sdenton4

28 ms·

481.
by sdenton4 2y ago
In fact, dragons are important stabilizing influences in dungeon economics. The hoard of gold isn't inflationary until the adventurers liberate it and start buying wands and stuff.
482.
by sdenton4 2y ago
The softmax value y is a linear combination of the vectors you're attending over: y = a1 v1 + a2 v2 + ... + an*vn where a_i >= 0 and sum(a_i) = 1. Then y is a convex combination of the v_i, and sits in the convex hull of the v_i.
483.
by sdenton4 2y ago
Many, certainly many primates. Our ability to understand animal intelligence is limited by or ability to ask appropriate questions, so we tend to underestimate intelligence outside of the primate space. Here's an excellent book on this
484.
by sdenton4 2y ago
I mean, humans do this too... If I tell an interviewee that they've done something wrong a few times, they'll have less confidence going forward (unless they're a sociopath), and typically start checking their work more close
485.
by sdenton4 2y ago
It's perfectly happy to make an imperial storm trooper riding a dragon, for what it's worth
486.
by sdenton4 2y ago
And worse, we pheasants are in constant danger of being eaten by starving peasants.
487.
by sdenton4 2y ago
They're trying to simulate real accounts by following people. Blocks are likely a strong signal for anti spam. So if the spammers notice that you block bots, they won't want to burn accounts by following you, and will instead foll
488.
by sdenton4 2y ago
One year later: "To interact with our support bot, please enable screen recording and keyboard logging." "Thanks! Now go to the settings and enter your password." "Thanks! All your Bitcoin are belong to us!"
489.
by sdenton4 2y ago
Of course, all of this pales next to the invention of the actual human brain, which was not driven by humans or AI.
490.
by sdenton4 2y ago
And only wears the underwear on 92% of days!
491.
by sdenton4 2y ago
Fusion power is just eight minutes away...
492.
by sdenton4 2y ago
The question is: For a given problem in machine intelligence, what's the expected time-horizon for a 'good' solution? Over the last, say, five years, a pile of 50+ year problems have been toppled by the deep learning + data +
493.
by sdenton4 2y ago
I think you misinterpret the 'don't know how' part. LLM's are fantastic for boilerplate (which is a big part of getting things started). From there, I might not know the right incantation for (say) handling a button clic
494.
by sdenton4 2y ago
"an output encoding and associated group actions under which the loss is invariant" What loss is that, exactly? One of the difficulties in speech processing is that we generally don't have a great model for speech quality or
495.
by sdenton4 2y ago
You can turn any lossy compression scheme into a lossless scheme by encoding the error... The lossy scheme should be aiming for low error, making the error encoding progressively smaller as the lossy scheme improves. What's harder to d
496.
by sdenton4 2y ago
Yeah, Jax provides a one-to-one reimplementation of the Numpy interface, and a decent chunk of the scipy interface. Random number handling is a bit different, but Numpy random number handling seeeeems to be trending in the Jax direction (ex
497.
by sdenton4 2y ago
Why not Jax?
498.
by sdenton4 2y ago
It wouldn't be terribly surprising. Forking everything provides a liiiitle bit of protection against things like the 'left pad' incident.
499.
by sdenton4 2y ago
One factor in the emphasis on role-playing is that "kill monsters and watch numbers go up while navigating a mostly linear narrative" is quite well addressed by video games... We go to the table to do things we can only do at the
500.
by sdenton4 2y ago
As long as the mosquitos contain real dinosaur blood, it's all good.
501.
by sdenton4 2y ago
(not to be confused with AlphaProto, which is helps with Google's core business of turning protocol buffers into differenter protocol buffers.)
502.
by sdenton4 2y ago
Let v = high violence, d = dem mayor, r = republican mayor. You seem to want to compare P(v|d) to P(v|r). We have little to no data for comparison because there are so few examples of cities with republican mayors. Those that exist are amon
503.
by sdenton4 2y ago
You're using raw counts /of cities/ in a top ten list, when the actual leadership of cities is heavily skewed towards Dem mayors. I point out that this is similar to comparing raw counts instead of per-capita.
504.
by sdenton4 2y ago
That's not what I said? Reread the comment, you might learn something.
505.
by sdenton4 2y ago
There are far more cities with Dem mayors than Republican mayors, so comparing raw counts is pretty meaningless. (Similar to counting comparing total homicide instead of per capita.) There are 10 Republican mayors amongst the 50 largest cit
506.
by sdenton4 2y ago
Hey, turns out there's a whole book about this! https://en.m.wikipedia.org/wiki/The_New_Jim_Crow And if moving pictures are more your thing, there's a movie, too! https://en.m.wikipedia.org/wi
507.
by sdenton4 2y ago
There's some argument that Einstein was in the right place at the right time. Mercury was wobbly, and there was about fifty years of non-euclidean geometry research built up, including (eg) Riemann breaking ground on differential geome
508.
by sdenton4 2y ago
It's a bit like asking why we still blame lead for lead poisoning 70 years after it was painted on the wall...
509.
by sdenton4 2y ago
Curious if you might have situations where the really good image for publication is taken from a higher quality device, while the day to day work is done with an array of cheaper devices.
510.
by sdenton4 2y ago
Are you proposing functional programming? There's plenty of examples of languages going that route (including languages with varying levels of compromise on 'pure' functionality for performance).

More ›