when AI makes STEM/Coding brehs obsolete......

[the bossman]

it's not. chatgpt 5 is worse than 4. gemini created their first capable model.

yeah like what?

Breh stop waving your old man stick in the air. Take a step back and let's look at the big picture. This is not even debatable what's happening today. Let's put the hype aside and look at what's actually happening.

Virtually every image that any model ever generated always had those fukked up looking human fingers and was horrible at rendering text. Until nano banana pro dropped three months ago. Now that marketing manager might not need to hire freelance designers as much as he used to in the past. Analysts don't have to depend on design teams as much

Then even a year or two ago AI could help you code but you still needed to understand proper code syntax, install the right toolchains, manage the dependencies, debug whenever it got stuck, it struggled across multiple files, etc. Them days are gone now. Claude style code agents with opus 4.5 got completely non-technical ppl building full blown usable prototypes now over a weekend. Anybody in whatever company can build their own dashboard or internal tools for whatever they need

Are most of those apps secure, scalable, resilient? Not really. You still need your senior developers. But now you don't have to depend per wait on the engineering team just to get a prototype going. The effect on junior developer roles has been documented for past few months now.

Does the current state of where things are at today justify mass layoffs? No. But if this pace keeps up then there's no way you can say that there wouldn't be fundamental changes to a lot of white collar job roles

 

[Still Benefited]

A.I is mostly hype,i do believe the jig will be up soon.

 

[O.T.I.S.]

Trust me. You dont want llm replacing devs. The world will collapse.


BTW we dont have ai, we have advanced chatbits

Thank you

And nowhere near AGI last I heard

This shyt about money right now, but could be a beautiful and useful thing with guidelines and guardrails. Especially economically

 

[cyndaquil]

If you work or worked in tech for a siginificant amount of time and understand software development (like me or clearly @Tair) or @Ty Daniels it's SO easy to see who's skimming hype/marketing headlines and who actually understands what's happening.

Understanding a) how software development works b) how (in)efficient AI is today to generate revenue for companies just by itself c) how important humans are to lead both developers and ensure code quality is high and best practices are followed, and that entire development teams including UI/UX designers, developers, animators you name it are working on the right things (this is my world) and ensure that the CEO/leadership vision is executed to the max level of efficiency to push up key performance indicators (like daily active users, lifetime value, average revenue per user etc.). is VITAL. That's not replaced with a "prompt" or an "agent that never sleeps".

YES you can code an by vibe coding.
YES there's some powerful stuff out there (for video editing, for managing agents, taking notes, creating assets, saving time, making quick prototypes)
NO no AI is beating in the field +10 years worth of human insight into what to build when, how to run a team properly, how to go from vision to revenue etc.
NO your UX or UI is not optimized just because an app was "delivered" by AI and you said "ok make the UI better"
NO you can't just sit back. You need to understand what users love, hate, need etc. to be able to even dream of optimizing KPIs AND beating the competition.

I can keep going brehs but you get the picture.

This is facts. But damn is everybody a developer on here?

 

[cyndaquil]

Current SWE will move into robotics/AI training/engineering. They won't have to work on finding errors and smaller tasks that take up a majority of their time, and they'll get to the "real work" faster, which will ultimately get them doing more work, not less.

Other STEM majors will work on interdisciplinary fields in robotics (from microfluidics/nanotech to wireless power transmission, and a bunch more areas)

If anything, AI will open up more areas for people to get into, not less.

But I find it hilarious that people think STEM cats just memorized some equations for plugging and chugging.

Cuz honestly there are a lot folks who don't work as or with SWEs professionally who took a boot camp during the pandemic or a course on free code camp or got stuck in tutorial hell on youtube and think software engineering is just simple HTTP methods doing CRUD on the backend (if they ever even got to that point) and a few simple html components and css on the front-end.

Type of people who think Netflix streaming a movie on their tv is as simple as just clicking play on Stranger Things and sending a GET request to a single server

 

[Sleepy Floyd]

AI will always be useless

Can't even get a fast food order right

 

[Mission249]

Cuz honestly there are a lot folks who don't work as or with SWEs professionally who took a boot camp during the pandemic or a course on free code camp or got stuck in tutorial hell on youtube and think software engineering is just simple HTTP methods doing CRUD on the backend (if they ever even got to that point) and a few simple html components and css on the front-end.

Type of people who think Netflix streaming a movie on their tv is as simple as just clicking play on Stranger Things and sending a GET request to a single server

For the last decade, 99% of the industry got flooded with CRUD and web app devs who can't do shyt else. The dark secret about the industry is that many were getting way overpaid for repeating mostly the same boilerplate in different biz domains and not giving a shyt about the craft - just the bag.

Notice none of the following devs are constanty being AI hypesters: OS/kernel, hardware, firmware, 3D graphics & video games, etc. They're far more measured.

If Claude can regurgitate some React app, and the dev claims he's 10x, should I expect to see 10x more levels in the next Grand Theft Auto?

