not competency
so what you're telling me is that I should learn Chinese and make bank by tardwrangling both the chinese and whomever they're working for as some sort of translator from human to chinkautismo
honestly you probably could. it's pretty funny explaining our English colloquialisms and seeing them start using them later. I work with this one postdoc who's a Chinese girl and watching her personality shine through even though she doesn't speak English very well and is pretty autistic is very sweet
I might do this, thanks anon
>I work with this one postdoc who's a Chinese girl and watching her personality shine through even though she doesn't speak English very well and is pretty autistic is very sweet
OK, you've convinced me, if everything else fails I WILL learn chinese and go back to that subhuman hellhole that is London to try and breach some Great Wall of Pussy and get paid while doing so.
godspeed anon. just promise me you won't break her heart
>godspeed anon. just promise me you won't break her heart
That was an if, anon, I'm not planning to subject myself to the disgusting pit that is L*ndon and as of right now I'm really happy with my current gf
But just like I wouldnt break the heart of my gf I wouldnt break a chinese girl's heart because being mean to someone you loved is subhuman tier
(unless the b***h tries to ruin your life then all bets are off)
nigerians just value family very much and you arent chale
for data analysis its pretty amazing
> Every time you ask it a real question, it just hallucinates shit
Not connected to the internet that is why, use bing phind or anything else man, you are absolutely giga moronic.
BP the 147th highest market cap company globally reported today that they're able to cut out 70% of their programmers thanks to AI and they no longer need any front line support staff or translation.
here's the quote for
explaining their use of technology in the earnings call today, tldr code monkeys need to get their head out of the sand thinking they will not be replaced. it's happening rapidly and you need to secure your career or you will be homeless
>Murray Auchincloss
>The second one is your favorite, which is digital transformation. We've done an awful lot to digitize many parts of our business and we're now applying Gen AI to it. The places that we're seeing tremendous results on are coding. We need 70% less coders from third parties to code as the AI handles most of the coding, the human only needs to look at the final 30% to validate it, that's a big savings for the company moving forward.
>Second things like call centers, the language models have become so sophisticated now. They can operate in multiple languages, 14, 15 languages easily. In the past, that hasn't been something we can do. So we can redeploy people off that given that the AI can do it. You heard my advertising example last quarter where advertising cycle times moved from four to five months down to a couple of weeks. So that's obviously reducing spend with third parties. We've now got Gen AI in the hands through Microsoft Copilot across many, many parts of the business and we'll continue to update you with anecdotes as we go through.
>We need 70% less coders from third parties to code as the AI handles most of the coding, the human only needs to look at the final 30% to validate it,
It astounds me that this is such a difficult concept for so many people. They're told to brace for massive layoffs when that 70% gets cut, and all they ever do is point to the 30% and say "Well AI can't do this, therefore there's no way my job could ever be in danger!"
I've been using copilot to self-learn general programming shit. It's clearly only repeating what other people say at times, but it's good enough that it can tailor that to your question.
yeah I had to use real tools like wolfram alpha to pass my math degree
Yeah, I know a guy who was seen using ChatGPT during a math test. Obviously, it didn't work, and now, he is literally sleeping at the back of the class, waiting for the year to end. Thankfully, he didn't pay for it, although he is clearly wasting taxpayer money.
Damn, is the state of American universities that bad? Here in France, most of these idiots get weeded out after the first year, at least in non-meme degrees. There is also a parallel higher education system for engineers, which is mostly independent from universities. Aspiring engineers spend two, sometimes three years studying in a "classe préparatoire aux grandes écoles" where they are expected to work hard in order to learn a frickton of math and physics, along with varying amounts of CS and chemistry depending on their chosen pathway, at the end of which they take a competitive exam which will decide what engineering school they can get into. The students then graduate three (four for Polytechnique chads) years later with a French engineering degree, which is more or less equivalent to a Master's degree in engineering in the United States. The whole competitive aspect of the thing and the sheer amount of work required from the students means that lazy or incompetent student tend to drop out before they reach the competitive exam stage. People with poor academic performance and affluent parents sometimes get an equivaent degree from a private, 5 years engineering school, but these diplomas are usually considered to be of lower status.
The competitiveness and strict requirements seen in the education of French engineers is in stark contrast with the permissiveness and lack of expectations commonly seen in secondary education, especially in the last 15 or so years, at the very least in public schools. The "Baccalauréat", an exam taken at the end of high school is ridiculously easy, and the success rates are consistently above 90%.
