We mourn our craft

> perpetual subscription services for the same software we used to “own”.

In another thread, people were looking for things to build. If there's a subscription service that you think shouldn't be a subscription (because they're not actually doing anything new for that subscription), disrupt the fuck out of it. Rent seekers about to lose their shirts. I pay for eg Spotify because there's new music that has to happen, but Dropbox?

If you're not adding new whatever (features/content) in order to justify a subscription, then you're only worth the electricity and hardware costs or else I'm gonna build and host my own.

Than your younger self that got all excited about the Internet.

The quality of local models is still abysmal compared to commercial SOTA models. You're not going to run something like Gemini or Claude locally. I have some "serious" hardware with 128G of VRAM and the results are still laughable. If I moved up to 512G, it still wouldn't be enough. You need serious hardware to get both quality and speed. If I can get "quality" at a couple tokens a second, it's not worth bothering.

They are getting better, but that doesn't mean they're good.

As a very old school programmer who taught myself assembler in 1982 on an 8-bit 4K micro, I don't see much to mourn here.

* People still craft wood furniture from felled trees entirely with hand tools. Some even make money doing it by calling it 'artisanal'. Nothing is stopping anyone from coding in any historical mode they like. Toggle switches, punch cards, paper tape, burning EPROMs, VT100, whatever.

* OP seems to be lamenting he may not be paid as much to expend hours doing "sleepless wrangling of some odd bug that eventually relents to the debugger at 2 AM." I've been there. Sometimes I'd feel mild satisfaction on solving a rat-hole problem but more often, it was significant relief. I never much liked that part of coding and began to see it as a failure mode. I found I got bigger bucks - and had more fun - the better I got at avoiding rat-hole problems in the first place.

* My entire journey creating software from ~1983 to ~2020 was about making a thing that solved someone's problem better, cheaper or faster - and, on a good day, we managed all three at once. At various times I ended up doing just about every aspect of it from low-level coding to CEO and back again, sometimes in the same day. Every role in the journey had major challenges. Some were interesting, a few were enjoyable, but most were just "what had to get done" to drag the product I'd dreamt up kicking and screaming into existence.

* From my first teenage hobby project to my first cassette-tape in-a-baggie game to a $200M revenue SaaS for F100, every improvement in coding from getting a floppy disk drive to an assembler with macros to an 80 column display to version control, new languages, libraries, IDEs and LLMs just helped "making the thing exist" be easier, faster and less painful.

* Eventually, to create even harder, bigger and better things I had to add others coding alongside me. Stepping into the player-coach role amplified my ability to bring new things into existence. It wasn't much at first because I had no idea how to manage programmers or projects but I started figuring it out and slowly got better. On a good day, using an LLM to help me "make the thing exist" feels a lot like when I first started being a player-coach. The frustration when it's 'two steps forward, one back' feels like deja vu. Much like current LLMs, my first part-time coding helpers weren't as good as I was and I didn't yet know how to help them do their best work. But it was still a net gain because there were more of them than me.

* The benefits of having more coders helping me really started paying off once I started recruiting coders who were much better programmers than I ever was. Getting there took a little ego adjustment on my part but what a difference! They had more experience, applied different patterns, knew to avoid problems I'd never seen and started coming up with some really good ideas. As LLMs get better and I get better at helping them help me - I hope that's were we're headed. It doesn't feel directionally different than the turbo-boost from my first floppy drive, macro-assembler, IDE or profiler but the impact is already greater with upside potential that's much higher still - and that's exciting.

  > The answer they always give is that compilers are deterministic and therefore trustworthy in ways that LLMs are not.

I don't see this as a frequent answer tbh, but I do frequently see claims that this is the critique.

I wrote much more here[0] and honestly I'm on the side of Dijkstra, and it doesn't matter if the LLM is deterministic or probabilistic

  It may be illuminating to try to imagine what would have happened if, right from the start our native tongue would have been the only vehicle for the input into and the output from our information processing equipment. My considered guess is that history would, in a sense, have repeated itself, and that computer science would consist mainly of the indeed black art how to bootstrap from there to a sufficiently well-defined formal system. We would need all the intellect in the world to get the interface narrow enough to be usable, and, in view of the history of mankind, it may not be overly pessimistic to guess that to do the job well enough would require again a few thousand years.
  - Dijkstra: On the foolishness of "natural language programming"

His argument has nothing to do with the deterministic systems[1] and all to do with the precision of the language. His argument comes down to "we invented symbolic languages for a good reason".

[0] https://news.ycombinator.com/item?id=46928421

[1] If we want to be more pedantic we can actually codify his argument more simply by using some mathematical language, but even this will take some interpretation: natural language naturally imposes a one to many relationship when processing information.