bob ✅ @parataxis@icosahedron.website
best mesothelioma lawyer dallas truck accident lawyer truck accident lawyer houston loisville car accident lawyer san diego water damage. @bobpoekert on twitter
My understanding of the internet last week:
(you)
|
| safe?
|
(WPA wifi AP)
|
| evil
|
| evil
|
| evil
|
(cat videos)
My understanding of the internet this week:
(you)
|
| evil
|
(WPA wifi AP)
|
| evil
|
| evil
|
| evil
|
(cat videos)
Why does cream cheese expire but not peanut butter?
who should I vote for in the vanguard proxy election?
there's a surprisingly small amount of stuff on doing optimization over bit vectors instead of vectors of reals. there's a bunch of stuff on SAT but that's not the same thing (I'm doing unconstrained optimization)
I'm probably going to write a bespoke optimization algorithm for my particular problem but doing that feels wrong, like I must be missing something
you can see the path of totality in traffic patterns!
on a traffic map that i helped build. AAAAAAAAH!!!
Prompt:
Imagine that people had a sixth sense that allowed them to feel the presence of all the objects in their environment at all times, even the ones that they couldn't see, feel, or hear. Some people are born without this sense. These people are considered disabled. Design an assistive technology for these people.
speech takes too much brainpower to process to be ubiquitous like that. what would an audio equivalent of shorthand look like?
if apple designed a system of primal pre-linguistic grunts and squeals for conveying information what would they sound like?
AR glasses with visual overlays are a ways off but the tech exists to make AR hearing aids now
there are a lot of interesting things you can do with bone conduction headphones + an IMU + OpenAL-style 3D sound
giving directions by having a "pong, pong" noise that you walk towards is the first thing that comes to mind
you could also give objects in the physical world auditory "auras" by putting faint audio sources in them
(you'd need to be doing SLAM, too, to be able to do that, ofc)
why did I let the keras documentation convince me to use rmsprop instead of adagrad?
o gods of gradient descent, please accept this offering of a regularizer, that you may bestow on us low validation error
reminder that empathy only leads to compassion for members of your tribe; it leads to schadenfreude for members of other tribes
Narratives: they're bad
there was an opportunity to make identity atomized and contextually-dependent and the industry actively decided to do the opposite
social media was a mistake. or at least, having user profiles and making them central was a mistake
PERSON A: [tribal signifier]
PERSON B: haha, yes, I also [tribal signifier]
people who can't relate to other people without binning them into some category first are extremely fucking exhausting
