my favorite podcast seems to be going through a coldplay phase and i'm not here for it. :(

AI I'd like to see Show more

sf transit fail Show more

sf transit fail Show more

ugh. the partner of a friend of mine has fallen for a windows tech support scam to the tune of $400.

now the scammers are trying to double-dip in order to "process a refund."

sorting this mess out is not what i'd planned for today, but this shit is annoying/frustrating/enraging.

ah well, maybe the scammers will learn a valuable lesson. never know who your victim might count on as a friend and advocate.

idly wondering who now owns carter's old peanut farm

"Human artists can’t draw a billion beautiful snowflakes in ten minutes, but it’s pretty common to have them fall out of the sky."

people.well.com/conf/inkwell.v

kid post, american gun nonsense Show more

please don't make me be the one to feed all the previous years' the-well-state-of-the-world into some neural net to see what pops out.

kid post Show more

i'd watch the fuck out of a movie detailing how alexander godunov's character in the money pit became his character in die hard.

so this just crossed my tl: twitter.com/dakami/status/1073

...and the three words my eyes first fell on were:

NAKED
WARFARE
MEAD

so uhh, you heard it here first. 2019 will be a year of bacchanalia.

postscript:

huh. ars posted about this very subject today:

arstechnica.com/science/2018/1

and the phrase "smoke and mirrors" came to me while I was thinking more about this, so I checked my gphotos for pictures of smoke and was similarly disappointed.

seems like you can simply file smoke and mirrors under "light plays tricks" - which i guess should've been obvious, as that's kind of the point?

in classic affirmation bias, i take this to mean i'm on to something and that this is a genuinely hard problem.

otoh, as i mentioned to a friend while discussing this last week, maybe this is already a solved problem and i just haven't gotten around to reading all the siggraph papers from years ago.

anyone know? anyone actually working on this?

so today it occurred to me, hey, i've got tens of thousands of photos with google and they do image recognition. let's see what they find when i search for "mirror."

there were under 20 results and most didn't contain mirrors. they were windows, shiny metal or laminates that reflected light but not image, and open doors.

half of the mirrors they found were side-view mirrors off of cars and were probably classified based on shape alone.

i've been thinking about image recognition and machine learning recently. probably in equal parts due to tumblr's porn ban and the whole direction of the self-driving car market.

specifically, i've been stuck on the concept of mirrors and how to tell them apart from doors or windows (especially when items reflected in them don't appear elsewhere in the image, so no edge detection).

kid post Show more

kid post Show more

things that shouldn't need to be said: if someone dies in your custody, you're a pretty fucking shitty custodian.

Show more
Mastodon

Follow friends and discover new ones. Publish anything you want: links, pictures, text, video. This server is run by the main developers of the Mastodon project. Everyone is welcome as long as you follow our code of conduct!