I think you’re mistaking improvements in tooling as improvements in the LLMs. LLMs are plateauing. The idea of exponential growth is an illusion. We took 20+ year old technology, geared it toward text (the LLM), and trained it on the entire Internet. Then, it’s popularity grew exponentially.
This is the hype narrative that Altman, Dario, Jensen, etc. push. They are trying to convince everyone that what we have is the Model T Ford of AI. Just imagine where we’ll be in 6 months!
Yup. The problem was never the hardware. It’s the stack. We’re all carrying supercomputers in our pockets now.
Feels like so many (dare I say most) programmers don’t even understand the work that the OS does to make the C programming environment as nice as it currently is, let alone a 200 line TODO app that uses 6GB of RAM.
That’s fucking nuts because any device that can connect to a cloud is powerful enough to run an operating system (they will probably not give user access to the non-volatile memory). Just not the bloated AI spyware box that they want.
I work in embedded, and you would be surprised how much work a CPU under 100mhz can do when there isn’t an operating system/browser in the way. We need to get back to basics in the software industry. We’ve been going down the wrong road for a looooong time, and the AI bubble is only accelerating us in that wrong direction.
I feel like this has more to do with what field you work in and what language/paradigm you use. Especially if you’re working within some bullshit walled garden, you may not have a choice. I’m a terminal jockey myself, but I mostly program in C, so my code is procedural and to the point. Maybe I might want some fancy smart refactoring feature if I worked in a language where half the code is boilerplate or glue.
If I have the choice though, I don’t see any advantage to an IDE. It’s like the combination of many tools rolled into a single, bloated UI with about 60% of their original functionality. And I guess it lets you build “projects” and choose which files will be built. That part never made sense to me. I don’t need a program for that! Just delete it dog. It’s in the repo!
IDE:
I’m my opinion, these programs are just better as separate programs.
(Rant) One thing that grinds my gears… Some IDEs will leave you with the dumbest possible directory structure imaginable. Like actively hostile toward us terminal jockeys. Remember, we are repeatedly typing these things out like cavemen. For example, c/c++ developers who put their headers in a separate, but identical directory structure. Oh and let’s do full taxonomy and go 10 directories deep. And what the hell, capitalize random letters and throw in some with spaces into the directory names for good measure. These things don’t have to matter to IDE people, but it is something to be mindful of.
AI hype in a nutshell
gcc main.c
They are also dev friendly too,
Not saying you’re wrong because I don’t use it, but from the outside, they appear actively hostile toward developers.
100%. In my opinion, the whole “build your program around your model of the world” mantra has caused more harm than good. Lots of “best practices” seem to be accepted without any quantitative measurement to prove it’s actually better. I want to think it’s just the growing pains of a young field.
She already exists! I swear!
Yeah, I had a silly hack for that. I don’t remember what it was. It’s been 3-4 years since I wrote bash for a living. While not perfect, I still need to know if a pipeline command failed. Continuing a script after an invisible error, in many cases, could have been catastrophic.
Woah, that ((i++)) triggered a memory I forgot about. I spent hours trying to figure out what fucked up my $? one day.
When I finally figured it out: “You’ve got to be kidding me.”
When i fixed with ((++i)): “SERIOUSLY! WTAF Bash!”
I was never a fan of set -e. I prefer to do my own error handling. But, I never understood why pipefail wasn’t the default. A failure is a failure. I would like to know about it!
I’ll give you my vim when you pry it from my cold, dead hands.
Easy win for companies that didn’t buy into the hype. I’m the only dedicated software dev at my company, so there was no middle manager to foolishly think a chat bot could do my job. We are a small company that can compete with big players, and those big players appear to be floundering. Now, we are expanding.