Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
Sha1rholder
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
6 ms
·
61.
▲
by
Sha1rholder
2mo ago
ANOTHER.md YETANOTHER.md YETANOTHERAGAIN.md
62.
▲
by
Sha1rholder
2mo ago
If you feel that my explanation about vaccines as a modern way to defense diseases is intended to offend your self-esteem, go ahead. I don't waste my time on quibbling.
63.
▲
by
Sha1rholder
2mo ago
That's a weird self-defense.
64.
▲
by
Sha1rholder
2mo ago
There is definitely a relationship. But I personally believe that once the training corpus reaches a certain scale, the returns exhibit diminishing marginal effects, to the point that multiplying the data volume cannot surpass something ess
65.
▲
by
Sha1rholder
2mo ago
> That is its legitimate input. Things that make you healthier are legitimate input, not things that are "natural". Living naturally doesn't usually mean living healthier. From another perspective, humanity has already tra
66.
▲
by
Sha1rholder
2mo ago
It sounds very hypocris to want your things better and tolerate neighborhood things worse.
67.
▲
by
Sha1rholder
2mo ago
Your immune system evolves every minute with everything throwing monkey wrench to it. And it's far from perfect at all.
68.
▲
Indie dev claims Gemini leaked scarily specific details in private documents
(gamesradar.com)
2 points
by
Sha1rholder
2mo ago
|
0 comments
69.
▲
by
Sha1rholder
2mo ago
It works on Macbook but not on Mac studio? How?
70.
▲
by
Sha1rholder
2mo ago
"Oh we're so responsible. Open source unsafe!"
71.
▲
by
Sha1rholder
2mo ago
In what aspect is Playboy magazines superior?
72.
▲
by
Sha1rholder
2mo ago
What do you think botanists and ecologists were doing before large-scale tree planting began?
73.
▲
by
Sha1rholder
2mo ago
I do but I do regret.
74.
▲
by
Sha1rholder
2mo ago
A "smooth" editor with uncountable bugs and without smooth scrolling... I agree they should focus on basics.
75.
▲
by
Sha1rholder
2mo ago
> Anthropic is unusually willing to say AI might kill us all. That they are building something dangerous and they know it. > > This is not normal corporate behavior. Normal corporate behavior suppresses downside. Bro is really
76.
▲
by
Sha1rholder
2mo ago
Crazy number, crazy software, crazy company, crazy government, crazy scandal, crazy consequences
77.
▲
by
Sha1rholder
3mo ago
> Roughly everyone can now benefit from its content. You mean every Anthropic?
78.
▲
by
Sha1rholder
3mo ago
Mark TLS 1.2 as legacy means not recommend to use in the future, not abandoned or unsupported. It's not even a breaking change like this RFC
79.
▲
by
Sha1rholder
3mo ago
The article’s length planning and argumentative process are too bad to feel like AI slop. It's human slop.
80.
▲
by
Sha1rholder
3mo ago
Friendly reminder: - Capitalism took 300 years to change the world. - The Age of Exploration took 200 years. - Electricity took 100 years. - The oil industry took 80 years. - Computers took 50 years. - The internet took 20 years. - Mobile i
81.
▲
by
Sha1rholder
3mo ago
I don't see much necessity. TLS1.3 is not something that hard or costly to support. I'd suggest just mark TLS1.2 as legacy and make some practical constraints to TLS1.3 applications/implementations to avoid replay attacks.
82.
▲
by
Sha1rholder
3mo ago
For service providers, highly intelligent AI agents with broad permissions, large token budgets, and purchasing power may not be fundamentally different from humans, since both can contribute value.
83.
▲
by
Sha1rholder
3mo ago
Every technology has both pros and cons. Those who develop standards must weigh both sides rather than considering only one of them. In countries where the internet is subject to censorship, as well as in regions with underdeveloped infrast
84.
▲
by
Sha1rholder
3mo ago
Forgive me but I thought of the Drop function in Microsoft Edge when I saw the title.
85.
▲
by
Sha1rholder
3mo ago
ANOTHER.md
86.
▲
by
Sha1rholder
3mo ago
VPNs get blocked? Your VPS needs an ip update.