budgie
A little off-topic, but I feel like people here have a good head on their shoulders.
I used to be quite good at making software. Was lucky enough to do engineering work at a “fairly high level.”
For most of my career, I was driven by mastery. Doing hard things, being perceived as smart, etc.
It’s not that things are no longer hard. People just don’t know enough anymore to recognize when something’s hard or when someone is skilled.
This attitude created a plague, where recognition no longer exists due to mass ignorance. It’s ruined any desire I have to do the thing I was once passionate about.
I’ve moved away from making software for software’s sake. I’m focusing more on business modeling and supporting decision-making now.
Even so, making software through AI, just as the means to accomplish other things, with real commercial value, has become insufferable to me.
The situation is made so much worse from the constant pressure to cut corners, close your eyes, and prompt. I swear I spend more time cussing out AI for doing things wrong than if I’d written the thing myself.
For the record, I still think quality matters. The hard part is convincing anyone else to give a ****, because everyone else stopped caring long ago.
I know this is all a symptom of something much worse. And for the first time, I’m genuinely scared for our future.
If one can already feel this purposeless, one can only imagine what things will be like in 5-10 years.
Purposeless or not, I’m tired of being gaslit about the capabilities of AI. I’ve spent A LOT of time building custom agents, AI pipelines, prompting frameworks, and even fine-tuning my own models across domains.
Fine-tuning tiny models is the ONLY thing that’s come close to impressing me.
So imagine that with each new release of a multi-trillion parameter “frontier” model, I feel more emotionally abused. I really can’t tell the difference between one release and the next, and I’m very skeptical that other people can.
It’s also pretty obvious to me that token prediction will not be the innovation leading to AGI. The written word is an extremely lossy compression of actual thought. Again, I feel gaslit when people say otherwise with such confidence.
But hey, large models are the only thing that justifies the current spend on components, compute, and datacenter builds. I’m not sure ANYONE can track the circular and incestuous funding relationships between these companies anymore.
All this to say, that large models aren’t the way.
I’m not even sure how to feel at this point.
I wasn’t planning on writing this much or this incoherently. But my thoughts are out there now. If you’re reading these words, thanks for giving me a moment of your time.
Where Next?
Trending in Discussions
Other Trending Topics
Latest on Elixir Forum
Sponsor Spotlight
Supporting innovation across the BEAM ecosystem.
Categories:
Sub Categories:
Forums
Popular Tags
- #ecto
- #liveview
- #troubleshooting
- #learning-elixir
- #library
- #deployment
- #erlang
- #testing
- #genserver
- #mix
- #absinthe
- #remote-other
- #otp
- #plug
- #how-to-question
- #macros
- #postgres
- #elixirconf
- #channels
- #exunit
- #discussion
- #code-sync
- #podcasts
- #javascript
- #onsite
- #dialyzer
- #docker
- #authentication
- #umbrella
- #full-time-contract
- #podcasts-by-brainlid
- #ecto-query
- #ai
- #elixirconf-us
- #blog-post
- #elixir-ls
- #phoenix_html
- #iex
- #graphql
- #genstage
- #websockets
- #supervisor
- #advent-of-code
- #distillery
- #processes
- #api
- #forms
- #elixirconf-eu
- #metaprogramming
- #hex
Our Sponsors
Build Elixir applications with speed and confidence.
Supporting innovation across the BEAM ecosystem.
We build reliable cloud platforms for business-critical systems.
Error tracking, logs, uptime, and Just Enough APM. Built for developers.
Elixir, Erlang & BEAM conferences from Code Sync - join us in 2026!
Producing high quality Elixir screencasts since 2017.
**Building the future of software with AI, trust, and human connection.**
The deployment platform built for Elixir: PaaS ease, VPS control.
The team behind Membrane, Popcorn, LiveDebugger. Available for hire.
Courses that'll move you from confusion to "Aha, now I get it!"
Showing Posts 1 to 10- Show Best Posts
- Show All (oldest first)
- Show All (newest first)
martosaur
You’re definitely not alone in these thoughts and I feel you. I’ve never in my life seen so many smart people admitting they have no idea of what lies ahead or what the path forward is.
I don’t have much of an advice though. I personally try to spend more time outside, indulge in recreational manual programming and writing blog posts. Cheers!
