Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
ThouYS
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
37 ms
·
151.
▲
by
ThouYS
4y ago
CMake truly is a perfect fit with C++. Both suck, but nothing less sucky seems to do the job
152.
▲
by
ThouYS
4y ago
exactly that. It often seems correct, until you look up the actual answer. I was asking it how to unload models from triton server using the REST api, and the results seemed sensible. However after googling the actual API, turns out ChatGPT
153.
▲
by
ThouYS
4y ago
Isn't it ironic that we can give concrete "virtually any form", and yet most new buildings are boring cubes? Old brick houses used to have a lot more ornamentation. And if you think about it, it makes sense. To ornament, simp
154.
▲
by
ThouYS
4y ago
sweet! if you ask me, differentiable rendering is the next big thing (tm).
155.
▲
by
ThouYS
4y ago
lol, I won't donate to wikipedia this year. Why on earth do they give grants out anyway? The only thing I want is wikipedia.
156.
▲
by
ThouYS
4y ago
If only they weren't in the business of screwing their customers...
157.
▲
by
ThouYS
4y ago
fantastic post!
158.
▲
by
ThouYS
4y ago
good for you dude
159.
▲
by
ThouYS
4y ago
I don't like it
160.
▲
by
ThouYS
4y ago
I don't get the concrete failings of nim from the readme. Even though I have one of my own: They should have adopted a python-style import system and not a C-style omnibus import system. You don't always have a properly configures
161.
▲
The Trimodal Nature of Software Engineering Salaries
(blog.pragmaticengineer.com)
3 points
by
ThouYS
4y ago
|
0 comments
162.
▲
by
ThouYS
4y ago
yawn
163.
▲
by
ThouYS
4y ago
Keenan Crane simply makes the most beautiful papers out there
164.
▲
by
ThouYS
5y ago
Not necessarily, since models that predict correctly can still be wrong. Science is figuring out the real mechanism
165.
▲
by
ThouYS
5y ago
Bootstrap resampling is such a black magic thing
166.
▲
by
ThouYS
5y ago
in the case of python: "better is better" ;)
167.
▲
by
ThouYS
5y ago
I've made very good experiences with Sourcetrail
168.
▲
by
ThouYS
5y ago
what's your take on university as opposed to learning by doing? how do they complement each other?
169.
▲
by
ThouYS
5y ago
ridiculous, as if the device wasn't pretty expensive already. laughable price for the pens as well
170.
▲
by
ThouYS
5y ago
I guess pretty soon we'll wonder how anyone was ever ok with using raw non-sparsified models
171.
▲
by
ThouYS
5y ago
This! Writing your flashcard makes you think the subject through, and sometimes you need to try a few times until you get a good flashcard. This is the learning experience, the rest is just preventing you from forgetting it.
172.
▲
by
ThouYS
5y ago
I thought so too, but looking through the V repo looks like it has come a long way
173.
▲
by
ThouYS
5y ago
Intriguing idea
174.
▲
by
ThouYS
6y ago
This is the never ending cycle of "too much order" - "too much chaos". It takes a lot of experience to be able to judge how much chaos you want. That is all.. experience. I don't think any theoretical theory can tel
175.
▲
by
ThouYS
6y ago
Depends all on the implementation. DARPA seems to (have) work(ed) fine
176.
▲
by
ThouYS
6y ago
Amazing interview, but horribly "optimized" webpage. Impossible to meaningfully archive, because elements that are not visible on screen get instantly hidden.
177.
▲
by
ThouYS
6y ago
hear hear
178.
▲
Mars University
(robrhinehart.com)
1 points
by
ThouYS
6y ago
|
0 comments
179.
▲
by
ThouYS
6y ago
insect.sh does the job for me
180.
▲
by
ThouYS
6y ago
I share your sentiment, but to make it more concrete, here are a few quickly glanced things that give good/bad indications: + >500 participants + Report not only aggregated values, but also bar charts with confidence interval (you c
More ›