klysm

31 ms·

601.
by klysm 2y ago
Canvas throws away a lot though, especially w accessibility
602.
by klysm 2y ago
I think there’s still a dependency graph
603.
by klysm 2y ago
I don’t really agree with the characterization. That can matter at scale and the other thing to consider is latency
604.
by klysm 2y ago
Vectors seem like kinda the easiest thing to store in a lot of ways
605.
by klysm 2y ago
> Replicating a real file system is not that hard Ummmm what? Replicating a file system is insanely hard
606.
by klysm 2y ago
Sending your api key in the header is equivalent to basic auth.
607.
by klysm 2y ago
Inflation might need to be factored in
608.
by klysm 2y ago
How could it be a static target though? That doesn't seem practical.
609.
by klysm 2y ago
I'm curious what the specific bounds were before and after this proof.
610.
by klysm 2y ago
The use of the Fibonacci sequence is so pseudo-intellectual. It's completely arbitrary, but use of the Fibonacci sequence makes it sound smarter or justified somehow.
611.
by klysm 2y ago
That’s terrible for developers though. It’s already a moving target and having several moving targets to pick from doesn’t sound fun
612.
by klysm 2y ago
I don't see any value in having this standardized. Different products need vastly different kinds of feature flagging for widely different reasons. Just build what you need when you need it!
613.
by klysm 2y ago
I've gotten a lot better at it by being boneheaded about not giving up. By bashing my head into IAM, I've gotten pretty good at figuring out the right set of permissions, roles, etc for cross account stuff to work. It's defin
614.
by klysm 2y ago
I'd imagine it just comes down to whether or not the extensions are allowed or not
615.
by klysm 2y ago
I’ll happily die on that hill with you
616.
by klysm 2y ago
Sounds reasonable, but since the language layer has no knowledge of signal handlers or what that means, it would be a separation of concerns problem. I'm sure you could get clang to do it, but still a tricky thing to design around. Ult
617.
by klysm 2y ago
Debian has a fairly famous one: CVE-2008-0166
618.
by klysm 2y ago
The company with the most valuable weight vectors
619.
by klysm 2y ago
One of this particular class!
620.
by klysm 2y ago
I didn’t say pissing off drivers was a bad thing!
621.
by klysm 2y ago
I was about to ask if there was a term for this
622.
by klysm 2y ago
Middle-upper suburban logistical warfare. A game theoretic amalgamation bound to piss off drivers.
623.
by klysm 2y ago
A lot of my progression as a software engineer has come from decoupling my ego from the things I produce. That’s hard to do because the code you produce is a very strong reflection of your mental models and it’s justified to feel some conne
624.
by klysm 2y ago
Conceptually I expect compiling NNs to hardware to eventually be done at a small scale at first. Imagine you have a relatively simple task with a fairly small NN which is being used a low latency, mass production application. If you could c
625.
by klysm 2y ago
Yes this is exactly what rebase is for, but a lot of developers do the merge based workflow.
626.
by klysm 2y ago
Let's say I'm working on a feature branch, and end up realizing I should really be opening two PRs because this one PR is going to be two hard to review. It's much more difficult to reorganize commits on your feature branch i
627.
by klysm 2y ago
Sure, but that is not the same as open source
628.
by klysm 2y ago
Looks cool, but I’m still team everything-in-Postgres
629.
by klysm 2y ago
I expect this one to perform much better due to the headline being more comprehensible
630.
by klysm 2y ago
True, but that’s not how most sensors actually work. For example consider a weighing scale. If it says 10.1kg, why would we use a normal distribution?

More ›