make3

12 ms·

391.
by make3 2y ago
you know how people judge you harshly if you're mean to service employees, I think that colleagues and professional acquaintances will expect the worse out of you if you say bad things out of other people, because they're also in
392.
by make3 2y ago
using Jython prevents you from using the regular Python ecosystem though
393.
by make3 2y ago
because it's a subjective very negative judgement about another human, and in polite conversation one should assume baseline positive things about other humans. it's basically like calling someone an asshole. as HN as a culture of
394.
by make3 2y ago
"generalize to its dataset" is a contradiction, especially as these models are trained in the one epoch regimen on datasets of the scale of all of the internet. if you think being able to generalize in ways similar to the whole of
395.
by make3 2y ago
also, not really true, right, even though it sounds intellectual and strong to say. these algorithms are trained to generalize as best as they can to unseen text, and most often don't ever see any data point twice, except for data that
396.
by make3 2y ago
the tensorflow and jax thing is 100% over, everyone uses pytorch both in academia and the industry, at least in the western world
397.
by make3 2y ago
@dang why is my post flagged while what I said is true and relevant?
398.
by make3 2y ago
I would assume that lawyer companies might want to take this on without charging the person just because the potential upside of winning the lawsuit or settling it is so big. I agree with the general case that the legal system is really not
399.
by make3 2y ago
why is it crazy? Java, c-sharp, and more all have normal threading enabled, why couldn't Python? clearly it's going to be massively useful.
400.
by make3 2y ago
women who are comfortable with/not grossed by an app like seeking will be terrible I think, comfortable with being a product/ service that men have to pay for, would not want to date men who would have to do something like that wi
401.
by make3 2y ago
the first n bytes are likely by far the most often accessed in practices, specifically for sorting & filtering, etc. Storing them inline is likely a huge optimization for little cost.
402.
by make3 2y ago
that's just crazy bs, starting from open source code and adding specific features needed for a project is a very common strategy, doesn't mean at all that the tool wasn't good to begin with
403.
by make3 2y ago
That's not the GitHub concept / almost trademark of "fork" anymore though, which is what your parent was talking about
404.
by make3 2y ago
things are so interdependent that in this scenario you might now just end up crashing the system if either Windows or Ubuntu are down instead of just the one of them you chose
405.
by make3 2y ago
you don't mention the top one, pytorch, which is behind most AI
406.
by make3 2y ago
Hugged to death from my perspective. Here is a backup: https://archive.ph/kOE3Q
407.
by make3 2y ago
but then you can't pip install it, & you have to re build your whole environments, etc
408.
by make3 2y ago
it would be fun to make an open-ai gym wrapper on this & train an rl agent to hit the corner every single time
409.
by make3 2y ago
follow the format
410.
by make3 2y ago
99%+ of people aren't writing kernels man, this doesn't mean anything, this is just silly
411.
by make3 2y ago
knowing how careful the NumPy devs are, this was likely a very well pondered decision & all of these deprecations likely have been announced for a long time. Seeing knee jerk reactions like this is annoying.
412.
by make3 2y ago
the free version has pretty low bitrate as far as I can tell
413.
by make3 2y ago
people talked about this elsewhere, the AI part is fully optional and disabled by default
414.
by make3 2y ago
it would be cool if the url was just the chain of the moves. then you just paste the move chain in the URL and you get the board state.
415.
by make3 2y ago
(A H200 is a five digits datacenter GPU without a display port, it's not what they mean by PC, but your general point still stands)
416.
by make3 3y ago
no it empathizes the importance of training smaller models for longer, like the Mistral "overtrained" models
417.
by make3 3y ago
I'm white myself, but people focusing on the one tiny thing that could be seen as racism against whites but is maybe accidental when there's so much extremely serious non-white racism everywhere every day since forever is just of
418.
by make3 3y ago
no, AI like chatgpt delivers fuck tons of content by having been trained on it but that content, that cost money to produce, doesn't get the money back
419.
by make3 3y ago
rag is also so much faster and cheaper and energy efficient than running a giant transformer over every single document, it's like not remotely close
420.
by make3 3y ago
They just see things that help disadvantaged people as taking away from their conditions, without understanding that their condition is a privileged one

More ›