HN user

blake_himself

21 karma
Posts3
Comments27
View on HN

Dude, we evolved on different continents for 50,000 years. We look different. Our bodies' biochemistry differs (different medications for whites + blacks for blood pressure). Why do you think brains came out the same? They didn't, and all the hard data, and all the evidence of everyone's eyes - how black vs white, or other populations, live, shows that. It's ugly, it's not nice, people don't know what to do about it, it's a can of worms, but it's true.

There was an article somewhere recently, of someone proposing changes to CPython for just that reason. Guido had already resisted several attempts at the kind of thing he was doing, from a desire to keep the CPython implementation simple.

I voted for it before reading it - it sucks. The title makes it sound like Glenn Greenwald is saying sorry - I don't see anything like that in there. His name gets a couple of mentions, but it's about some no-name former marine changing /his/ mind. I regret my upvote.

edit: Ah, I mis-read the title, I see he's saying sorry to Glenn Greenwald. No harm no foul, but I still wouldn't have upvoted.

Announcing Rust 1.0 11 years ago

Will Rust always be faster than C?

(By 'always' did you mean 'ever'?)

That may be so, but their advertising gives the wrong impression. They should keep their claims muted until they can back them up.

That point about employees not having friends at work being at risk is true in my case - if I don't have friends where I work I don't feel at home, and am indeed more likely to leave. Which is a force for anti-diversity, since I only rarely make friends with Asians (of either sort). And I'd bet that this is the rule, not the exception.

When you are playing chess, at some point you make a mistake, you may go back several steps that were “correct” to find the one that was wrong. When you fall off a bicycle, you think of when you lost your balance. Deep learning does that. The credit assignment in a deep learning exercise can be tens, even hundreds, of levels deep.

This sounds like reinforcement learning. Anyone know what he's talking about, some RNN with 'tens, even hundreds' of levels of feedback that hasn't died away?

For NLP. Presumably you could extract it from sentences with techniques like Socher's:

http://www.socher.org/index.php/Main/ParsingNaturalScenesAnd...

There was an example in Socher's paper of a tree of logic - ands, ors, nots (not shown on that site, but in the paper) - which logic, he showed his RNN technique can represent. The OP paper I imagine is more along those lines. Being able to capture full logic like that gets you that much closer to being able to extract the full meaning from language, not just words that provide a general flavor.