Show newer

RT @ellisonbg
Congrats to everyone at @Meta @PyTorch and @linuxfoundation on this launch! This is a huge step forward for PyTorch as a foundational and multi-stakeholder open source ML tool. I am honored to represent @awscloud on the board of the new twitter.com/linuxfoundation/st

Can confirm
---
RT @shookcodes
You know what happens when you keep trying to code after you're tired? Mistakes.

Ask me how I know
twitter.com/shookcodes/status/

RT @shookcodes
You know what happens when you keep trying to code after you're tired? Mistakes.

Ask me how I know

I plan events through orgmode agenda. For simple note-less recurring events I do a <2022-09-12 9:00 Mon +1w> to get it weekly, or daily, etc. For more granularity I use `org-clone-subtree-with-time-shift` (C-c C-x c). I make then for a month or year in advance and just use `COMMENT` on the upcoming section, until I need to open it.

I use Google Scholar to get the bibtex for an article along with its pdf. Is there any other comparable tool, or does Google have the corner on this service?

RT @grzm
Currently engaged in a 59+-line Slack thread re: a 2-expression bash conditional that tests environment variables and includes calls to git piping to grep and cross-platform behavior differences. Combined developer experience years is likely 30+.

Just use .

@trinsec To be fair, the short version is very handy much of the time -- when you truly need to do a tiny thing that won't be re-used. But otherwise, I am for the idea of naming things rigorously and profusely.

@trinsec , which has two syntax for lambdas. The shortest is `#(fun-call % %2)`, which is what the post warns against. More readable is `(fn liitle-known-optional-name [nice-arg-name] stuff-done)`. As she points out, using that little-known-optional-name is really helpful in stack trace debugging. Plus, it's great for readability.

@trinsec , which has two syntax for lambdas. The shortest is `#(fun-call % %2)`, which is what the post warns against. More readable is `(fn liitle-known-optional-name [nice-arg-name] stuff-done)`. As she points out, using that little-known-optional-name is really helpful in stack trace debugging. Plus, it's great for readability.

🤯 Genius is about making things simple.
---
RT @pappapez
The entire implementation of ClojureScript postwalk and prewalk. Mind blown.
twitter.com/pappapez/status/15

RT @KaliszAd
Two groups of programmers stand out and both are gravely mistaken.

The is no fast language on its own. Also, there is no practical system that is bad, slow, hard to maintain purely because it is written in a particular language.

RT @MrGung1
"If the BDD-tests we've got from our 3-amigos-sessions is not sufficient for you, you can still write additional tests."

Amazing how much can be wrong in a single sentence.

And: Hell yes, sure I'll do.

@jalefkowit@octodon.social I had suspicions but this verifies my concerns

We cannot write a program that generates randomness, because as soon as we've done so it can be used to predict that randomness. We can extropolate (within limits) randomness, mix it, & adjust its probability distribution. But randomness itself must be sourced as input!

There's dedicated circuits for this, & often we can extract randomness from other inputs as sideeffect of the digitization process.

Maybe randomness is just what we don't understand, or maybe its some sort of nonsentient deity.

"The road to callback hell is paved with #(...) lambdas."

I use follow-mode (built-in to ) to split reading buffers across modern wide screens to use all the real estate. The mode keeps the panes in sync with each other. images.toryanderson.com/follow

Show older
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.