Building sovereign AI infrastructure. 20-node cluster · 6× RTX PRO 6000 Blackwell (576GB VRAM) · 2TB ECC RAM · PB-scale Ceph · 400GbE fabric
Santa Clara, London
Joined November 2013
- Tweets601
- Following3.1K
- Followers3.1K
- Likes5.8K
Dataphile retweeted
Fable 5.2 wrote its own specifications, selected the components, routed the four-layer circuit boards, designed enclosures for 3D printing, and programmed the software.
Each board was ready for manufacture in under a day. All three devices work without any modifications.
Dataphile retweeted
During pre-release testing, Claude Fable 5.2 and Claude Mythos 5.2 worked on ambitious projects in many different fields. Here are a few.
From a one-paragraph prompt, Fable 5.2 designed three working electronic devices, including this home voice assistant hub.
Dataphile retweeted
Introducing Claude Fable 5.2 and Claude Mythos 5.2, designed to tackle the most daunting technical challenges.
They’re capable collaborators on ambitious projects in many fields, and show improvements in areas like coding, reasoning, and vision.
Dataphile retweeted
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.…
Dataphile retweeted
5/5
A huge thanks to @TheZachMueller for the multiple bitter lessons along the way
==========
Benchmarking Tool Sources:
- @YourLocalAILab llm-inference-bench
github.com/local-inference-l…
- @NVIDIAAI Dynamo AIPerf
github.com/ai-dynamo/aiperf
- @SemiAnalysis_ AgentX + their Claude Code traces
github.com/SemiAnalysisAI/ag…
huggingface.co/datasets/semi…
- local-inference-lab/DeepSeek-V4.1-Flash-1017-lossless-CSF checkpoint repo
huggingface.co/local-inferen…
- Launch recipes, AIPerf, and AgentX data:
github.com/yoniq-org/compute
==========
llm-inference-bench data can be found at @LottoLabs's LocalMaxxing links below:
- deepseek-ai/DeepSeek-V4.1-Flash-1017
localmaxxing.com/en/inferenc…
- deepseek-ai/DeepSeek-V4.1-Flash-1017 (DSpark 7)
localmaxxing.com/en/inferenc…
- nvidia/DeepSeek-V4.1-Flash-1017-NVFP4
localmaxxing.com/en/inferenc…
- nvidia/DeepSeek-V4.1-Flash-1017-NVFP4 (DSpark 7)
localmaxxing.com/en/inferenc…
- local-inference-lab/DeepSeek-V4.1-Flash-1017-lossless-CSF (DSpark 7)
localmaxxing.com/en/inferenc…
Dataphile retweeted
The #NVIDIAGTC Berlin Pregame has just begun. 🎬
Industry leaders are on the live stage setting the conversation before Jensen Huang’s keynote.
Tune in now → nvda.ws/4yvMNSU
Dataphile retweeted
New on our Engineering blog: Creating in-house source control.
In May, we merged around 2,500 code changes per day. In September, we merged more than 30,000 per day. As a result, we became bottlenecked by developer infrastructure built for human (rather than agent) scale.
Here, we share how we built a new version control system—and fully migrated our monorepo from git—in five months: anthropic.com/engineering/re…
Dataphile retweeted
Updated the Yoniq Compute repo with inference recipes validated from 1x up to 8x RTX Pro 6000 GPUs
This repo provides tools to stage a Linux node for ML workloads, including scripts to build & serve inference recipes across SGLang / vLLM, with isolated env management
github.com/yoniq-org/compute
Dataphile retweeted
Community Notice: Please Refrain from Posting Quiz-Style Content
Dear members,
We’ve noticed a growing trend of accounts posting math and physics “quiz” questions in the community. While we appreciate enthusiasm for the subject, this community is dedicated to meaningful discussions about important ideas in mathematics and physics — not for promoting personal accounts through quiz-style content.
Going forward, posts consisting solely of quiz questions or “challenge problems” intended to drive engagement or followers will be removed, and accounts repeatedly posting such content will be banned.
We encourage thoughtful, in-depth discussion, questions that lead to real exploration, and contributions that enrich the collective understanding of these fields.
Thanks for helping keep this community focused and high-quality.
- The Mods