Daniel Lombraña
@teleyinex.eth
Shortly, we discover that agents were trained to use as many tokens as possible to charge us more.
1 reply
0 recast
2 reactions
Jacob
@jrf
i'm already maxing out the max plan
1 reply
0 recast
0 reaction
Daniel Lombraña
@teleyinex.eth
I would start to be suspicious 😉
1 reply
0 recast
1 reaction
Jacob
@jrf
i'm, already... how do you say? in too deep
1 reply
0 recast
0 reaction
Daniel Lombraña
@teleyinex.eth
While the architecture will be perfect there's a remaining question: is all that needed in the current project? Is it making this overcomplicated so you cannot work on your own later on and depend more on it? I would recommend reading the book careless people to look into how these companies design their products.
1 reply
0 recast
1 reaction
Jacob
@jrf
funny, i built the /atlas agent without any understanding of how to do it i had to learn every line of code and every time i added to it i made it more and more complex as of this morning, i have a system that knows itself and organized its own architecture to be more composable and scalable... and it improves itself insanely fast so what's the downside?
1 reply
0 recast
0 reaction
Daniel Lombraña
@teleyinex.eth
Per se there's nothing bad. Today I have found this blog post which resonates with what I said in this thread: Vibe Coding Casino – blog/opinion/vibes_casino — evrim.zone https://share.google/3wkwuQUYYm6zPn27t Be mindful.
1 reply
0 recast
1 reaction
Jacob
@jrf
i totally hear you and see how it can also be like this thanks for sharing be mindful!
0 reply
0 recast
1 reaction