qarl2

10 ms·

241.
by qarl2 4mo ago
It doesn't matter why. Many people choose to do it that way. And if you want them to use your software, you need to enable them. But to answer your question directly - I suspect they simply don't care, and just take the easiest
242.
by qarl2 4mo ago
Oh yes, I agree completely. But apparently Paul Graham does not - and his whim is law.
243.
by qarl2 4mo ago
Ah. That's a shame... as there is no button or indicator for "mild". Making the behavior for "I disagree" and "this is erroneous" the same seems like a problematic design.
244.
by qarl2 4mo ago
Why am I being downvoted for stating my reasonable opinion?
245.
by qarl2 4mo ago
> That does not mean "I used it via emailing it". There is no ambiguity - he was asked specifically about this. On the contrary - I think the most reasonable interpretation of his words is that he did use it via emailing it. B
246.
by qarl2 4mo ago
He said: > Author here. It was usable like any Openclaw agent. For example, I used it to ask it questions about the VPS, to summarize emails, etc.
247.
by qarl2 4mo ago
From his explanation in these comments, he claims the agent did respond in the beginning but it became too costly, so he just manually checked it after that - did the agent correctly catch malicious messages? It did not reject everything, i
248.
by qarl2 4mo ago
I think what he's saying is that initially, it could respond, and did respond with useful behavior. But after a bit the cost grew so high that he just checked whether the attacks would have worked, without doing the costly response. I
249.
by qarl2 4mo ago
As someone who has spent years in the software industry in silicon valley - My experience does not match yours.
250.
by qarl2 4mo ago
Sure. Agreed. I do it myself. But sometimes people don't do that. Sometimes people don't care about that at all. And it's not just sometimes - it's most people. So. They have their option too. And it's at the
251.
by qarl2 4mo ago
Yes, if you want to, you can do that. Understand that 99% are comfortable trusting downloads. They know that it's just as easy to sneak backdoors into source code as it is to sneak backdoors into executables. See also: XZ hack.
252.
by qarl2 4mo ago
To the people downvoting me - I'm sorry, but it's true. I know ChatGPT seems like it's non-deterministic - but that just a user preference. The only real issue is if you have many people on the same server, the GPU contention
253.
by qarl2 4mo ago
> Why would I? Because then you can install it without depending on a package manager?
254.
by qarl2 4mo ago
You can simply look at the installer by leaving off the "| bash".
255.
by qarl2 4mo ago
My dude - if you're going to trust them then you're going to trust them. You think it's hard to obfuscate shell calls from inside a built executable? What it tells us is that you're probably searching for reasons to grou
256.
by qarl2 4mo ago
You understand that computer programs are deterministic unless someone explicitly injects non-determinism in them, right? Even when they implement LLMs.
257.
by qarl2 4mo ago
> no one made that claim! You did not literally make that claim but your cost argument hinges on it. Without it, then Claude does about the same as a human and only costs $100. Apparently I'm reading your comments more thoroughly th
258.
by qarl2 4mo ago
Right. I am not claiming Claude Code creates perfect software. I am refuting your claim that using it would be cost prohibitive. I took the 10 most difficult patches from the git history - the ones that took the most back-and-forth to fix
259.
by qarl2 4mo ago
I looked at the git history. The first three years were wasted waiting for a human to pick it up. He then very slowly submitted patches over 2 years. Claude Code doesn't need to be interested to work.
260.
by qarl2 4mo ago
HackerNews is now censoring my replies. I did the math - all of these patches would have cost around $100. Let's see if they'll let this account through.