Five prevalent themes in the discussion
1. The enduring value of deep, low‑level system understanding
Many participants stress that knowing how machines and OSes work is their core passion and remains essential for sound decision‑making, even when LLMs handle routine coding.
“My strongest passion in computing is simply learning how the machine and OS work. The code I would write would usually be experiments, not projects or tools.” – techgnosis
“Know what to ask is about as important as the answer. You can't do that very well without domain knowledge.” – sampullman
“People who know what they are talking about worry about all of the details from the big picture down to the small scales.” – atrettel
2. Grief over the loss of craft and fear of deskilling
A strong undercurrent of sadness frames LLMs as threatening the fulfillment derived from deep engineering work, likening the shift to an industrial revolution that devalues skill.
“Grief is exactly what many engineers (not just programmers) are going through right now… coming to terms with the death of something we loved.” – CrLf
“It’s pretty depressing all around… Llms have made people into over confident morons.” – righthand
“This reminds me a lot (mostly for worse) of the industrial revolution… displacement of code… by cheaper, shittier, more homogenous code.” – isityettime
3. LLMs as tools that still require human judgment, direction, and the ability to ask the right questions
Several commenters argue that LLMs accelerate work but need skilled humans to guide them, verify output, and focus on higher‑level problem formulation.
“ChatGPT is a fantastic encyclopedia for knowledge retrieval… we've built an amazing library registry and need to up our librarian skills.” – tomrod
“The point of the knowledge is that it helps you ask the right questions, not so much the ability to answer them.” – fancyfredbot
“Maybe people will be hired for the question they can actually ask and not for the raw knowledge they have.” – oxmo456
4. Debate over the quality and performance of LLM‑generated code
Opinions split on whether LLMs produce buggy, inefficient “slop” or can, with proper prompting, achieve meaningful optimizations and even surpass human performance in narrow tasks.
“LLM code is abysmal if you look at the details… Performance engineer jobs are not going anywhere.” – anonymous908213
“LLMs have more patience to investigate performance issues and fix them than humans.” – azakai
“I pointed frontier models at some PyTorch code… it optimized it 10x fold.” – sowhat1
5. Emergence of new roles and a shift in the software‑development workflow
Participants see a transition from manual coding to activities like architecture, prompt engineering, reviewing LLM output, or focusing on outcomes rather than low‑level implementation.
“I view it as a different layer of abstraction… I can continue to craft my Rust… and those skills help in the day to day… yet I'm also able to compete in the market.” – tomrod
“How are you keeping your ‘human’ addition to the loop valuable, is it through the time spent on the software craftsman hobby?” – spacephysics
“There will start to be a demand for ‘hand‑made’ software that someone took time to make as delightful as possible to use.” – sean2d