Oz ⚡️ (@UTXOz) header
Oz ⚡️ (@UTXOz) profile picture

Your personal Self-Custody coach Physicist & Software Engineer building with Bitcoin, Nostr and AI. Building @BitTaskerdotcom Creator of https://teapaw.org/t.co/qNiNZoSAd8

Milky Way 🌌
Joined April 2022
Photo by Oz ⚡️ (@UTXOz)Photo by Oz ⚡️ (@UTXOz)Photo by Oz ⚡️ (@UTXOz)Photo by Oz ⚡️ (@UTXOz)Photo by Oz ⚡️ (@UTXOz)Photo by Oz ⚡️ (@UTXOz)Photo by Oz ⚡️ (@UTXOz)Photo by Oz ⚡️ (@UTXOz)Photo by Oz ⚡️ (@UTXOz)Photo by Oz ⚡️ (@UTXOz)
CypherFlow.ai is LIVE! 🎉 Your private AI playground: ✅ #Bitcoin micropayments (Tokens/sat) ✅ NIP-60 Lightning & Ecash wallet ✅ NIP-44 encrypted chats No subscriptions, only AI for Sats. 👇 🌐 cypherflow.ai #AI #nostr #ecash #cashu #lightning
3
2
14
1,087
Oz ⚡️ retweeted
🇯🇵🎵 Il Y A 47 ANS aujourd'hui sortait un HYMNE de la scène pop japonaise : « STAY WITH ME », de Miki Matsubara. La chanteuse, décédée il y a 22 ans d'un cancer, a laissé derrière elle un hit planétaire à plus de 500 MILLIONS de streams sur Spotify.
25
632
81
6,180
1,643,860
This is the way ☸️
Steering one attention head causes Qwen to turn into Buddha
1
57
I am sure that is real money
BREAKING: 🇺🇸 Over $3 trillion erased from the US stock market today.
1
86
Oz ⚡️ retweeted
肯定
39
1,023
34
5,733
95,027
Oz ⚡️ retweeted
adding her to my tailnet to show that i'm serious about the relationship
38
293
22
3,431
74,840
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
2,359
7,276
5,725
55,926
9,843,366
Oz ⚡️ retweeted
Sorry what the fuck
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
16
21
3
883
113,771
Oz ⚡️ retweeted
132
2,889
91
30,855
594,008
Just used my first bagged full reset on #codex Token printer go BRRRRRRRRR
86
Not a single one mentioned shipping 2000 prs in a month
Oh wow… not a single one mentioned anything about work. Now excuse me while I ball my eyes out.
70
The universe doesn't care that us, psycodelic monkeys, have 10 fingers
meanwhile, over in physics land the fundamental unit of length is this monstrosity the universe does not care about your aesthetics
45
Oz ⚡️ retweeted
human: "god, why did you do that to integer multiplication?" god: "do what?" human: "nlogn. didn't you craft math to be elegant? beautiful?" god: 'the fuck's a log?"
32
180
8
9,674
294,219
Oz ⚡️ retweeted
“there aren’t any cryptography results” “there aren’t any cryptography results”
30
337
28
6,402
227,659
How about solving the millenium problem of Bluetooth headphones connecting to the right fucking device? @OpenAI
1
1
83
Bro, the only valid technical indicator is Insider Information™ , don't get fooled, you are not on the right telegram group
1
79
Sadly, neither am I. So just stack sats and be humble
15
Oz ⚡️ retweeted
opentunnel - public urls for anything running on your machine tunnel services exist but this one is different - e2e encrypted, relay can't see shit - sdk to embed this functionality in your apps will be integrated into opencode tomorrow
121
148
27
3,284
141,539
Oz ⚡️ retweeted
This made my morning. ☀️ Recent results like 3-sum in n^1.999 and multiplication in n(logn)^0.999 many many worry that maybe mathematics are not so beautiful after all and "it's all square packing" when you dig hard enough. But as so often in math, an ugly result just means you picked the wrong perspective.
7
17
1
284
24,637
2
36
2
392
6,632
Oz ⚡️ retweeted
This is what opus 5.5 says about the “surprising” results for physics from these math proofs
We’re releasing a broad range of new mathematical results produced by an internal frontier model. We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and public recommendations to inform how we release these results. github.com/openai/math
15
18
12
257
51,104