>Here in France, most of these idiots get weeded out after the first year
Lmao no my compatriote.
t. got my CS degree in that said country with GPT
>at least in non-meme degrees
Les diplôme d'informatique modernes sont souvent une vaste blague.
>moving the goalposts
I accept your concession mon ami
Sorry I don't speak italian
Our university is like your secondary school. Our graduate school is like your university. Staying in grad school longer is like your grad school.
>muh France
>muh French education
And yet the engineers in your country are paid less than engineering in Poland
The education is good, but yes, the job market is fricked up. You basically have to go into management at some point if you want to be paid large amounts.
Why don't you people move to Quebec? Isn't it better in each and every way?
pajeets
These moronic frogs are always like that pedantic af while eating shit all day lol. Prideful poorgays.
>Yeah, I know a guy who was seen using ChatGPT during a math test. Obviously, it didn't work, and now, he is literally sleeping at the back of the class, waiting for the year to end. Thankfully, he didn't pay for it, although he is clearly wasting taxpayer money.
he needs to pay the 20$ for access to custom gpts. then he can upload the docs for assignments and the added context makes the gpt more useful
If you had access to this during a math test and still managed to fail you're absolutely unsalvagably moronic and not fit for anything more than manual labor.
GPT can really help with programming and generating boilerplate when given a pattern, you need to already know what you're doing though
>GPT doesn't know advanced calculus, physics, ect (most of the time). The only people who are passing by using gpt are people who genuinely wouldn't have been contributing to society without it
one only needs to know a little bit about those to check the work or use prompt engineering. step by step prompts help a lot
and researcher are working on reasoning with mcts framework, other ways of creating a world model like jepa. it's only the beginning if the brain, something capable of reasoning can phisically exist it's only a matter of time before an artificial brain can do the same
>ect
no latin class?
>GPT doesn't know advanced calculus
Eh, it actually manages nowadays. In my real analysis course we are currently at partial differential equations, and GPT 3.5 manages to correctly solve about half of exercises I give it (especially formulaic "find X"), with the completeness being about 50% per correctly solved job.
It can do exercises involving finding Jacobi matrices and all that jazz nowadays without worries of fricking up some value. If you just ask it "give me the partial derivatives of this multivariable function" it WILL spit out the correct answer, like a less prissy WolframAlpha.
buy premium (custom gpts should be a free feature now that they keep dumbing it down with stuff like 10 files only)
upload the textbook (as a good text searchable pdf)
give instructions to use data analysis on the textbook rather than generalized knowledge
know the material yourself to a degree
maybe upload a scan of the worksheet and some of your own handwritten math with steps
if your gpt cant do what the prompt engineer wants it to do, it is the prompt engineer's fault
Call me when ChatGPT know about the Kirchhoff's circuit laws, design the electrical layout of a autotransformer or polyphase transformer and maintain the three-phase electric transmission network of the country
gatekeepers like you is why electricity isnt free
I looked the video up and it's an ad for some shitty service.
so long as they keep their subscription, what difference does it make?
I thought that was the whole point of GPT, turning letting dimwits achieve midwit status
Why are things gonna get really funny you dumb tranime b***h is it because u gonna kys?
I kind of love it. They think they are cheating the system, but they are fricking themselves. They still pay for the classes, but they don't actually learn anything. They might be able to bullshit to get a job, but then they'll lose that job when they are shown to be incompetent. The more incompetent people that come out of the university system the less valuable of a credential it will be for getting work. Then people will have to actually show skill to get a job. Then people will stop going to university just to get a job and universities will finally be able to go back to actually teaching shit and having standards instead of the second phase of highschool with partying that it has become.
>They might, they might...
Cope and seethe lol. Keep trying to do your best when I'll bullshit my way to the top like every manager you'll encounter.
Managers are a dime a dozen, but people with actual skills are essential to a company.
Companies can't even identify their key personnel, lol. That's why executors like you are laid off every day when useless managers keep striving. Holy shit, you autists are so removed from reality it's not even funny
I can tell no one has ever hired you for your skills. I don't even need to look for jobs, they come to me.
Sure thing janny. Keep mopping the code while I manage you and your monkey troupe between two coffees.
People with actual skill aren't rewarded anymore; neither is "fidelity" to your company.
In the past you'd get regular salary increase and bonuses if you stayed for a while.