AstonJ
As Alex has said above, you’re definitely not alone, and this is one of the reasons I hope language and framework teams will continue to, or start focusing on, programmer happiness - it could well be something that ensures their survival in an age of AI-controlled-by-a-handful-of-companies.
Also, it might cheer you up a (tiny) bit, but I added the rant tag to your thread, making yours the first to make use of our rant icon
(It’s an experimental feature, only to be used for threads like this - not to rant about anything that is tied or linked to a person/team/project etc!!)
volcov
I find myself thinking about this several times a week, I’ve lost count of how often it’s come up in therapy, haha!
There are moments when I wonder if I’m just being the kind of person who “doesn’t like change”, but often, that doesn’t feel like the whole story.
I’ve been experimenting with AI outside of work to see if I can have a bit more fun and gain a different perspective on personal projects, since I’m the one calling the shots there, rather than following company guidelines. I think that makes the feeling a bit better, but it’s still there.
This week, I noticed something else along the lines of “AI has taken away some of the joy of building software”, something that isn’t about the relationship between me and the AI, but between me and my team. I don’t know if this happens to you guys too, but I used to really enjoy those team discussions where we’d try to design something or solve a problem together. Increasingly, though, I feel like the team talks less to one another, when we do talk, it feels more like we’re acting as “bridges” for our AI agents, comparing the results I got with “my Claude” versus what you got with yours.
I don’t know, weird times, hahaha!
martosaur
You’re spot on about team relationships. Coding was definitely a social activity for me and this aspect was disrupted by LLMs. I’m working on a longer form blog post exploring exactly this aspect, hopefully it will get more attentions in the larger AI discussion.
arcanemachine
Today in particular, this post resonates with me.
I’ve been using AI in a legacy codebase, and the complexity has slowly just bogged down more and more to the point where the model (Opus 5, Fable 5.1) is just giving garbage answers, and flip-flopping at every turn.
I’m finally to the point where I need to step through things, carefully, and untangle the slop web before things degrade any further.
The “taste” problem is still very real (i.e. these incredible pattern matching machines do not have it), and I strongly believe that there is a time and place to use these tools, but they are not the be-all-end-all.
Ultimately, I do not think the craft of software development will reach its end with these tools. They are a lever, and they can move you fast in a direction, but it is not always the right one. There is a time and place to use them (and they are incredible when they get it right), but man, I am just exhausted from this stuff.
mudasobwa
Absolutely, and this is obvious to anyone who had a quick glance at the tech under the hood.
But your whole point sounds to me as “the drill cannot make square holes, that’s why it’s of no use whatsoever.” LLMs are great at performing rough work according to a clear technical assignment, why don’t use it as such?
I elaborated on this here.
mudasobwa
In my experience the rule “garbage-in-garbage-out” works not only when one states tasks for models, but also for human beings.
Flip-flopping is 95% on the harness, not on the model itself.
Awlexus
I agree with your point overall, LLMs are good at the boring stuff, not quite as good at the hard stuff such as architecture, although they can be decent advisers.
I can’t help it, but I have to point out there are drill bits for making square holes
mudasobwa
Yeah, I worked with mortise bits myself and that’s exactly why I put it that “punny” way
OndrejValenta
I wouldn’t say I hate programming, but I find it increasingly more futile.. The common theme now is “why would I use something someone else created and put his energy and life time (the most precious commodity) into, when I can ask Claude to do the same thing, or even worse straight out steal the bits I need and throw the rest away”…
A) this is absolutely demoralizing from time and money perspective
B) it’s absolute non sense to think that a solution, be it a library or SaaS that other spent months and years developing, can be replaced by Claud over a weekend
C) they don’t realize that they now have transferred the responsibility for the knowledge base on themselves, but they think that the AI will solve that for them
D) the same goes for tests, fine tuning, responsivity, performance… man, there is so much they have to do, but they delegate it to AI and are done with it, like the AI is something you can count on 1000%
And don’t even let me started on rates.. In Czech republic, the rates have now fallen 50% for senior programmers, they are lower than in 2015, and we should be ready to go even lower..
I’ve seen that happening already, in China there are millions of programmers, artists, animators without work, all replaced by AI. The product is ****, but it’s cheap, and cheap is what greedy managers and CxOs want.