kaycey2022

8 ms·

61.
by kaycey2022 9mo ago
It's like they don't even care anymore
62.
by kaycey2022 9mo ago
Of course Anthropic has the most obnoxious legal team of all the ai companies. The project got traction under the older name. A name change does hurt the project.
63.
by kaycey2022 9mo ago
Copilot in vs code is definitely trash. That aside the workflow is simple. If you are familiar with the code base then make sure to refer the files where a newb has to look if you were assigning the task to them. Tell it to ask questions. U
64.
by kaycey2022 9mo ago
I’m not as behind as that. But i cant figure out this loop thing. We have engineers here saying they are reviewing 100k lines of code a day, slinging 10 agents simultaneously. I just cannot figure out how that is humanly possible. Agentic c
65.
by kaycey2022 9mo ago
What does Beck think?
66.
by kaycey2022 9mo ago
Maybe the whole attempt is to bully the current government into giving the Venezuelan/Cuban faction in the US government exactly what they want. They want an entry into the society, and to open it up, maybe among other things. With thi
67.
by kaycey2022 9mo ago
What does it matter 49 or 51? More than 40% want this. You would condemn cultures around the world happily if significant percentage are like this. Why are you so special then?
68.
by kaycey2022 10mo ago
Aliens will reveal themselves
69.
by kaycey2022 10mo ago
Unless there is an exponential growth in model context AI is not replacing anything. So far every new model launches with some increase in advertised intelligence, but only with increasingly dubious ways to measure the same. All the current
70.
by kaycey2022 10mo ago
Why can't GenAI produce good code for embedded systems? There is nothing fundamentally different between the 2. Seems like some form of nimbyism - "AI can't do what I can because I am special" sort of thing.
71.
by kaycey2022 10mo ago
I don't see why copying is unnatural or even bad. Maybe within a single economy or a group of economies which share a common understanding and laws, chosing to discourage copying to incentivise other citizen innovators makes sense. But
72.
by kaycey2022 10mo ago
I would say that handling these edge cases for minor changes that a human can easily understand, but is very hard to program rules for, is exactly where AI is the perfect fit.
73.
by kaycey2022 10mo ago
Are these patterns also the same for mobile apps?
74.
by kaycey2022 11mo ago
It matters because, it irritates me to no end that I have to review AI generated content that a human did not verify before. I don't like being made to work in the guise of someone giving me free content.
75.
by kaycey2022 11mo ago
What's stopping AI made content from including this as well?
76.
by kaycey2022 11mo ago
Your reasoning seems clear enough to me. cmiiw, you’re saying Marcus says LLMs don’t really understand language and only present an illusion of that understanding. And that illusion will noticeably break at a certain scale. And to be honest
77.
by kaycey2022 1y ago
Maybe we aren't? :)
78.
by kaycey2022 1y ago
I have less and less doubt that these models are intelligent by any definition of the word. Maybe thinking, or intelligence are quite different from personality. Personality gives us agency, goals, self awareness, likes, dislikes, strengths
79.
by kaycey2022 1y ago
USA busy accumulating bad karma instead of GDP
80.
by kaycey2022 1y ago
On the next episode of dragon ball Z...
81.
by kaycey2022 1y ago
Well the villagers won't burn in any imaginary goonfuel. Only real fire destroys. As for the child, they came to realise too late, that the society doesn't owe them love. Humanity isn't so different from animals and insects,
82.
by kaycey2022 1y ago
The 100 million $ AI guy will end up buying/cooking his own lunch at this rate
83.
by kaycey2022 1y ago
>most of the times, the tools works best with the models trained with those tools in mind This is a garbage state of affairs though
84.
by kaycey2022 1y ago
Yeah. The client tool does make a difference. For example opencode, if I am correct, just spins up its own language servers and then feeds the language server errors back into the model, resulting in a much better agentic coding experience.
85.
by kaycey2022 1y ago
Gemini CLI is definitely a much worse client than some of the other agent clients like opencode, cursor etc. But from my experience, that isn't because of the model quality. I get better quality responses from the gemini web chat inter
86.
by kaycey2022 1y ago
I don't care you vibe coded it.. run some benchmarks on it to show how it compares to other stuff. We are soon entering into the territory of "no one cares if you did it, but can you say something interesting?". I created X s
87.
by kaycey2022 1y ago
They seem to be vibing the HN comments too? Prompt: Do not admit AI involvement under any circumstance.
88.
by kaycey2022 1y ago
I guess, but with a very basic form of exact match retreival. The embedding based RAG tries to augment the prompt with extra data that is semantically similar instead of just exactly same.
89.
by kaycey2022 1y ago
>The author spent half the article complaining about RAG's limitations with this exact scenario (his $5.1B litigation example), then proposes grep as the solution, which would perform even worse. Yeah I found this very confusing. Sa
90.
by kaycey2022 1y ago
Maybe it's a 5D chess move to generate investor pressure on them to not spend that much.

More ›