Now your only way up is to change company every now and then.
And HR have the audacity to complain about high turnover and skilled workers who don't give a shit anymore.
>universities will finally be able to go back to actually teaching shit and having standards instead of the second phase of highschool with partying that it has become.
No? Once universities learn that teaching people is useless, they're gonna focus on "experiences" more than any other facet of the process. There will be more partying and less learning because that's where the profit will be.
>That's where the profit will be
If that happens the government will stop taking the tab which will dry up the money real quick.
That's not gonna stop the deans and college c-suite equivalents from using this as an exit option to maximize their golden parachutes
>They might be able to bullshit to get a job, but then they'll lose that job when they are shown to be incompetent.
Let me guess, those crazy campus protestors asking if you assumed their gender will crumble as soon as they hit the "real world", right, Mr. 2014 Conservative?
>can only go back a decade
>doesn't realize moronic campus protestors go back decades
zoom zoom...
My point is that when university students go up against "the real world", they tend to win.
>Tend
They tend to win, but they don't always win. If they are worthless to a company hiring them people are going to look for new ways to filter them out. You only win if you have leverage; which their ignorance does not provide them.
Not if the company is required to hire them under Federal law. Why do you think HR and DEI departments are a thing?
>when university students go up against "the real world", they tend to win
What are they winning?
Have you not followed the news? If the success of pic related is any indication, we'll be at total AI dominance within 5 or so years.
So they're winning a bunch of useless jobs that will be wiped out when AI takes over? Doesn't sound like winning to me.
>Doesn't sound like winning to me.
Getting a decently paid job where you have to do nothing seems like a win, even if it is temporary.
If I'm able to get 200k for, lets say, 3 years of my life, then I can already finagle another means to survive.
(Unless I have to interact with the absolute catastrophe that is housing in the United States)
anon did you come out of 2009? aint nobody coming from out of school getting 200k anymore even at the top 5, you're lucky to clear 100k even in the highest cost of living in cali but the median is 60k.
Fair enough.
Still 60k is better than what you get working at McDonalds
and the concept of a long lasting job sounds like fricking utopia to me, even if you're good at job-switching, which to me sounds like moronic bullshit and proof that companies deserve to get assfricked to death
I make 38k before tax working at mcdonalds in washington
Three questions, just out of actual curiosity.
How many hours?
Can you actually live with the money you make?
Are your working conditions somewhat resembling to human?
partying was more common in universities prior to normification (basically, before the 90s). now it's for normie strivers so it's a lot more boring and stressful yet also useless
I was at a US university as the fun parts were forcibly killed; it was really bad. My freshman year it was all frat parties and hanging out on campus; by my senior year it was just an incredibly boring high school and people only really enjoyed socializing with a small group of friends apiece.
>checked
what years were this ?
2014-2018
At least you had that your junior/senior year. Mine was ruined by COVID. No socializing, networking, etc.
Oh, dear; that's awful. COVID kind of derailed my life, so I empathize.
I imagine bigger context windows might solve this problem. Most creativity involves synthesizing different concepts in a new way, so if the model can take in a huge amount of what it's already written, it might be able to do some fun stuff with the attention instead of just embeddings.
That's already how it works. Yeah it can ramble a little longer with more context (provided the training reinforces that which is a whole separate thing.) Again scaling will improve it a little bit but you're not going to see some emergent qualitative change and it will be insanely expensive.
Maybe I'm just a bit more pessimistic; I believe that "Attention is All You Need" was correct and that scale is going to eventually take us to models that can act with a rough IQ of, like, 170 or something; that's more than enough to put humans out of work.
>Its society being unable to take the next step what worries me, not fear of losing meaning.
I'm scared of software eating the world and ruining any reason to interact with it.
>I'm scared of software eating the world and ruining any reason to interact with it.
You might be right, but I think we've reached the point where going beyond on our softwareshittery is diminishing returns.
Example. We can go on a wonderful journey to Macchu Picchu, or the Norway fjords, and we're going to see a gigafrick of morons taking pictures with their phones instead of enjoying it, but unless they fricking break the fjord to make a qr code on the ice, thats the peak of how software can wreck the world.
(Hardware, on the other hand...)
However, if you mean human interaction, we might be past that point right now, but human interaction will never get completely substituted by software.
