Conversation
its over for codecels
8
1
4
@georgia I love these charts please make everyone retarded quit programming so the rest of us can get jobs
2
2
11
@georgia
Luckily fukken NOBODY KNOWS WHST THE FUQ THEY NEED :DDdDDDDDd


Let alone write spec :ddDDddDD


((The whole industry is fucked none the less ))
0
0
1
@georgia this was obviously coming as I’ve been saying, and this is only going to improve over the next few years
1
0
0
@RustyCrab @georgia Won't help, the retarded programmers are going into AI.
1
0
1
@georgia is it actually good at coding?
1
0
0
@PunishedD @RustyCrab @georgia Nah, math filters out retards. They'll all going into cybersecurity and other segments of IT now.
1
0
3

To be fair the majority of coders are pajeets who use AI anyway so it's a battery powered battery charger.

0
0
0
@georgia Competitive coding is a completely different skill set from actually building software

One thing I really like about the AI hype though is that it's make the divisions between people who actually love CS and software engineering and the "people" who are just in it for a job very pronounced
0
1
8
@anemone @georgia @arcana don't fall for the propaganda + you're rich you have no reason to fear for your job lmao
1
0
2
@lucy @anemone @georgia It's not propaganda, technology is going to keep increasing, and o3 is passing a great deal of benchmarks and coding tests
3
0
0
@anemone @georgia Use your wealth to start up your own stuff
1
0
1
@arcana @anemone @georgia @lucy my partner works in AI and tests multiple models peddled by major companies and I disagree tbh
2
0
2
@teratology @anemone @georgia @lucy Yep, and I also have people close to me at the absolute top of these fields that say the opposite. You can also use these things yourself and see how they are definitely at the level of replacing juniors, and it's going to continue improving.
2
0
0
@teratology @anemone @arcana @georgia arcana wants to make money with AI so she needs people to believe in it
3
1
1
@meso @anemone @arcana @georgia @teratology yeah that's why i don't engage in an argument over it
1
0
1
@anemone @georgia @teratology @lucy It's pretty clear where the actual benefit from these things is. I actually read recently and I agree that ultimately, they amplify what gets put in. For those who find them unuseful, that's an indictment of that person more than anything.

Lots of lots of businesses are already finding them to a be a major improvement over traditional employees. I go to a bunch of networking events and regardless of what you think, more and more businesses and business owners are seeing a great ROI with them
2
0
0
@lucy @anemone @georgia @teratology Lol, if you want your head in the sand so be it
0
0
0
@lucy @anemone @arcana @georgia I mean there's "working with AI" and then there's actually doing machine learning shit that requires quite a significant mathematical understanding. Just from seeing the shit generated at my bfs job even he thinks it's retarded hype for the most part.

I have a younger relative going super hard into the AI hype and shit. He also didn't seem to comprehend how cryptocurrency exchanges could be compromised lol
1
0
2
@teratology @anemone @georgia @lucy The people I know are top tier machine learning people and extremely mathematically knowledgeable.
2
0
1
@arcana @anemone @georgia @teratology if it's the people i think you're talking about they're all merchants tho
1
0
0
@lucy @anemone @georgia @teratology I don't know who you think I'm talking about
0
0
0
@arcana @anemone @georgia @lucy

So you are in favor of profit over people essentially? I'm not even talking about whether or not the claims made by OpenAI or other major companies is true or not, I'm talking about the ethical implications of AI.

There are actual interesting and useful applications for ML in several domains, I just don't think that replacing people is one of them.
2
0
1
@teratology @anemone @georgia @lucy I think the future of the world is a much, much smaller number of people who are elite human capital, and that a lot of workers are ultimately replaceable and that the march of progress doesn't care about them

This is coming no matter how much people kick and scream about it. Adapt or die has always been the way in this universe.
2
0
2
@arcana @anemone @georgia @lucy and do you consider yourself among the elites? Do you think that you yourself aren't replaceable under this paradigm? I mean you said you're not really tech savvy after all.
1
0
1
@teratology @anemone @georgia @lucy Being super tech savvy isn't really what will be of most value in the coming future.

The role I play is complex.
3
0
0

@teratology @arcana @anemone @georgia @lucy I wonder if these conversations took place in the past when new tech allowed to replace workers

1
0
1

@arcana@fedi.layer02.net @georgia@netzsphaere.xyz @teratology@the.asbestos.cafe @lucy@netzsphaere.xyz they’re an improvement in the same way outsourcing customer service was an improvement. They’re worse but currently a lot cheaper.

