if i lived in sf i would have been a “chinese model” for halloween. but i am not. you are welcome
2
0
72
2,835
im usually pretty brave on a citibike but biking down 5th ave/wsp during halloween in full costume might have been peak Courage for me
4
0
30
1,797
it’s november 1st, it’s god’s day. i’m counting shapes in the sky, i’m handwashing cashmere. this life is beautiful for it is mine and mine alone
1
0
63
2,496
sometimes i ask kevin what its like to have the coolest pm ever and he's like idc look at my deepinfer megamoe h200 kernel
1
1
63
3,812
because of my undying hatred of fast fashion my halloween costume consists of a cut up rick skirt and tom ford gucci bamboo
8
0
181
5,042
a week straight of perfect weather in new york city i know god will forsake me soon
10
83
1,464
22,360
brainstorming about clef merch and i was telling @kevin_flansburg about how the chore coat is overplayed. so now we're making clef napoleon jackets
1
0
40
2,567
michelle retweeted
the scariest thing you can be for halloween is brutally honest and emotionally vulnerable about what you want
9
249
1,831
34,082
bleach taught me everything there is to know about aura. a multi-generational run for tite kubo. greatest of all time
16
1
75
4,201
michelle retweeted
When I said hunch could learn in real-time I wasn't kidding. It's like teaching a kid!
Behind the scenes, @Cloudflare’s Clef is helping refine the local model so it learns in real time, while @googlegemma’s EmbeddingGemma2 runs live on the CPU. 🧵
Behind the scenes, @Cloudflare’s Clef is helping refine the local model so it learns in real time, while @googlegemma’s EmbeddingGemma2 runs live on the CPU. 🧵
It's been a while since I showed any Jev stuff but I've been cooking on a way to use @GoogleDeepMind's EmbeddingGemma2 as an insanely fast real-time /v1/systemone endpoint (20ms on text, 100ms on images).
It's trainable in real-time, can distill an upstream teacher model, and is Hard to believe a @googlegemma model of this size can do so well with the right harness!
29
57
702
29,945