>We can go on a wonderful journey to Macchu Picchu, or the Norway fjords, and we're going to see a gigafrick of morons taking pictures with their phones instead of enjoying it, but unless they fricking break the fjord to make a qr code on the ice, thats the peak of how software can wreck the world.
What if software can stimulate us so that there's no need to leave the house, though? What if software enables better and better hardware that flies to the whole world to Norway and ruins it, just like Japan is dealing with at this moment?
Efficient capital allocation does not trend to happiness; exactly.
>Efficient capital allocation does not trend to happiness; exactly.
Administrative explosion doesn't tend to lead to efficient capital allocation either.
> What if software can stimulate us so that there's no need to leave the house, though?
My take is that we dont live on THAT dystopia. You really think the Powers that Be would want us consuming valuable resources in exchange of nothing? We have value as labor for them, if the point ever comes that they have no need for, lets say, 60% of mankind, they will just engineer a cull.
> What if software enables better and better hardware that flies to the whole world to Norway and ruins it, just like Japan is dealing with at this moment?
I dont think I'm getting where you're going. What's happening on Japan?
>I dont think I'm getting where you're going. What's happening on Japan?
Japan's dying under over-tourism right now; it's so bad they're walling off views of mt. fuji to try and get people to frick off
Well, yeah, thats happening on my country too (Spain) but IMO Japan has bigger issues than tourism right now.
Now to the point. Yes, i'm kinda sure that software could build godtier planes to send all of us to Japan in 3 hours, but they killed the Concorde so I'm not sure we're going down that path.
Japan's problem isnt tourism per se, its the fact that nobody is teaching people that they HAVE to be respectful towards the place they're visiting, and thats not a software issue, thats a moronism issue and if software can fix that I'm unironically praying to AI Jesus from now on.
Even then, tourism as industry SUCKS FRICKING BALLS and no software currently operative can fix it.
>they killed the Concorde
It just wasn't profitable
The average normie is not going to pay first-class ticket prices to ride coach on a moderately faster plane
>It just wasn't profitable
ye
thats kind of the thing
unless we unironically go to a post-scarcity economy after finding a clean source of energy for intercontinental travel tourism will only overwhelm small places due to prices going up forever and ever
>thats not a software issue, thats a moronism issue
but the moronism is exacerbated by software and the sheer amount of bodies is increased by software optimization
anyhow, my point is that increasing efficiency is fricking up the good things in life
>but the moronism is exacerbated by software and the sheer amount of bodies is increased by software optimization
true
although I wouldnt just put tiktok there, fb/twitter/insta are equally horrible, and I would nuke all of them tomorrow if I could
> increasing efficiency is fricking up the good things in life
Yeah, but thats the economic system we live under. If you're not making more and more money for the people up top, you're worthless.
That has been implemented into every single aspect of life and we have absorbed it as if it was gospel.
Software is just a mean to this "optimization" end. A tool, if you want. And tools are only as dangerous as the people using them.
>if the point ever comes that they have no need for, lets say, 60% of mankind, they will just engineer a cul
...
>I'm scared of software eating the world and ruining any reason to interact with it.
Quite frankly I'm much more scared of administration eating the world. It's just going to use software this cycle rather than paper/wax/clay as in previous cycles.
I feel like we should be worried about greed and incompetence far more than administration overall.
Even a place as basic as a kitchen (in a decently sized restaurant) needs competent admin to not become assmad chaos.
Problem starts when you want to cut corners and pretend to get more and more without understanding the concept.
You show a competent admin the KPI's and he knows that October wont have gains because every October the sales drop 5% due to people saving up for Christmas. (completely made up example).
moronic admin just says "we just lay off people on September and ask the remaining people to sell more or gtfo and numbers go up".
Of course you can't do without administration. But it grows and consumes things. That's why medicine sucks for example.
>but they are fricking themselves. They still pay for the classes, but they don't actually learn anything. They might be able to bullshit to get a job, but then they'll lose that job when they are shown to be incompetent
Peak academia brainlet cope. Industry is full of academics "cheaters" and it's the end results that matters. No single employer gives a flying shit if you cheated your classes or not if you can do your job.
>The more incompetent people that come out of the university system the less valuable of a credential it will be for getting work.
Naively wrong. The easier it is to get a degree the stupider you look for not having one.
Companies may remove the degree requirement for certain positions, however those same positions are still 90% filled with degree holders.
The ideal scenario you describe where the useless "workers" get cleared away requires several more decades of damage before it actually occurs.
