Based on your comment you’re probably not a very good principal engineer ;)

Hence, you are back in the group of those who should benefit from LLMs. Following your own logic :)

Ps: please don’t take it seriously

No worries. True, you need to learn new skills to work properly with Claude. However, 30 yrs of coding experience come in handy to quickly detect it is going in the wrong direction. Especially on an architectural level you need to guide it.

Embrace

I urge you to actually try these tools. You will very quickly realize you have nothing to worry about.

In the hands of a knowledgeable engineer these tools can save a lot of drudge work because you have the experience to spot when they’re going off the rails.

Now imagine someone who doesn’t have the experience, and is not able to correct where necessary. Do you really think that’s going to end well?

I love it. I can't stand this sentiment and this type of technologist pompous ass. You are why software mostly sucks. You have no imagination. Hopefully the models make your limited, extraordinarily overvalued skill set the last 20 years completely democratized. We will see who is the idiot going forward.

I review PRs daily and people are pushing changes that have basic problems, not to talk about more serious flaws. The amount of code an engineer can produce is higher, but it's also less thought through.

There will be more code with lower quality. If you want to be valued for your expertise, you need to find niches where quality has to stay high. In a lot of the SaaS-world, most products do not require perfection, so more slop is acceptable.

Or you can accept the slop, grind out however more years you need to retire, and in the meanwhile find some new passion.

I thought this was parody until the last sentence.

CC is not nearly that good. It may never be. It's an amplifier not a replacer.

On the plus side you're retiring soon... imagine if your were a graduate today

I think it’s important for you to understand that there were always way more people who loved programming than were able to work professionally as high-level coders. Sure, if you spent most of your working life writing code, you’d be very proficient. But for many, many others, they haven’t been able to spend the time developing those muscles. Modern LLMs really are a joyful experience for people who enjoy software creation but haven’t had the 10,000 hours.

I think that the biggest difference is between people who mostly enjoy the act of programming (carefully craft beautiful code; you read and enjoyed "Programming Pearls" and love SICP), vs the people who enjoy having the code done, well structured and working, and mostly see the act of writing it as an annoying distraction.

I've been programming for 40 years, and I've been on both sides. I love how easy it is to be in the flow when writing something that stretches my abilities in Common Lisp, and I thoroughly enjoy the act of programming then. But coding a frontend in React, or yet another set of Python endpoints, is just necessary toil to a desired endpoint.

I would argue that people like you are now in the perfect position to help drive what software needs writing, because you understand the landscape. You won't be the one typing, but you can still be the one architecting it at a much higher level. I've found enjoyment and solace in this.

No offense but you sound more like a “principle coder”, not a principle engineer. At least in many domains and orgs, Most principal engineers are already spending most their time not coding. But -engineering- still take sip much or most of their time.

I felt what you describe feeling. But it lasted like a week in December. Otherwise there’s still tons of stuff to build and my teams need me to design the systems and review their designs. And their prompt machine is not replacing my good sense. There’s plenty of engineering to do, even if the coding writes itself.

> Any idiot can now prompt their way to the same software

If you really think it's the reality, then your expert knowledge is not that good to begin with.

I'm not sure why you feel devalued or let down, LLM code is a joke and will be a thing of the past after everyone has had their production environment trashed for the nth time by "AI."

Completely the opposite experience here! I am a tech lead with decades of experience with various programming languages.

When it comes to producing code with an llm, most noobs get stuck producing spaghetti and rolling over. It is so bad that I have to go prompt-fix their randomly generated architecture, de-duplicate, vectorize and simplify.

If they lack domain knowledge on top of being a noob it is a complete disaster. I saw llm code pick a bad default (0) for a denominator and then "fix" that by replacing with epsilon.

It isn't the end, it is a new beginning. And I'm excited.

From my point of view, having the llm as co-pilot is like having the junior engineer that the team would never justify the budget to hire. I get quite a bit more done when I can assign the tool a task to work on, work on something else in the meantime, and come back in 5 or 10 minutes to check on its progress and make adjustments.

There are many aspects of software engineering that are fun, but the pure mechanical part gets sold quickly; there are only but so many times you can type "emplace" and feel fulfilled. I'm finding that co-pilot is extremely good at that part.

Really? I love LLMs because I can't stand the process of taking the model in my brain and putting it in a file. Flow State is so hard for me to hit these days.

So now I spec it out, feed it to an LLM, and monitor it while having a cup of tea. If it goes off the rails (it usually does) I redirect it. Way better than banging it out by hand.

I think you've got this backwards!

I've been working with computers since an Apple ][+ landed in our living room in the early 80s.

My perspective on what AI can do for me and for everyone has shifted dramatically in the last few weeks. The most recent models are amazing and are equipping me to take on tasks that I just didn't have the time or energy for. But I have the knowledge and experience to direct them.

I haven't been this enthused about the possibilities in a long time.

This is a huge adjustment, no doubt. But I think if I can learn to direct these tools better, I am going to get a lot done. Way more than I ever thought possible. And this is still early days!

Just incredible stuff.

> My experience is that people who weren't very good at writing software are the ones now "most excited" to "create" with a LLM.

I consider myself to have been a 'pretty good' programmer in my heyday. Think 'assembly for speed improvements' good.

Then came the time of 'a new framework for everything, relearn a new paradigm every other week. No need to understand the x % 2 == 0 if we can just npm an .iseven()' era ... which completely destroyed my motivation to even start a new project.

LLMs cut the boilerplate away for me. I've been back building software again. And that's good.

Indeed, and I noticed companies now are focusing on hiring coops, paying them peanuts and just use AI, and have maybe one senior and one wrangler (engineering/project manager), that’s basically what I have noticed what neo-teams are.

