Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
jgilias
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
10 ms
·
61.
▲
by
jgilias
4mo ago
inb4 someone calls Bessent to explain how this can be used in fentanyl production.
62.
▲
by
jgilias
4mo ago
“You’ll own nothing and be happy” Thanks, but no thanks!
63.
▲
by
jgilias
4mo ago
Oh boy. I see this so much.
64.
▲
by
jgilias
4mo ago
Cool. But. Most of the “impressive” stuff is not “the model” but “the harness”. Spinning up the subagents and teams of lower models, letting them explore, do adversarial coding. It’s all in the harness. Granted, Mythos might be better at th
65.
▲
by
jgilias
4mo ago
One could argue that scalability matters more when software is expensive to make, as you need to reuse it to make the cost worthwhile. So, for context, two data points I have that make me want to argue for the opposite side. First, some tim
66.
▲
by
jgilias
4mo ago
That was in the world where making software was _prohibitively_ expensive.
67.
▲
by
jgilias
4mo ago
> And we all know the demand is drying up. I don’t think the data really supports this? Last I checked at least.
68.
▲
by
jgilias
4mo ago
If they then stop fulfilling any of the inclusion criteria (profitability, say), they should be removed, yes. Isn’t that obvious?
69.
▲
US Military personnel are being targeted using location data
(reuters.com)
14 points
by
jgilias
5mo ago
|
4 comments
70.
▲
by
jgilias
5mo ago
We’re going to be fine. I see daily what non-engineers that build stuff with AI come up with.
71.
▲
by
jgilias
5mo ago
Nah, he’ll just get some deal where the FBI can’t investigate him and his family for anything ever. Yay for democracy and rule of law.
72.
▲
by
jgilias
5mo ago
Sometimes writing a lengthy answer is absolutely necessary to convey stuff correctly and precisely. But if a human is writing it, every sentence has a meaning, and it might just be worth reading. Because a human is unlikely to waste their t
73.
▲
by
jgilias
5mo ago
But it does. A whole class of runtime errors you can trivially produce in safe Go — null pointer dereferences, unchecked type assertions, missed enum cases — are unrepresentable in safe Rust. Also, the type system is a lot more expressive,
74.
▲
by
jgilias
5mo ago
Yes, but it comes with much better “built-in” guardrails to rein in the autocomplete. Especially if compared to something runtime-surprise-prone-if-lovable like Ruby.
75.
▲
by
jgilias
5mo ago
Not sure about the angry part, but vague sometimes works really good. The important part is to have enough good context pushed into the context window beforehand (codebase explorations, docs, etc). Then a vague prompt of the general directi
76.
▲
by
jgilias
6mo ago
I agree, yeah
77.
▲
by
jgilias
6mo ago
If they basically generated a desktop for themselves, what’s the chance they didn’t generate the article? I think pretty slim. Also, reading it is probably not the intended use. It’s probably: “Hey Claude, give me a TLDR of this”
78.
▲
by
jgilias
6mo ago
Out of curiosity, what are the other two realms? (I assume it’s two)
79.
▲
by
jgilias
6mo ago
It’s not sorcery tech at all. Nothing in their “goblin post mortem” is surprising the least bit if you have a working high-level mental model of what an LLM is. It’s a fancy autocomplete that takes a bunch of text in and produces the most “
80.
▲
by
jgilias
6mo ago
And the vast majority of current and past countries with the word “democratic” in their name weren’t actually democratic.
81.
▲
by
jgilias
6mo ago
If you know the ID, I think you can get Lego ship it to you.
82.
▲
by
jgilias
6mo ago
OpenAI is not the worst it could or would send it to.
83.
▲
by
jgilias
7mo ago
Not my project! I’m just monitoring it from time to time, as I’m really curious as to when/if someone manages to prompt inject it!
84.
▲
by
jgilias
7mo ago
And there’s no reason why they couldn’t vibe fix the issues if there was a process to report the bugs. Fixing issues like that could also be something that’s fully automated. Provided there’s a good test suite (not a given).
85.
▲
by
jgilias
7mo ago
Not sure about Waze, but the combination of AI + TikTok/shorts/etc sounds like “not great” for long term brain health.
86.
▲
by
jgilias
7mo ago
The point was - some people are running hugely outdated stacks for “reasons”.
87.
▲
by
jgilias
7mo ago
Go ahead, try it out: https://hackmyclaw.com/
88.
▲
by
jgilias
7mo ago
I remember reading somewhere that Deutsche Bahn is running Windows 3.1 for something still?
89.
▲
by
jgilias
7mo ago
This page linked in the bottom explains a lot: https://www.kicad.org/blog/2025/06/KiCad-and-Wayland-Support...
90.
▲
by
jgilias
7mo ago
Getting more convinced day by day that bureaucrats of all countries are all pretty dense.
More ›