Yeah we all hear the main arguments... AI is bad because of slop content, stealing from creators, brain rot & brain damage, privacy concerns and most importantly... how billionaires are just using it for their own selfish reasons
โBut I'm asking about YOU ๐ซต personally. The individual. What do you really think about AI? Do you care or are you indifferent? Has it actually affected your day to day life?
I went through a phase where I was doing a lot of vibe coding. I thought that I was smart enough to be working around the known problems and limitations of it. I was not. Pretty much everything "I" had made turned into an unmaintainable nightmare once it passed a certain level of complexity, and in practice when you're trying your best to actually make vibe coding work on a project larger than one file you spend so much time reviewing every little thing, rewording your specs etc that you might as well just write the fukken code yourself.
So I deleted all of those projects, and now I use AI for brainstorming (I talk into speech to text, usually while doing something else, then give the text to an LLM and say "organize these ramblings into a design document for me") and simple specific task automation ("comment my code" and "update my readme"). I figure the one thing LLMs are demonstrably good at is summarization, so pretty much everything I ask them to do is along those lines.
That said, if I could press a button that would make all of the ai companies crash and burn but it meant that I would have to back to doing those things manually, I would press it in a heartbeat. The externalities that the tech world is forcing into all of us for this thing that makes summarizing documents more convenient are way too much.
As for anything else, generative images or video or whatever... I just have zero interest in making or consuming them. Sure it hits my feed like it does for everyone else but I'd say at this point my reaction is about 1/10 "haha" and 9/10 "ugh" when I see them. There was already so much human made art out there that you could be looking at new things all day every day and you wouldn't be able to keep up.
edit: also, regarding the coding: if your goal isn't "learn to code" but rather "make a quick python script that needs to work exactly one time", then vibe coding can be a valid use case. but if you are trying to learn then prompting cannot teach you in the same way that you will never learn to make high quality digital art by prompting for image generator.
I find it's perfectly possible to write large maintainable projects using these tools. I have a Rust project I built with LLMs that's over 150k loc now, and it's structure is a lot better than anything I would've ended up on my own. One of the things I do is ask the LLM to come up with a phased plan for introducing new features. I also ask the model to make mermaidjs diagrams I can inspect and come up with file layout up front. Then I get it to make a branch for each phase and implement a focused feature. Then I can review it and I have good context for what it's supposed to be doing, and it's scoped so that the code is manageable enough to fit in my head. Doing that alone gets you a long way. Another thing I do is ask it to make refactors by looking through the code base and finding repeating patterns in code that can be consolidated, or large files that need to be split up. If you do this regularly, you end up with much cleaner code, and the agent is much better at doing that at scale than you could by hand.
Testing is another really important aspect. I always ask the model to do TDD, and then add end to end integration tests for features. For web apps, using playwright storybooks is really effective. You can define exactly what the user workflow is and then have the model test it through a headless browser end to end. This creates a contract where you know what that functionality is actually working end to end.
As long as you don't just let the model crap out tons of code unsupervised, and box it in sufficiently with a contract, then the code is no worse than what a human would produce. And I'd argue that it's often better.
Openspec is along these lines and works well ime
Yeah it's a nice tool, you really just need a bit of rails to keep the model on track I find. They're good at doing well defined tasks, and if you have a plan where they just check steps off as they go, things tend to work well. I also found beads is a really handy tool for task tracking cause then you can just file stuff in there and instead of writing tasks in markdown you have a real history along with the status of the tasks.
good to know, I'll check out beads. I was forced to start using Claude at work and had mixed impressions but once we started using openspec to keep it on rails as you say, it got hard to deny its capabilities.
I find I kind of look at the whole LLM + agentic harness setup as a genetic algorithm. Your tests and specs are the fitness function for the program you're evolving, and the LLM is the mutator. At each step it generates some output, it gets tested against the fitness function, the LLM gets feedback and iterates on it. Eventually something working falls out in the end. The better you can define the selection criteria the more you box the agent in the better results you get.