sophie (@sophemera) header
sophie (@sophemera) profile picture

sunset painter • founding design eng @hey_aristotle • waterloo se

sf ོ
Joined December 2024
Photo by sophie (@sophemera)Photo by sophie (@sophemera)Photo by sophie (@sophemera)Photo by sophie (@sophemera)Photo by sophie (@sophemera)Photo by sophie (@sophemera)Photo by sophie (@sophemera)Photo by sophie (@sophemera)Photo by sophie (@sophemera)Photo by sophie (@sophemera)
most AI products want to look like the future, which usually means high tech, neon glow, cold. i knew i wanted Aristotle's brand to convey a real philosophy... so like all good art, i started with a feeling! here's my whole branding journey (1/)
13
3
7
105
9,130
CS Lewis believed that we write not to explain things, but to understand things. The writer goes to the blank page to crack a puzzle that matters to him. This is why streaming slop and AI art feel so empty: there is no artist behind the scenes trying to figure something out
18
202
17
1,430
32,038
My 20 Favorites
621
20,045
1,108
306,464
13,658,999
We usually think of “emotional unavailability” as being cold and withdrawn. But highly agreeable people are also emotionally unavailable. They, too, withhold their true emotions, they just do it in service of being liked.
25
308
60
3,545
118,633
Introducing: Timbit Viewer Explore the beauty, texture, and design of every timbit, do it at your own pace, take your time and go to the next one only when you’re ready. stuff.internetvin.com/timbit…
51
30
17
426
24,924
Whenever I hear someone say ceramic tiles are outdated, I think of the wonderful work of Riwal Ceramika.
8
104
6
930
19,669
Sleeping Beauty borrowed its palette from a medieval manuscript. Eyvind Earle looked to the Tres Riches Heures, the early 15th-century book of hours, for the cool blues, sharp greens and flat gold light of the film's forests. Eyvind Earle after the Limbourg brothers, Sleeping Beauty palette studies. 1959 / c. 1412-16.
1
46
5
369
6,732
These four vintage photographs show a 1920s Art Deco mannequin head captured under four completely different lighting setups to demonstrate how direction and shadow alter perception.
4
75
3
783
15,002
Tall twin waterspouts over San Francisco Bay near Brisbane, CA lasting several minutes. Never stop chasing. #cawx
89
882
144
6,582
280,373
sophie retweeted
software factories feel like a temporary problem. the models will learn to build them for you the more interesting question to me is how to build a workspace where agents can collaborate with you
592
345
189
7,573
573,898
sophie retweeted
sketchbook study with a cheap little multicolor pen i got :3
433
8,765
610
111,176
2,926,398
sophie retweeted
the ultimate challenge is to train yourself to expect surprise. it is almost a contradiction in terms, to set your expectations so as to have them defied, but it does seem to be possible, and you can see this in people who constantly and nimbly adapt to chaos of the world
2
2
1
20
1,096
how lucky i am at this moment in time to be surrounded by objects that were designed by humans
5
130
masked decoding on gemma 3 12b to control text rivers if a token isn't the length i want it to be, i mask it out added a lora for width tags so the model is aware of text length + coherence RL afterwards
147
354
99
6,173
811,734
all of my favourite writers are at inkhaven, so every day i am Feasting
1
4
436
another banger take in @dynomight7's latest
1
151
sf t&t is basically heaven.. i got so nostalgic and homesick i started crying
11
4
6
130
13,977
i grew up going to a t&t every week, it's canadian :)
4
262
sophie retweeted
one thought about the proliferation of AI art is that communication derives much of its significance from what is *not* said. when you communicate you choose to say one thing and choose not to say others. ted chiang, eg, has a famously small body of work, he's in some sense only chosen to say a handful of things. there's a certain meaninglessness to the mass production of even supposedly high-quality AI art (or math, etc) that i think derives partly from this; the sheer volume of the slop wave makes it hard to tell what, exactly, is being *said* at the end of the day
9
1
1
79
3,209
sophie retweeted
80
2,577
91
31,460
453,775
sophie retweeted
i'm finding it increasingly important to not ask LLM-s for copy / visual mocks because the basic patterns and cliches will poison my brain and make it harder to design / write beyond the average
61
32
6
626
21,777