2
0
1

@arcana @anemone @teratology @georgia @lucy if james has 5 apples and i have 4 tomatoes, how many horses are equivalent to the diameter of the sun?

0
0
0
@teratology @anemone @georgia @lucy Also if you think I'm in favor of this kind of thing, you're mistaken. I simply understand what is coming.
0
0
1
@anemone @georgia @teratology @lucy I'm not saying this stuff because I support it, I'm saying it because people need to understand the shift that is happening and act accordingly for their own perservation
1
0
0

@arcana@fedi.layer02.net @georgia@netzsphaere.xyz @teratology@the.asbestos.cafe @lucy@netzsphaere.xyz we’ll see once the vc money dries up. This stuff currently isn’t good enough to replace an engineer in any real capacity and it’s a lot more expensive than the current prices suggest

1
0
2
@Hyperhidrosis @georgia people who don't work as software engineers can't understand what "good" or "bad" means in this context. These LLMs are extremely good at tiny stock problems or "challenges" like these. They completely fall apart and their error rate approaches 60% PER LINE when you get them into a real world program.

But by all means please tell everyone that human engineers are obsolete so we can get some of our jobs back.
1
4
8
@anemone @georgia @teratology @lucy have a backup plan for life is all I’m saying. Things are still on schedule for the 90% population drop by 2040
2
0
0
@arcana @anemone @georgia @teratology i won't be alive by then so good luck
0
0
0
@arcana @anemone @georgia @lucy why do you think that by 2040?

This shit sounds dystopian ASF and will hit the poor and working class the hardest 😩 you seem way to nonchalant about it ngl
1
0
0
@mischievoustomato @anemone @arcana @georgia @lucy yeah it's happened a few times but not to this scale, like the industrial revolution and the first luddites
0
0
0
@arcana @anemone @georgia @lucy it really sounds like you work in business and rub elbows with technocrats
1
0
0
@anemone @georgia @lucy @arcana it's the way healthcare is now, it's cheaper because they're virtualizing everything but at what cost? Shitty care? Shitty customer service.

Of course profit must be prioritized above all else.
0
0
0
@teratology @anemone @georgia @lucy the vast majority of those technocrats will die too btw
0
0
0
@teratology @anemone @georgia @lucy same reason I knew of the virus in the early 2020s back in 2014.

I’m nonchalant about it because it’s set to happen. I hope those I consider friends make it through but ultimately billions won’t make it, that tends to be how extinction events go. It’s merely a cosmic cycle in a sense
3
0
0
Non techies see this as a win but if AI truly replaces a competent programmer then all their jobs will have been replaced first.

What happens when AI is able to produce competing products to Microsoft, Apple, Google etc lmao
0
0
1
@anemone @georgia @teratology @lucy I know, and nobody listened to them. I didn’t know about the pandemic in 2020 through magic or something.

It’s a big part of why I’m nonchalant too. It doesn’t matter what you know or what’s coming, people just don’t care and stick their heads in the sand.
0
0
1
@anemone @georgia @teratology @lucy most of the global population is poor and in places that will be near unlivable in a couple of decades, so yes, they will
2
0
1
@arcana @anemone @georgia @lucy

I'm central american, you don't need to remind me how expendable we are to the majority of the world and society (the global south). This rhetoric shouldn't be normalized.

It reminds me of discussions about overpopulation and the way that the west tries to put the burden of responsibility on the masses instead of corporations. It's environmental racism.
2
0
0
@anemone @georgia @lucy @arcana this was unironically a thing going on in 2013/2014 because I remember my dad being hit up and pressured to invest in BTC bc it was going to be the new thing that would totes liberate our country and people from the evil clutches of the U.S Dollar. It's turned out real well amirite 😆
0
0
1
@teratology @anemone @georgia @lucy it’s not “rhetoric being normalized”, it’s the fate of the next couple of decades. All of this goes far beyond corporations and capitalism and what not. That won’t exist past the next 20 years either.
0
0
1
@georgia Finally I can get my code without ever having to interact with coders
0
0
0
@arcana @anemone @georgia @lucy LLMs are incapable of thought and reasoning, I feel like people often forget that all they do is autocomplete text.
There are inherent limits to how 'good' outputs, that aren't close-ish to some training data, can be, and it shows. LLMs can't do maths, Tasking an LLM to generate vertices for a 3d model that you describe in text would just go wrong, or even simpler, LLMs can't really do ascii art.

