Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
IceHegel
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
11 ms
·
91.
▲
by
IceHegel
3y ago
In agree in the abstract. In practice, we are very over-indexed on these deconstruction type issues by selectively reopening old wounds, often along racial lines, which isn't leading anywhere good, defiantly not to mars.
92.
▲
by
IceHegel
3y ago
> "mob elects a leader" That's just democracy. The mob and democracy are referring to same thing with different attitudes. > "Populist ideologes often are popular, for the subset of people that stand to benefit fro
93.
▲
by
IceHegel
3y ago
This is so washed up as a take. Let’s go to mars instead of pouring energy into blaming the British for everything.
94.
▲
by
IceHegel
3y ago
democracy for me but not for thee
95.
▲
by
IceHegel
3y ago
Because he’s not popular at Harvard or the NYT.
96.
▲
by
IceHegel
3y ago
It's interesting that despite the distain for Modi among US libs, he's by fair the most popular democratically elected leader. https://www.businesstoday.in/latest/politics/story/pm-modi-m... Bukele
97.
▲
by
IceHegel
3y ago
At what point does an open border sorta make the government illegitimate? The state is just ignoring its own laws on this issue for one reason or another.
98.
▲
by
IceHegel
3y ago
A parameter refers to any trained value in the model. If you initialize it to a random number at the start of training - it's a parameter. In the context of transformer language models, that includes the weights and biases in the feed-
99.
▲
by
IceHegel
3y ago
I know a lot depends on architecture and number representation, but do people have a sense for how big a compute cluster is needed to train these classes of models from 1.5B, 3B, 7B, 13B, 70B? Didn’t Meta say they trained on 2k A100s for LL
100.
▲
by
IceHegel
3y ago
I like how this guys writes. Simple, causal, clean.
101.
▲
by
IceHegel
3y ago
Though it generally seems to have gotten worse, I've found it's gotten better at programing in Swift. It used to be really bad, now it's somewhat useable.
102.
▲
by
IceHegel
3y ago
Agree, sharing is what should be restricted.
103.
▲
The China Convergence
(theupheaval.substack.com)
3 points
by
IceHegel
3y ago
|
0 comments
104.
▲
by
IceHegel
3y ago
They did win the war with Athens…
105.
▲
by
IceHegel
3y ago
GDP as the prime measure of national prosperity is nothing short of a joke. This indicator is really just a proxy for unchecked consumption. It fools us into thinking we're tracking economic health. If you allow yourself to imagine, sa
106.
▲
by
IceHegel
3y ago
Interesting that in Mexico and Peru the wallet without money was returned more often than the one with money. The gap was larger in Mexico.
107.
▲
by
IceHegel
3y ago
I want to be able to try something without signing up. I'll sign up if it's good.
108.
▲
by
IceHegel
3y ago
Is the latex code on github? The margins are a bit off.
109.
▲
by
IceHegel
3y ago
I, uh, respectfully disagree. 1. There is still no direct Ancestor. Despite extensive sampling and analysis of wildlife, no direct ancestor or intermediate host of SARS-CoV-2 has been found. The closest virus to SARS-CoV-2 is RaTG13, found
110.
▲
by
IceHegel
3y ago
It seems like lots of places that used to build don't anymore (USA anyone) and detailed analysis about why a particular nation or sector is having issues is always helpful. That being said, I think there is also a background issue of m
111.
▲
by
IceHegel
3y ago
A PhD, like all degrees, is just a piece of paper. This piece of paper doesn't mean you'll say true things or do good science, it's just a shortcut way of letting people know "I've spent a lot of time on this."
112.
▲
by
IceHegel
3y ago
For all those criticizing the debate format as unsuitable for truth discovery, I would remind them that no form of information is immune from manipulation. Recall the highly cited The proximal origin of SARS-CoV-2 . https://www.
113.
▲
by
IceHegel
3y ago
This is probably true on a micro level but not on a macro level. There are probably a lot of things worth debating in the world of climate science, but the number of people who are allowed to even ask questions, is tiny. Basically, the only
114.
▲
by
IceHegel
3y ago
I like the tech but the design is about 6/10 atm. One great designer could make it a 9/10.
115.
▲
by
IceHegel
3y ago
It’s sad of course, but the early days of flight were also quite dangerous but they lead to rapid innovation and the stable industry of today. It’s a tragedy, but high risk activities are by definition going to have a few.
116.
▲
by
IceHegel
3y ago
It's not just the connector. It's the whole charging experience. NACS + Supercharger network is 10x better than the alternatives both in speed of charging and ease of use.
117.
▲
by
IceHegel
3y ago
> drug deaths have increased each year without fail So circling back to my original claim, it's certainly true in the US (see NIH source above). As for Portugal, it appears to be technically wrong because even though drug deaths a
118.
▲
by
IceHegel
3y ago
I've compiled some data. Regarding Portugal (which decriminalized all drugs in 2001), according to the Drug Policy Foundation, a pro-legalization advocacy group, drug deaths initially dropped (2003-2010) but have since returned to pre-
119.
▲
by
IceHegel
3y ago
This screams late empire to me. The whole liberalization idea is extremely regressive. The high functioning and rich bathe their brain in some nice chemicals while others OD or become homeless zombies because they can’t self-regulate. Does
120.
▲
by
IceHegel
3y ago
You say it is preferable today. In a way we all must agree it’s preferable. Unlike the state’s “social contract” which you cannot exit, any of us could - for the price of a bus ticket - find ourselves back in the wild. None of us go. But ju
More ›