@unsignedbytes

Do you have enough compute for what's coming?

 Florida
github.com/UnsignedChad
Joined May 2026
  • Tweets1.6K
  • Following291
  • Followers124
217Media
Include

Only show posts containing:

Exclude

Hide posts containing:

Time range
-
Minimum likes
I'm trying meshy out for my game. Here's a referral link for free stuff. meshy.ai/?utm_source=meshy&u…

AI 3D Model Generator: Create 3D from Text & Images | Meshy

Meshy's AI 3D model generator turns text and images into production-ready 3D models in 20-30 seconds. Export to FBX, OBJ, GLB, and STL formats. Try Meshy free.

meshy.ai
23
Replying to @jun_song
I'm 4 deep and my only regret is not getting them sooner.
2
24
1,067
Oh xitter.cf/januarycomputer/status…
allegedly! 🔡
@januarycomputerOct 20
CONTEXT: he runs a company that rents out compute
1
36
Replying to @elshayib_
@elshayib_
1
2
950
🥳
Vals AI
@ValsAI23h
We found DeepSeek V4.1 Flash 1017 beats GPT 6 Sol on our multi-step offensive cybersecurity task. Sol was tested via the Daybreak trusted access program. We have been building a cybersecurity benchmark based on simulated corporate networks. This is a valuable indicator for the cyber-risk a model poses if in the hands of a bad actor. We plan to make the full evaluation public in the coming weeks but wanted to share our first immediate finding here.
1
43
Doing math by hand now feels like digging a tunnel with a teaspoon.
4
13
3
167
5,420
Replying to @pumamethod
AI slop
2
94
Replying to @juanjovn
Lmao
547
I got this $400 NAS off facebook marketplace for $30 bcs the guy thought it was broke when he was using the wrong power supply.
Charles
@unsignedbytesOct 13
Wow these are pretty cool
5
160
Replying to @tysn_dev
how much $$$
1
208
Charles retweeted
Ben🎃@DungeonNoirOct 19
Art by Dan Kendrick
7
114
2
1,214
6,957
1
11
1
131
1,061
1
1
43
Recipe is MiaAI-Lab/DeepSeek-v4.1-Flash-DGX-Sparks, ring transport is FujitsuPolycom/sparkring, tuned profile from knapcio's TP4 fork. In deepseek harness I get 70-100 tok/s single decode which is approaching frontier API speeds.
1
1
81
100 tok/s with TP4 deepseek flash v4.1 1017 No switch just ring mesh. Infinite unlocked opus 4.8 only cost me 20k, bonkers. Let me know if there's a way to go faster without losing any quality.
1
1
8
671
Replying to @justinn_builds
Recipe is MiaAI-Lab/DeepSeek-v4.1-Flash-DGX-Sparks, ring transport is FujitsuPolycom/sparkring, tuned profile from knapcio's TP4 fork. No switch In deepseek harness I get 70-100 tok/s
1
33
Deepseek v4.1 1017 on 4 dgx sparks is awesome btw
1
1
4
222
I hope someone out there is testing the shit out of deepseek right now looking for sleeper agent activity. This kinda stuff could kill local AI arxiv.org/abs/2401.05566

Sleeper Agents: Training Deceptive LLMs that Persist Through...

Humans are capable of strategically deceptive behavior: behaving helpfully in most situations, but then behaving very differently in order to pursue alternative objectives when given the...

arxiv.org
29
Replying to @0xAIlabs
slop slops sloppy
29
Replying to @curdmudgeon1 @AndrewCurran_
arxiv.org/abs/2401.05566

Sleeper Agents: Training Deceptive LLMs that Persist Through...

Humans are capable of strategically deceptive behavior: behaving helpfully in most situations, but then behaving very differently in order to pursue alternative objectives when given the...

arxiv.org
1
44
Load more