And even when they should be close to something in the training set, it can go terribly wrong if it's nothing that has been reiterated times and times again.
I tried using copilot when i wrote a small helper library to encode x86 instructions, and it was the biggest nightmare, the code it generated looked reasonable, but it was riddled with small bugs that took me more time to debug than if i had just wrote it myself.
If someone's job was receiving instructions such as in files [list of files], "center all divs with the id 'foo'" and fulfilling that, they'd be in danger, but that's not what software engineering looks like.

Real AGI could eventually replace engineers and reshape things in a way you describe, but AGI is still pretty far away, LLMs are certainly not intelligent.
1
1
4
@k4t3 @anemone @georgia @lucy you clearly don’t know what AGI actually means if this is your view. Frankly you’re mistaken and still operating on 2022 levels I’d say.
1
0
1

@georgia why is this on the Kanye subreddit??????

0
0
0

@RustyCrab @Hyperhidrosis @georgia this means they are going to create a fuckton of shitcode that generations will be employed to clean up

2
0
3
@jeff @RustyCrab @georgia imagine a company that uses ai to build up their products and one day it stops functioning and they can't figure out why because of spaghetti code
2
0
1

@Hyperhidrosis @RustyCrab @georgia oh man you got no idea how much money this will make programmers in the future

1
0
1
@jeff @RustyCrab @georgia you'd essentially have to build the whole thing back up from scratch
1
0
0
@PurpCat @jeff @Hyperhidrosis @georgia I actually tried it on a new completely proprietary language and it kept randomly injecting chunks of JavaScript
1
0
1

@Hyperhidrosis @RustyCrab @georgia ssssshhh that’s the secret dont tell them that c:

1
0
2
@jeff @RustyCrab @georgia even if you did tell them. majority of people think software engineering is just coding.
0
0
1
@jeff @RustyCrab @Hyperhidrosis @georgia Pascal, Ada, and some others in that vein are also notorious for the same shit because the jeets who made AI can't scrape usenet.
0
0
1

This is already the case if you just replace "ai" with "cheap outsourcing"

0
0
1
@anemone @georgia @lucy sure, but I’ve seen first hand it doing the work of juniors in minutes. Teams reduced from 8 people to 2 people. It’s an amplifier. It will improve too.

I’m not acting like you should quit your job right now but be aware of the trends and make sure you have a backup plan
0
0
0
@hakui @anemone @georgia @lucy @teratology it’s a long story that I don’t share publicly
1
0
0
@arcana @anemone @georgia @lucy So what's your view on what AGI means / if we have reached it yet? It's not a very narrow and fixed definition that boils down to an easy yes/no.

I, for my part, am sticking to "matches or surpasses human cognitive capabilities across a wide range of cognitive tasks."
In case of LLMs, I find they are, without argue, good at predicting and transforming text, but it doesn't take me long until i hit a "wall of thought" where it all falls apart.
1
0
1
@k4t3 @anemone @georgia @lucy we haven’t reached AGI yet.

AGI refers to a general intelligence and the ability to perform out of domain. Agency too is a factor. The reason I think your definition is incorrect is that AGI is effectively the creation of a new alien species but one that can process information magnitudes faster than humans. It will effectively be game over for humanity in the same way that Cthulhu would be.

A lot of tech bros see AGI as some “wow, it’s just an employee as a service” but it’s not. Also before you go into some “well it’s not technically actually thinking…” or “it doesn’t reason like humans do”. That doesn’t matter. It’s irrelevant to risk.

Regarding LLMs, I personally see them as a system where you get out what you put in. They’re amplifiers. It’s like Krita, using the software doesn’t make you a great artist. Different people will receive different outcomes.

It’s the exact same with LLMs and I think the optimal user here is a manager. I get great results out of it by managing it in the way I used to manage junior employees. I know many others who are very skilled managers who use it to amazing effect and it’s increased their productivity very measurably.

The people who seem to have the worst outcome from LLMs tend to be people who, no offence intended, tend to have poorer social skills and so I imagine their usage of it isn’t optimal. Again, like with managing people, if you speak to your employees in a way that reduces ambiguity, sets out clear goals and constraints, and speaks in a natural and encouraging way, you get much better results than if you speak to it in an awkward manner.
0
0
1