4 ms·
Quitting programming as a career right now because of LLMs would be like quitting carpentry as a career thanks to the invention of the table saw.
by simonw 1y ago
Quitting programming as a career right now because of LLMs would be like quitting carpentry as a career thanks to the invention of the table saw.
- chii 1y agoMore like quitting blacksmithing due to the invention of CNC.
- globular-toast 1y agoCNC produces something no blacksmith could. The same cannot be said of LLMs.
- bluefirebrand 1y agoI agree, but it doesn't matter what we think Execs are convinced LLMs produce something that no programmer ever could. Or at least the same thing, faster than any programmer ever could So they will drive us off a cliff chasing it
- globular-toast 1y agoLuckily people who know better are allowed to found companies and compete with them. I look forward to large companies with several levels of clueless execs going under, especially because I'll still be here to pick up the pieces in exchange for a tidy sum.
- guappa 1y agoThey are, but right now it's a hard sell with VCs
- globular-toast 1y agoYeah. I would recommend everyone to budget for being able to "weather a storm" for a few years regardless of what they do. If you have liabilities requiring you to maintain your current income indefinitely then you need to fix that. Personally I sleep at night knowing I could take a 33% pay cut tomorrow and not even make any lifestyle changes, let alone struggle with any liabilities. Live way below your means. Be Stoic.
- imtringued 1y agoForging and subtractive manufacturing are different techniques.
- globular-toast 1y agoThis comment seems especially good considering it's written by one of the authors of Django. I started my career making websites in raw HTML, CSS and PHP. Stuff like Django, Bootstrap, React etc. came a long and made a lot of the stuff I used to do trivial. Not once did I fear for my career, rather I got excited because now I could do so much more. I wonder what it is about LLMs that has got people scared rather than excited.
- saltcured 1y agoFor me, at a macro (societal) level, I'm kind of a mild "doomer". My worry is triggered by way our world is becoming "software-defined" at the same time AI hype is driving lots of people into insane delusions of productivity and quality. I do not look forward to a "vibe-defined" technological environment in the future. I really hope this hype cycle comes crashing down before it damages our shared world, having displaced properly engineered systems that had disciplined human oversight. The way AI products (not just LLM) are being hyped disturbs me. For example, I do not like the trend to alpha/beta test safety critical car control systems on public roads. I do not like the way the public is embracing these things and blithely trusting them, and rapdily lowering their vigilance. To me, this seems like a latent disaster, where global socio-technological systems may collapse once this feedback loop proceeds far enough. At a micro (personal career) level, I just recoil at the whole LLM paradigm. I am seriously looking at my finances and contemplating a hard "no" to retire. I know myself pretty well. I have spent a career becoming a very senior "individual contributor" in an academic R&D environment. I am not a manager, and would not want a managerial job. I do not enjoy nor am I good at delegating work to lower-skilled juniors. I have had excellent collaborations with true peers over the years. I reject the mindset that an LLM can be such a peer. It lacks the hysteresis, learning capability, and actual agency to form a trustworthy working relationship. I refuse to spend my energy on pretending it is more than it is. Part of my issue is that, despite doing R&D prototyping, I am a very top-down thinker. I do a lot of ruminating through all the levels of a system before I build it. I know I depend on this to get aspects of quality out of anything I build. I accept that I am not a good code reviewer, e.g. to take random code and find flaws or rework it. I need to trust my partner not to be gas lighting me with fiction nor sneaky flaws. When in doubt, I will "nuke it from orbit" and start over in a green field manner with top-down designs where I can gauge the risk. I would much rather write from scratch than take someone's janky code and rework it. I would never take such code and then stake my own reputation on it being functional. Doing so with a stateless generator is nightmare fuel. I would rather walk away become a janitor or auditor who has to monitor the output of the LLM crack monkey...
- camgunz 1y agoMaybe you dashed this off and I'm digging into a comparison that wasn't meant for deep scrutiny. But ripping wood is an extremely small part of carpentry. The AI hype is literally "have an LLM eat a bug report on GitHub, wait a little while, review the bug fix PR, also this is the least capable this will ever be." That's kind of the whole thing. Sure there's system design, requirements gathering, blah blah blah, but the whole product is legitimately "SWE too expensive? Get Claude!"
- simonw 1y agoThat's my point: I'm pushing back against the AI hype here, and arguing that the part of our job that can be done by an LLM is far from being everything we do.
- camgunz 1y agoFly.io just shipped Phoenix.new which is legitimately "do the whole thing super fast (as long as you could already mostly do the whole thing at regular speed)". I really don't think this is hype.
- simonw 1y agoPhoenix.new is great, but you still need to know a lot about web development to use it effectively. https://simonwillison.net/2025/Jun/23/phoenix-new/ https://simonwillison.net/2025/Jun/23/phoenix-new/
- camgunz 1y agoYou're right and I agree: in order to successfully use LLMs to build software products of any significant complexity, you need an experienced dev at the helm. My guess is this is the case for another couple years max, but who knows really. Reading your (as always, great) post, it seems like Phoenix.new did like 95% of the work? You had to coach it around FTS, but I also have to look that up every time so :shrug:. It's pretty bananas; I still maintain this is nothing like "traditional software engineering but one of your tools got way more efficient", i.e. you got a table saw. This is "I typed 'build a table' into a box, it couldn't figure out dovetailing, I taught it dovetailing, and I had a table." Regardless of whether or not you think that's good or bad, I think it's clear it's a huge change.