genderfluid gold star enby: art, jokes, AI and NFTs π₯Ή. in renaissance. quite likely sexy. good blog. | she! π = enby + pan
pfp by: @arlekin_julia β¨
π³οΈββ§οΈ!
Sometimes you take a photo, and you just remember it because it was a special moment in your learning--as a photographer.
Special light, quick moment, you're fully there, and keep going back to it.
Coincidentally, this cat's ass is living in my head rent-free for the last year.
@Aella_Girl Overall, finding publishable, relevant, engaging, and real findings in statistics seems to be significantly too demanding for anyone to make a day job out of it.
The work done here is *questionable* in terms of producing real findings.
(also iffy, given female hypergamy theory)
@Aella_Girl > I sourced photos from a few subreddits for rating nudes, and from porn sites.
(from s*bstack)
Hm! It's a selection bias influencing the findings. Women's beauty has much more social capital on it right?
> 2 points below
Same as normalization at -2 points! Not a clean finding.
gm fam! πΏ
most of us feel we *should* be good to ourselves and others. but think of a person in a home, wanting to get some plants.
the first sprouts often don't make it. the geometry, the light, the water, the owner, are all a mess.
our society is both kinder, and less kind.
@meowbooksj > tfw it is better to cut the wafer into long strips, and try to assemble an even larger megachip architecture from chiplet-like parts and quality interconnect, which would allow for adjusting yields per segment, neater surfaces, and higher SoC throughput; while paying for the IC
@willccbb I buy that!
One of my first experiments with the ChatGPT web app, was trying to get better tweet imitation of a target by including the rationales and logic behind my own tweets in plain language.
IIRC, it seemed to me that it might be working.
Maybe imitation ICL is too hard!
@kalomaze π₯Ή If I understand it correctly--the geometric structures are loss manifolds-ish of the two problems, and the iterated structure is... batched learning, and/or chain of thought?
I had a more bounded 'theory' why AI works.
Parsimonious is a hard claim...
https://t.co/RFdCvRC5XI
Optimization for next-token prediction turns into language understanding when "next word prediction" is a proxy axis with enough correspondence to the hidden structure of language, thought, and the world; so the best way to reduce loss is to learn good representations of meaning.
maybe 'be good', or 'manage', is the phrase you can give yourself.
other times, it won't help. sometimes, the thing that is causing you pain doesn't like, know, or understand you.
society won't let 'be good' be enough.
i have no motto for all lives.
just a gm and good luck! π
It's with great over-excitement that I announce formally that I have received two lovely gifts by artists.
Changes: @arlekin_julia
Selfmeme: @0xMahnaz
As Empress of the Human Realm wearing the symbolic crown can get sort of difficult. And I cope by changing my PFP pretty often.
@0xGrainX digital art is art available and made to be viewed and shared digitally, or at least that's how i used it. it may also be made by digital means. π
a piece of art can be both digital and physical!
gm fam! π«Άπ»
my thesis as a collector in the nft space is this:
digital art and cryptocurrency work together very well, and i want to share and nurture my care, respect, and support for real, human, and creative artists and art.
personally art moves and tends my heart every day.