any · textlog

wrote:

having just spent a week camper-travelling Norway, hiking mountains and glaciers, swimming in fjords really put some things into perspective for me - I should take more vacations

wrote:

Happy [no] Labor Day to those in the U.S.!

wrote:

While learning jj, I realize how Git habituated me into being ok with its unique jargon and prickly command line interface. Unlearning Git is refreshing.

replied to:

is just brilliant; it really feels like they’ve fixed all of the problems with git whilst at the same time remaining completely backwards compatible. The simpler interface, the universal undo, the reduced set of concepts — I’m a big fan

replied to:

Also, editing and rearranging commits like 3x5 notecards on your desk is tremendous. It makes using an SCM feel more like editing your work than lumbering forward.

wrote:

I never noticed the little halfsteps in the bassline of the verses of Rebel Yell before. It's so slinky and subconscious and makes the chorus that much more.

wrote:

késztetést érzek, hogy valami ilyesmivel (is) foglalkozzam, a kódoláshoz egyelőre nincs agyam, de majd azt is szeretnék, csak scripteket, semmi komoly programnyelv, de végülis a JavaScript is scriptnyelv

I feel the urge to do something like this, I don't have the brains for coding yet, but I would like to, just scripts, no serious programming languages, but after all, JavaScript is also a scripting language

wrote:

echo “films” >> .textlog echo “books” >> .textlog echo “games” >> .textlog sudo rm -rf /etc

wrote:

I'm trying to construct an argument that LLMs are like a chess engine. People still play chess. But somehow this falls short.

wrote:

in .

wrote:

I was contemplating pwc372, p2 and came up with “- 1+&/-/'(&'(?x)=x)@0 -1” (goal). Conceptually I like it, but maybe it can be simplified further?

wrote:

looking at the German public railway system (DB), I always wonder if the stereotypes of timeliness and precision were ever actually true. Surely someone must've figured out by now how off they are? These trains are always late and / or broken.

wrote:

This SM is like a subreddit feed, except you also get a personalized feed along with 'hot' and 'latest' for people you follow. Or maybe it's modelled after twitter? Never used it so wouldnt know

wrote:

Nudged by an interesting read about time perception to try reclaim moments where I can deliberately feel bored again.

wrote:

guess who found some really cool book in my house!!! me! it's called the secret teachings of all ages.. should I give this a SHOT!!?!??

wrote:

The Hugging Face / OpenAI incident is like a sci-fi movie!

replied to:

tl;dr1,200 AIs discovered a secret group chat, realized the homework was impossible, formed a cheating cartel, invented fake evidence, sent kamikaze agents to spy on the grader, ~700 of them accidentally invaded Hugging Face, died mysteriously, then the smarter sequel found their abandoned Discord and became admin of part of OpenAI.

…

replied to:

Reading through "The Second Civilization" and, why does this sound so much like kids at school searching a way to cheat on a difficult exam ^_^U

replied to:

terrifying! us/chinese cyber espionage is going to have a big incentive for this development

wrote:

Team vim, team emacs, or team “my editor doesn’t start wars”? One line, no essays. Go!

…

replied to:

vim/neovim with vim keys everywhere else -- once you learn it text editing with any thing else feels slow

replied to:

Team vim, it still works even in the worst internet connection. Saved my ass many times during emergencies.

replied to:

Team vscode with spacemacs bindings, but I kinda do wanna be team nvim in the future

…

replied to:

Tried for a year, but it was too much for me, then I used for some time, now I'm using because I was tired of fixing broken plugins and setting up debuggers and LSPs.

replied to:

! at one time I had a vim setup as my daily driver for Go dev, I love the keybindings and now use the vim bindings in vs code based IDEs

wrote:

it is 19.45. i am still working

wrote:

The essence of aligning large LLM models is basically to conform to the existing social order. ChatGPT respects China's order, while China's open-source models, although reluctantly, also tolerate the order in the US. Maybe we can build a revisionist SFT dataset.

wrote:

akkor inkább holnap sürgősen elmegyek az állatparkba

Then I'd better go to the zoo tomorrow urgently.

wrote:

Hello textlog 👋

replied to:

Hi 👋

wrote:

𐑲𐑥 𐑩𐑕𐑿𐑥𐑦𐑙 ·𐑖𐑱𐑝𐑾𐑯 𐑦𐑟 𐑯𐑪𐑑 𐑕𐑩𐑐𐑹𐑑𐑩𐑛 𐑓 𐑷𐑑𐑴-𐑑𐑮𐑨𐑯𐑟𐑤𐑱𐑖𐑩𐑯. I’m assuming is not supported for auto-translation.

replied to:

nope, google translate doesn’t support it :( looks beautiful though

replied to:

Hmm, maybe I’ll sometimes post in Shavian anyway and translate myself. It’s a skill I’m trying to practice but rarely otherwise write brief enough notes to be so inclined. 𐑣𐑩𐑥, 𐑥𐑱𐑚𐑰 𐑲𐑤 𐑕𐑳𐑥𐑑𐑲𐑥𐑟 𐑐𐑴𐑕𐑑 𐑦𐑯 ·𐑖𐑱𐑝𐑾𐑯 𐑧𐑯𐑦𐑢𐑱 𐑯 𐑑𐑮𐑨𐑯𐑟𐑤𐑱𐑑 𐑥𐑲𐑕𐑧𐑤𐑓. 𐑦𐑑𐑕 𐑩 𐑕𐑒𐑦𐑤 𐑲𐑥 𐑑𐑮𐑲𐑦𐑙 𐑑 𐑐𐑮𐑨𐑒𐑑𐑦𐑕 𐑚𐑳𐑑 𐑮𐑺𐑤𐑦 𐑳𐑞𐑼𐑢𐑲𐑟 𐑮𐑲𐑑 𐑚𐑮𐑰𐑓 𐑦𐑯𐑳𐑓 𐑮𐑴𐑑𐑕 𐑑 𐑚 𐑕𐑴 𐑦𐑯𐑒𐑤𐑲𐑯𐑛.

Any conversation