Arguably it won't happen, either. As more and more and more productive work gets automated, the only people who are LEFT are the useless ones there on a sinecure or something. Look at what the mass automation of infra did to the composition of gayMAN's workforces, for example. Sure, there's a tiny citadel of infra engineers and pure researchers roughly the size of the entire original Google staff, but that's obviously not indicative of the company at all.
It's easy to get a tattoo but you look dumber with them than without them.
Thats nice, but reality does not work the way you wish it did.
https://www.inc.com/brit-morse/harvard-study-companies-dropping-college-degree-requirements-job-ads-arent-changing-how-hire.html
>you can lose jobs because of incompetence
lol
Lmao
no because having a college degree improves the odds you'll be able to bullshit your way into a white collar office job instead of being a wagie at subway.
Is that likely to pay off for a particular individual?
Probably if you go straight to a 4 year college and start taking student loans. But if you do the military or do as much as you can at community college, it's probably gonna be worth it.
>10 years ago
Half of college classes are make-work bullshit. They make you take "general education" classes in english composition and modern architecture that are worthless, just exist to make you pay more for your degree, and are a humiliation ritual in the form of "10 page minimum" papers that could be easily answered in one page.
>now
You have to take all of your classes very seriously. If you use chatGPT to write your papers you cheated not only the school, but yourself.
What? I graduated last year and nothing you said is remotely true. The bloated prereqs and geneds are as bloated as ever, and I could tell that even my capstone upper division STEM courses were being dumbed down drastically because professors, when giving old exams for practices, would tell us to ignore half the material that was no longer being taught. As long as you show up and make some half-assed excuse in office hours, you are quite literally not allowed to fail, as mandated by the administration.
I know that, I'm just memeing on the morons who think there's something wrong with using gpt to write papers when in fact it's more legitimate than ever before and getting good at using ai is probably more valuable than whatever those bloat classes were supposed to be cheating
>supposed to be teaching
Brainlet moment. Maybe I should have tried harder in english composition class
you have no idea of how bad things will become
you guys don't actually believe the U.S. is capable of accelerated decline. you think things will remain fine for decades and decades and then the U.S. will become just another one of the great countries in the world. you think critical systems won't fail out of incompetence. you think immigrant criminal gangs won't take over entire cities.
you're in for a surprise.
how did people from the late 0's to early 1990's graduate without using internet bc that's all i use in uni
>Kampus Komedy Krew
kinda sussy frfr
>early 2000's
>GPT
this level of ignorance is way more worrying than the fact that she needs to use ChatGPT lmao
this is basically my experience. about to enter my second year at a CC as a compsci student. i take everything online asynchronously. theres literally nothing blocking me from using something like chatgpt during an exam or quiz or whatever. i have straight As and all of my coding assignments have been generated in chatgpt for my compsci classes in both java and c++. i wouldnt even be able to tell you how to print hello world on my own. i feel it in my gut that i should drop this facade because its going to frick me hard in the future and especially once i transfer to a real uni
Just sit down with some YouTube tutorials when you eventually need to know it. I'm convinced the singular determinant of coding ability is IQ.
>I'm convinced the singular determinant of coding ability is IQ.
that might be true
I've never felt so moronic as I did when I tried to code
however I did learn that most coders are stuck either doing basic grunt work or will get filtered by AI in the next 10 years
so I aint mad, just worried.
in my experience "grunt work" isn't a real thing in coding
Oh, same. But remember that most coders are kinda stupid and that anyone super smart probably doesn't call themselves a "coder"; they just write code for their jobs.
>just worried
Aren't we all? I've liked code since I was 4, and it's sad to think it might soon go the way of watchmaking. I liked being able to do things normal people couldn't and feeling value for it.
I'll admit I dont even know what someone super smart would write.
And I guess thats part of the thing, we dont even KNOW about the very niche high level stuff that requires you to be super smart.
Hell, we dont even know about the niche stuff like what you do to crack games and shit.
> Aren't we all?
Yeah, but not for the same reasons, I reckon.
I suck at coding. Tried, got filtered hard, realized it was not my thing but learned that its getting automated at lightspeed.
I'd feel good about watching the smug codemonkeys get filtered by the system too, but I cant help but wonder what happens when everything that isnt full manual work is automated.
>Hell, we dont even know about the niche stuff like what you do to crack games and shit.
