localhoster

Yavos - yet another vibe coded shit

Yavos - Yet Another Vibe-cOded Service :)

Idk if this article was vibe written or the author just "got adjusted" but it's clearly is, and it's unreadable. Man this becomes anmoying

Tbh I always wondered how are we still matching routes using regex and not something like a radix tree? That would eliminate these kinds of issues no?

Couldn't read a single char on my phone because of how abnoxious this website it. So many ads popping up like mushrooms after the rain. At this state I wish AI take over news site because this is unbearable.

This is the stupidest thing I have heard today. I love it!

It's so hard to trust a vibe coded software with something with a thing as valuable as data.

I see no reason why would anyone even bother checking this out, while there a trusted, battle scared, and non vibe-coded alternatives such as postgres, duckdb,boltdb,SQLite.

Sorry mate.

I often wonder to myself if we are the teachers protesting against the calculator.

The tool is undoubtedly useful, and it has been said that calculators also removed the need to do manual calculations.

I think the difference between LLMs and calculators is inherently small. Both are algorithms. And algorithms have already were making decisions for us before LLMs, so LLMs are a natural next step of the path we were already in. Deterministic or not, I think that's irrelevant.

So, in a sense outsourcing your learning is just the natural next step. If you look about it this way, you were already doing that. We don't need to learn how to play on tools to listen to music. We don't need to learn how to do complex calculations to do them. It's just not right to look at LLMs as different from other decision making, study decapitating, tech we already have.

Ok so in all honesty I have it a try. I think you are burning "input tokens" rather than output ones, which fill your usage quicker.

Suggestion to improve? :praying_hands:

Dumbest point ever. There is no value for this issue. I don't agree with the way they did the rewrite, but they did the rewrite, and this post contributes nothing, beside making the author seem childish. If it had any real contribution I would have waved it off, but it really doesn't. This tribalism and "I'm better than you"-ism and the same reason everybody hated the stack overflow community, and the rust community as well.

My dumb ass read it like "postgres for de win" and I got baffled by the fact that clockhouse, of all companies, said it.

Another one bites the dust

Companies will tell you that it allows them to examine many directions faster, and in a sense they are right.

Not doing something is much faster than doing it fast. Sit on your ideas, think them trough, dismiss early, or choose to move to implementations.

another behavior I noticed is that even you plan with an agent than a lot of business logic leaks to the code.

some states, for an example, are meant to be assumed from the data shape, rather than the actual state fields, but damn they like adding a state field.

I think the issue that majority of agentic developers don't understand their design fully. Those holes, perviously, were have to be closed by a person eventually, and inherently you gain understanding in the process.

This is not the case eny more. I never thought LLMs write bad solutions, but when you let it think for you you loose something important, understanding. And when something brekas, some people consider this being a proof that the llm is the problem. And in a sense it is. But you are also to blame. And that's exactly the issue with this all industry - in order to move faster you don't need to "type" faster. A person bashing his hands on the keyboard randomly at top velocity can get pretty high wpm. In order to move master you need to _think_ faster, be _snappier_ and _sharper_ and most people aren't.

I like how one of my colleagues phrased it in a company meeting. He asked one of the c levels "for llm s to 10x me, I need to let them make all decitions and dictate intentions. By percentage, how much of the code you want me to _know_?"

Almost accurate It's not "the route" But "a route"

Tiring reading this gpt (prob Claude) "enhanced" (generated) articles that are meant to ultrapump the AI industry. I don't know if it's the current lingering economy trying to hang on AI as it's savior, or people making written proofs of "ai hype" for their internet persona, but it's tiring.

I want to read about the issues with this new magical technology, and how are we trying to solve them, rather than MORE HYPE

Let's make it very clear - this is the original creator of redis, or one of them.

He is not "your avg dev" and it took him 4 months with llm.

This is not a seal of approval for you to go and command all your developers to move to Claude code/codex/any other ai coding tool fully.

I'm looking at you - any avg CEO of a startup.

Listen, I don't like AI as much as the next person, but I find ed zitron very hard to listen to or read through. Too much yapping, to much of repeating himself.

Forgot m-dashs and the arch.md file

Huh I thought they meant like a new paradigm.

Ok this was a nice read. Very expressive and I agreed with the majority of the points. I'm not going to lie, I expected "and this is why we built <insert generic gpt wrapper>", but it never came. Which leads me to my question, and forgive me for my impatience, but _what_ are you selling? Is it a platform? Ide? Books? So many words and I ended with more questions than I started. On any who, good read indeed.

I rent in my country (first world) for 1000ish. I get a 60 sqmt home with garden And two rooms. It saddens me that the usa startup culture gave us this.

So I think theres some more nuance than that. A lot of the times, the abstraction is solid enough for you to work with that code area, ie tracking down some bug or extending a functionality. But sometimes you find yourself at a crossroad - which is either hacking around the existing implementation, or rethink it. With LLMs, how do you even rethink it? Does it even matter to rethink it? And on any who, those decisions are hidden away from you.

Hard to read this article as it was written by Claude as a report after the migration that Claude did for you. If an llm helped you migrate and save this much money, kudos. But if you decide to write about it at least proof read it and remove redundant parts and llm storytelling.

> "A week of back and forth, 2.3 billion tokens, $2,283 in API costs, and about ~20 hours of me unsticking it from dead ends. It popped calc."

Corrent me if I'm wrong, I'm not a security researcher, but 20 hours, a week of work, 2283$ spent and over 2 trillion tokens, is not very 10x-ing as we were promised. Especially if you take into account that the guy is at least half capable for this take.

I dunno

Tbh I dont really agree with your statements. Especially with working with data, intention is key. By using an llm, by definition, you are loosing intention. And Thai puts you in a position where you have to 1) think of exactly what you look for. 2) able to understand what the llm generated.

You might say it "still less work" and that's true, perhaps, only for the first few times. After a while you _learn_ how to do it, and understand how to _think_ with the language of your data. With LLMs, you never get this benefit, and also loose your ability to judge the LLM's output properly.

But again, that might be enough on your case, or, you simply don't _know_.

This article surface every once in a while, and I love it. What the author suggests is very clever. I have implemented an extended version of that in Go as an experiment. Instead of using a trie, I used a radix tree. Functions the same, but it's much more compressed (and faster).

Was it ever? Even before llm, writing software, or at least web clients, was as easy as it can get.

Weird flex but ok