It is weird because I am the opposite. The symbols were never the objective for me but instead how they all fit together.

Now I am like a perfect weapon because I have the wisdom to know what I want to build and I don't have to translate it to an army of senior engineers. I just have Github Copilot implement it directly.

> Any idiot can now prompt their way to the same software

I have been thinking about the "same software"

Because I remember seeing Sonnet 4.5 and I had made comments that time as well that I just wanted AI to stop developing more as the more it develops, the more harm to the economy/engineers it would do than benefit in totality.

It was good enough to make scripts, I could make random scripts/one-off projects, something which I couldn't do previously but I used to still copy-paste it and run commands and I gave it the language to choose and everything. At that time, All I wanted was the models getting smaller/open source.

Now, I would say that even an Idiot making software with AI is gonna reach AI fatigue at one point or another and it just feels so detached with agents.

I do think that we would've been better off in society if we could've stopped the models at sonnet 4.5. We do now have models which are small and competitive to sonnet (Qwen,GLM,[kimi is a little large])

In my experience, the truly best in class have gone from being 10x engineers to being 100x engineers, assuming they embrace AI. It's incredible to watch.

I wouldn't say I'm a 10x-er, but I'm comfortable enough with my abilities nowadays to say I am definitely "above average", and I feel beyond empowered. When I joined college 15 years ago, I felt like I was always 10 steps ahead of everyone else, and in recent years that feeling had sort of faded. Well, I've got that feeling back! So much of the world around me feels frozen in place, whereas I am enjoying programming perhaps as much as when I learned it as a little kid. I didn't know I MISSED this feeling, but I truly did!

Everything in my daily life (be it coding or creating user stories — who has time to use a mouse when you can MCP to JIRA/notion/whatever?) is happening at an amazing speed and with provable higher levels of quality (more tests, better end-user and client satisfaction, more projects/leads closed, faster development times, less bug reports, etc.). I barely write lines of code, and I barely type (often just dictate to MacWhisper).

I completely understand different people like different things. Had you asked me 5 years ago I probably would have told you I would be miserable if I stopped "writing" code, but apparently what I love is the problem solving, not the code churning. I'm not trying to claim my feelings are right, and other people are "wrong" for "feeling upset". What is "right" or "wrong" in matters of feelings? Perhaps little more than projection or a need for validation. There is no "right" or "wrong" about this!

If I now look at average-to-low-tier-engineers, I think they are a mixed bag with AI on their hands. Sometimes they go faster and actually produce code as good as or better than before. Often, though, they lack the experience, "taste" or "a priori knowledge" to properly guide LLMs, so they churn lots of poorly designed code. I'd say they are not a net-positive. But Opus 4.6 is definitely turning the tide here, making it less likely that average engineers do as much damage as before (e.g. with a Sonnet-level model)

On top of this divide within the "programming realm", there's another clear thing happening: software has finally entered the DIY era.

Previously, anyone could already code, but...not really. It would be very difficult for random people to hack something quickly. I know we've had the terms "Script kiddies" for a long time, but realistically you couldn't just wire your own solution to things like you can with several physical objects. In the physical world, you grab your hammer and your tools and you build your DIY solutions — as a hobby or out of necessity. For software...this hadn't really been the case....until now! Yes, we've had no-code solutions, but they don't compare.

I know 65 year olds who have never even written a line of code that are now living the life by creating small apps to improve their daily lives or just for the fun of it. It's inspiring to see, and it excites me tremendously for the future. Computers have always meant endless possibilities, but now so many more people can create with computers! To me it's a golden age for experimentation and innovation!

I could say the same about music, and art creation. So many people I know and love have been creating art. They can finally express themselves in a way they couldn't before. They can produce music and pictures that bring tears to my eyes. They aren't slop (though there is an abundance of slop out there — it's a problem), they are beautiful.

There is something to be said about the ethical implications of these systems, and how artists (and programmers, to a point?) are getting ripped off, but that's an entirely different topic. It's an important topic, but it does not negate that this is a brand new world of brand new artists, brand new possibilities, and brand new challenges. Change is never easy — often not even fair.

I know that your post has lots of comments, but I'd like to weigh in kindly too.

> I've spent decades building up and accumulating expert knowledge and now that has been massively devalued.

Listen to the comments that say that experience is more valuable than ever.

> Any idiot can now prompt their way to the same software.

No they cannot. You and an LLM can build something together far more powerful and sophisticated than you ever could have dreamt, and you can do it because of your decades of experience. A newbie cannot recognize the patterns of a project gone bad without that experience.

> I feel depressed and very unmotivated and expect to retire soon.

Welcome to the industry. :) It happens. Why not take a break? Work on a side project, something you love to do.

> My experience is that people who weren't very good at writing software are the ones now "most excited" to "create" with a LLM.

Once upon a time painters and illustrators were not "artists", but archivists and documenters. They were hired to archive what something looked like, and they were largely evaluated on that metric alone. When photography took that role, painters and illustrators had to re-evaluate their social role, and they became artists and interpreters. Impressionism, surrealism, conceptualism, post-modernism are examples of art movements that, in my interpretation, were still attempting to grapple with that shift decades, even a century later.

Today, we SWE are grappling with a very similar shift. People using LLMs to create software are not poor coders any more (or less) than photographers were poor painters. Painters and illustrators became very valuable after the invention of photography, arguably more valuable socially than before.

What I keep hearing is that the people who weren't very good at writing software are the ones reluctant to embrace LLMs because they are too emotionally attached to "coding" as a discipline rather than design and architecture, which are where the interesting and actually difficult work is done.

[dead]

[flagged]

On the bright side, working in tech between 2006 and 2026 means you should be extremely wealthy and able to retire comfortably.