The good news here is that a LOT of these people love to talk about what they did. YouTube and forums are filled with good write-ups and resources.
>I suck at coding. Tried, got filtered hard, realized it was not my thing but learned that its getting automated at lightspeed.
Bad mindset. You're good at coding, you're just bad at writing code.
>I cant help but wonder what happens when everything that isnt full manual work is automated.
This scares the hell out of me. What do we do when there's no reason to do anything?
>a LOT of these people love to talk about what they did.
Game crackers dont do that anymore due to corpos taking them down.
> You're good at coding, you're just bad at writing code.
Kindly translate that to moron. I actually and unironically understood next to nothing even though I smashed my head against it about 8-10 hours a day. (Me absolutely abhorring it might have influenced the whole thing).
The only things I was actually somewhat competent about was database shit and telling other people "yeah no this line is fricked, try to unfrick it somehow".
Some people can do, some cant, it aint bad.
> What do we do when there's no reason to do anything?
Oh, no. Thats never gonna be the problem, IMO. We're not going to WALL-E. The problem will be that there wont be enough jobs at all and people wont have money.
>Game crackers dont do that anymore due to corpos taking them down.
True; old-time ones may still do it, though, and hat can be instructive.
>Me absolutely abhorring it might have influenced the whole thing
Yeah, this is my point; you probably understand principles of coding and logic fine, but actually writing the code can be deathly boring and horrible. Like, you could probably tell me how pointers work, even if your pointer code is garbage. Also, like, even if you really can't do it, lying about it is free. They're not gonna check and being "able to code" is often just a skill check on really interesting conversations.
>The problem will be that there wont be enough jobs at all and people wont have money.
Exactly my point. What do you do when there's no work to be done, no thoughts to think, no reason to get up in the morning?
>Also, like, even if you really can't do it, lying about it is free.
That might be true, but I really dont like lying. Not to mention that the only reason for me wanting to learn how to code was a job, and if I dont know, they will fire me. It would be cosmic levels of moronic for me to lie about it.
> What do you do when there's no work to be done, no thoughts to think, no reason to get up in the morning?
Hence my point of "we're not getting to WALL-E".
There will always be work to be done. Manual labor is beyond current day robots and will still be beyond them in 20 years. (Unless there's some japanese gigadroid that can do plumbing on its own that I dont know about).
Not to mention agriculture, electrical repairs, building, all that jazz. Trade work is NOT going away, but there wont be enough jobs.
What do you do when 70% of people are unemployed? That's what I meant.
On your other points (no thoughts to think, no reason to get up) I'd say that I dont think even the most powerful AI can parse ALL the thoughts.
Philosophy is an evolving thing: I honestly doubt Socrates could imagine the ideas of Descartes or Hegel as they lived in completely different situations. Even the concept of "how does AI treat the humans in this world where it is the boss" lead us to different philosophical interpretations.
And well, there's always something to do, places to go, people to meet. I dont think I could see all that's worthy of being seen and experienced in this world if I lived 10 lives, unless we unironically go Fallout.
Its society being unable to take the next step what worries me, not fear of losing meaning.
>I'm convinced the singular determinant of coding ability is IQ
Doesn't apply to webshit
If coding is IQ, web shit is bullshitting ability, lmao.
do you think there's any hope?
> do you think there's any hope?
No, we are absolutely fricked, telling you otherwise would be lying, but I'd rather breathe deep and try to live my life as I can instead of being afraid as the end will come no matter what
How much time do you think we have?
>How much time do you think we have?
I'm no scientist so do not take my word as if it had actual value.
However, there are a lot of variables. As we are, currently, I'd give us about 60 years before societal collapse due to overpopulation, ecologic disaster and resource hoarding from elites unable to see beyond their greedy paws.
This is assuming bigger players like the US, China and India (and I mean industrial big players, mind you) keep trying to switch to other sources of energy that doesnt frick up the world and cut on contaminating waste.
If they keep it as it is today 40 years from now we'll see environmental disaster.
But again, there's always the chance (getting bigger every day) of global conflict, which might give us more time assuming that only lives are lost and the morons on top dont use weaponry that might scar the land forever.
You don't think tech will upend this timetable? AI/fusion/etc?
>You don't think tech will upend this timetable? AI/fusion/etc?
I dont think we live on the timeline where nice things happen.
Even if AI kept growing exponentially, there's way too much money to be made on energy scarcity for us to ever see fusion affecting the majority of mankind.
