hagen_dogs

fluid + liquid = solid (85%) -- didn't expect that

blue + red = yellow (87%) -- rgb, neat

black + {red,blue,yellow,green} = white 83% -- weird

really good markdown editor, overwhelming amount of community plugins, all data's local and stored in plaintext. app itself is free, easy syncing option is the $4/month (with afew other nice to haves) though there are ways to do it for free (syncthing, but that's changing with the ST team not supporting andriod anymore and apple is weird too).

I've now been using it for ~3 years and I do everything in it. wonderful, wondeful tool

iiuc, it's both. drag / paste an image into obsidian it will automatically be stored in your attachments folder (regardless of what note you're in) and a preview is placed in the note itself

Probably pointing at exploding bottles; I've made probably 40 batches of ginger beer and when I would 'burp' the bottles I never had an issue. Decided once to try out not doing so and had six hours of clean up. But otherwise you're probably fine. Just let out carbonation morning and night, it takes a minute or so.

vim in obsidian?

I think it came from this paper, TinyStories (https://arxiv.org/abs/2305.07759). iirc this was also the inspiration for the Phi family of models. The essential point (of the TinyStories paper), "if we train a model on text meant for 3-4 year olds, since that's much simpler shouldn't we need fewer parameters?" Which is correct. In the original they have a model that's 32 Million parameters and they compare it GPT-2 (1.5 Billion parameters) and the 32M model does much better. Microsoft has been interesed in this because "lower models == less resource usage" which means they can run on consumer devices. You can easily run TinyStories from your phone, which is presumably what Microsoft wants to do too.

I'm curious to know if IA has ways of detecting whether archived pages were deleted. If rewriting history has big implications, doing so for "the internet" would similarly be big.