 

[thefloorislava]

Technology is being invented and implemented at an exponential rate. Much of that technology uses software.

Software is a landscape that rapidly changes, and A.I has exhausted all current scraping sources (stack overflow, reddit)

A.I can give wildly different answers to the same question depending on the wording of the prompt due to its sycophantic nature and struggles with consistency and solving novel problems.

So we have an industry/trade which deals mostly in novel problem solving, where solutions & frameworks that were the rage become obsolete just as fast.

My prediction: Standards for software jobs will increase, and teams will get smaller, but engineers with specialized skills (Embedded, Computer vision) and domain knowledge (Mathematics, Statistics, Biology) will flourish in select roles.

Basic Web development (CRUD apps, Simple Websites for Businesses) are either handled by A.I or small teams at most.

Developers working on complex systems i.e organizational intranet, sensors, large web services, robotics. Continue to flourish

 

[null]

A.I can give wildly different answers to the same question depending on the wording of the prompt due to its sycophantic nature and struggles with consistency and solving novel problems.

LLMs (not AI) can give different answers to exactly the same prompt (or attention) because it randomly selects tokens in the case of dead- or close-heat.

"When an Large Language Model (LLM) determines that the next token is part of a set of possibilities, it uses sampling techniques to select one based on calculated probabilities. Instead of just picking the most likely option, the model can choose from a ranked set of candidates, which introduces randomness and variability into the output"

Decoding Strategies: How LLMs Choose The Next Word

Large Language Models are trained to guess the next word. But when generating text, the combination of their probability estimates with algorithms known as decoding strategies is what determines how they actually choose words. Learn how decoding strategies work in this article.

www.assemblyai.com

that is what gives individuality of output when say telling a story or writing a script.

 

[bnew]

AI will always be useless

Can't even get a fast food order right

humans don't get fast food orders right everyday. we don't even know why various A.I systems fail to get fast food orders wrong. we don't know what modls they're using, their system prompts, agents or tools they make available to it.

 

[null]

humans don't get fast food orders right everyday. we don't even know why various A.I systems fail to get fast food orders wrong. we don't know what modls they're using, their system prompts, agents or tools they make available to it.

LLMs will refuse to accept corrections even when directed to do so. human staff don't do that. not for very long anyway.

Incorrect orders is nothing to do with models. All LLMs "hallucinate".

ChatGPT does so 30%+ of the time.

Which business uses staff who make errors 30% of the time?

"ChatGPT's hallucination rate varies significantly by model, with newer reasoning models like o3 (33%–51%) and o4-mini (48%–79%) showing high error rates on specific benchmarks, while GPT-4.5 has shown improved, lower rates (roughly 37.1% on SimpleQA). Hallucination rates often range between 3% and 91%, heavily dependent on the task complexity"

https://www.nytimes.com/2025/05/05/technology/ai-hallucinations-chatgpt-google.html

 

[Oldschooler]

yup
all these ceos and companies care about is how much money they can make and staff salary is usually the number 1 cost for most organizations

the lie of ai was selling the it will be like autopilot you still need the pilot but the autopilot is there to help ease things tools requiring constant human oversight, but navigating the unexpected, monitoring system performance still needing teh human touch

But what they want is no pilot at all just straight AI doing everything


im a blue collar worker delivering packages for ups screw y'all tech brehs, tech brehs lose their jobs, suddenly people ordering stuff and shipping is down, suddenly that ups gig they don't need all those drivers

Thought you worked at a hospital? As a driver you'll be replaced too eventually. The first to be replaced will be workers that rely heavily on the computer or software. Writers, translators, programmers, designers, etc. It's amazing how quick the technology has developed.

Humans are going to go back to basics until robotics is advanced enough to take over which will be centuries from now. The best careers to be in are careers that you are doing as much physical labour as possible. Nursing, surgeon, plumber, any sort of trade in the building industry really. The issue here is that like you mentioned previously the economy dies when spending stops. Spending stops or decreases when people lose their jobs or get undercut in salary. It doesn't matter that you still need a programmer to look over AI code now. What matters is that previously you had 10 software developers making $150k+ each down to 3. The rest maybe in India or China which don't help the local economy.

 

[MikelArteta]

Thought you worked at a hospital? As a driver you'll be replaced too eventually. The first to be replaced will be workers that rely heavily on the computer or software. Writers, translators, programmers, designers, etc. It's amazing how quick the technology has developed.

Humans are going to go back to basics until robotics is advanced enough to take over which will be centuries from now. The best careers to be in are careers that you are doing as much physical labour as possible. Nursing, surgeon, plumber, any sort of trade in the building industry really. The issue here is that like you mentioned previously the economy dies when spending stops. Spending stops or decreases when people lose their jobs or get undercut in salary. It doesn't matter that you still need a programmer to look over AI code now. What matters is that previously you had 10 software developers making $150k+ each down to 3. The rest maybe in India or China which don't help the local economy.

I do work at a hospital was just an example