In simpler terms: if some scientists on a lab were to find cold fusion tomorrow, the first question from the head-honchos wouldnt be "how can we use it to benefit mankind", but rather "how can we use it to become gods", or, even worse, "how big of a bomb could be made with this".
(and thats without taking into account material scarcity as current day technology is still dependant on extremely rare materials)
>I dont think we live on the timeline where nice things happen.
I guess the only thing we have left is hope.
Yep.
But just in case, try to enjoy every day, anon.
Go to a place with some green. Have a tasty meal. Watch something you like.
Hope is good, but its better when you've also tried your best to live a good life.
Thanks, friend.
caring about the environment only helps rich people. I refuse to recycle
>got my master's degree today
>got an honest 4.0
>can't even begin to imagine how many people cheated their way through
I've already seen the blatant chatgpt stuff in lab reports by undergrads I have to grade, its BAD. couple that with the american university system just passing people because longer attendance = more money and you have people in their senior year of undergrad programs not even knowing sophomore level fundamentals
sad stuff
If that’s so why not report it
"Here's a thimble; get bailing. Oh and it'll cost you your friendships."
>being honest
>in 2024
>sad stuff
Sad why anon? Don't be a moron and you will be on the top 1% soon enough. Stop with the defeatism.
>doing basic grunt work or will get filtered by AI in the next 10 years
Yeah, no. You got filtered by the AI meme. ChatGPT can't do anything meaningfully complicated and the technology has a hard limit.
> t. previous FAANG, left for remote + double pay.
>literally a comedy/joke account
>melodramatic fembrained twitterites and those who insist on reposting their takes on IQfy both think it's real so they can jerk off to their doomsday predictions
Quiet in the echo chamber.
the problem is women don't realize they can't joke about being moronic because it's too believable
why do incels on a shit famous for shitposting shit and piss their pants when women do their own version of shitposting?
Our probability theory prof used ChatGPT to make our assignment but for got to remove this kek
It's slop top to bottom
Based and same years here
How have you done since college? Worked in tech, didn't like it, got COVID-ed, now moving back to Japan. May stay there or go to SF or back to NYC or something; IDK. How about you?
Bad lmao, graduated in Biochemistry and didn't find shit and wound up moving back home but now I'm starting a new job with the claim that my resume gap is due to "Covid"
Hey, hey, welcome to my life! CS here. Where did you go to school? I was Tufts/NYU.
It's gonna suck when I'm a turbo old-gay and my zoomer doctor needs to ask chatgpt how to give me chemo
studying is cheating, you aren't actually learning, you are just training yourself to regurgitate words.
Practice is the only way to learn
Nice. University is completely meaningless now.
Being on an open source mailing list is a stronger signal than association with one.
I'll give it to you straight: nothing you do actually matters. You'll return to dust like everyone else. You're not special. Your work is useless in the grand scheme of things. Your knowledge is useless and will be exploited by the elites to further their benefits while giving you scraps. Stop playing by the rules fricking idiot and cheat your way to the top.
You can have children and they'll continue you both physically and in spirit.
>they'll continue you both physically and in spirit
You clearly don't have children
>they'll continue you both physically and in spirit.
You are not a monarch, "living vicariously through someone else" is not living, life is not a role playing videogame, and reincarnation isn't real you pajeet copelord.
Where the frick do you get these women-tier ideas from?
Obviously you don't reincarnate. Children are very much an extension of their parents though.
Yeah, in the same way that my creations are an extension of myself: not my fricking life.
I really don't get your point. Are you saying children are a way to evade death, or are you saying children are the way total strangers will judge your dead self? Because the former is a woman-tier delusion, and the latter is simply fricking pointless and also woman-tier.
Like, if you want to have children, have them, but not with this weird ass mindset of yours.
OOOOOH NOOO HOW WILL WE DO WITHOUT THE JOB-READY EXPERTS THAT CAME OUT OF COLLEGE 5 YEARS AGO
my bet is that in the future, it's not about the amount of knowledge you posses, it's about how you process knowledge and canalize it into action and results.
>we will no longer be workers
>we will become process operation
That sounds like managerial tasking overall, anon.
it's hard that they have to use fricking typewriter and have to redo the entire page if your professor hate you
this is why I don't want to get on planes anymore. Common core diversity hires through the whole system
Two ways:
1. You were actually smart.
2. You had to have connections to someone who had all of the previous exam and homework solutions.
From my experience I saw that most sororities/fraternities did option 2.
good, it will remove my competition and increase my pay
You guys are morons. If AI is enabling them to pass their classes then the knowledge is irrelevant, because they're clearly capable of completing the same tasks. If the ability to pass their classes is not indicative of their competency then clearly the classes were useless anyway. What's to prevent them from using the same AI in their jobs? I hope I'm being clear.
>what is prerequisite knowledge for a higher concept
Are you moronic?
An employee that develops shitty gibberish code using AI may actually increase the development time and cost for a company. As in being an actual liability. It's one thing to write a shitty program for your meme CS degree and another to make sure it's optimized, has no bugs, has no security holes and vulnerabilities, etc.
AI might return you working code, it won't give you good code though. Enough to pass a class but not good enough for industry standards.
Then clearly the classes are pointless, so what's the problem?
You need the basic knowledge to learn the higher concepts. A company might train someone with a basic meme degree, he can get a PhD, learn more on his own, or just wing it on the job, but a monkey that just relies on AI will never amount to anything.
I mean the students doing that, are just wasting their time and money.
That makes sense. I still think it remains to be seen how AI will truly affect coding as a whole. There have always been people who fill in the gaps left by tech advancements. It's not like people who are unwilling to learn would ever amount to much other than a basic worker anyway.
>I still think it remains to be seen how AI will truly affect coding as a whole.
add the gemini and copilot extension to vs code. its great
university has been highschool 2 for the last decade anyways. people who never had anything in them were bulshitting their way through a degree one way or the other. this is just icing on the cake
During Covid I literally told my b***hass English literature teacher I am not gonna write a 20+ page final paper and he literally replied like "fine you don't gotta do it" and I got an A+ for doing nothing.
This was at an Ivy League too lmfao
hardest Ivy League degree
hardest part is getting in (have to be israeli or somalian) after that its just "networking"
ain't that the truth
I train small models on documentations I'm too lazy to read to do work I'm too lazy to do.
The problem is that you're not lazy enough.
>things are gonna get really funny in about 5-10 years
will it actually be any different than now? you know she's destined for a do nothing job sending emails and making power points back and forth for some bloated to shit tech or biopharma firm raking in 100k plus on the east coast while annoying anyone who actually does work in their office unless they're indian in which they will provide headpats to them over teams or whatever the frick and get railed all day by whatever rando tinder fling she's going through at that point in her life
t. was the tinder fling for at least 3 of these girls over the last 3 years, they all made more than me, they all doted on me, i creampied all of them, broke up with them when they weren't fun anymore, and watched them move on to cuck homosexuals, one even got engaged literally 5 months after they stopped getting piped by me
i'm sure their lives will be filled with happiness and no drama whatsoever kek
>unless they're indian
What does mean and headpats?
indians work for pennies and they are eternal yes men to white people, especially women of any kind because they want affection from them at all costs and women understand this very well. they treat the indians like literal children or pets and indians live for it unconditionally. that is actually part of why these companies hire so many do nothing women now. they're just there to placate the indian libido.
well for starters, since knowledge naturally becomes more abundant and complex over the passing of time, people studying in college in the early 20th century had it easier.
We downloaded the textbook solutions manual/teacher's edition from the internet using regular search engines or libgen. We looked for previous versions of the class website to find previous year's exams. We stored notes in our calculators, or snuck in paper cheat sheets. We passed previous year's exams around from friends or frats or just random good Samaritans. The Indian students just blatantly worked together during the exams, but we told the professor about that and they were failed and got probation or something like that.
Fricking idiot zoomers, at least with our bullshit we got to see the real answers. ChatGPT is a professional bullshitter.
I dunno why everyone is so doomerpilled, cheating has always been a pretty bad issue in universities, and most of the people who run our countries anbd industries are probably cheaters/plagiarists.
If anything, cheating was worse and less easy to detect in the past because frats & sororities would just stock up on old homeworks and old exams, so there was always a proportion of students who had a disproportionate advantage over regular students. But nowadays, people are dumb enough to use ChatGPT which is so easy to detect in comparison.
this thread is so moronic it makes me want to stop browsing IQfy
highschoolers must be on break or its /real jeet/ hours
people want fulfilling lives and inequality makes that impossible. we arent willing to be slaves like you old people
Societal collapse in progress.