Stephen Roller

.

Cluster magicians and GPU whisperers, come join us! We’re looking for supercomputing engineers to build the infrastructure behind real-time interactive models, Tinker, and large-scale training: scheduling, storage, networking, reliability, and distributed systems at scale. Hiring in NYC and SF job-boards.greenhouse.io/thinkingmachin…

@willdepue `caffeinate -s`

Tired of holding your laptop half open to keep your agents running? Introducing AgentPlug: A USB-C dummy plug that keeps your Mac in clamshell mode by pretending to be an external display! No commands, no security worries (just pull it out to stop!), no hassle.

People talk, listen, watch, think, and collaborate at the same time, in real time. We've designed an AI that works with people the same way. We share our approach, early results, and a quick look at our model in action. thinkingmachines.ai/blog/interacti…

This is the kind of shit Godspeed You! Black Emperor would use as an album cover with a title like "And the Holy Flame Blinded us for 333000 Years Pt. II"

@BeidiChen imho hash layers are the logistic regression of routers — a surprisingly strong baseline that you absolutely must beat

@stephenroller That was a compelling debate on hash-layer MoE vs. trainable routers—especially valuable with so many experts weighing in now ~

Didn't have hash layers on my bingo card

DeepSeek-V4 uses our Hash routing approach developed back in 2021 -- see screenshot of their tech report! (Looks like a great model, congrats!) Bonus note: our same blogpost (& paper) back in 2021 also introduced 'looped transformers', but we called that staircase & ladder (see screenshot): parl.ai/projects/param… huggingface.co/deepseek-ai/De…

me: Make me the most AI slop image that ever AI slopped. The pinnacle of slop. A seminal work on AI slop. ChatGPT Images 2.0:

Long context windows are now available for select models on Tinker! - 128k tokens for Kimi K2.5 and GPT-OSS-120B - 256k for Nemotron 3 Super 120B and Qwen3.5 397B. For more details and pricing, see our full model lineup: tinker-docs.thinkingmachines.ai/tinker/models/

@bcherny cruel april fools day joke

Today we're excited to announce NO_FLICKER mode for Claude Code in the terminal It uses an experimental new renderer that we're excited about. The renderer is early and has tradeoffs, but already we've found that most internal users prefer it over the old renderer. It also supports mouse events (yes, in a terminal). Try it: CLAUDE_CODE_NO_FLICKER=1 claude

I heard ASL-5 is when the Claude code TUI stops flickering in tmux

this is what actual national suicide looks like btw

I couldn't agree more. This is exactly why we founded @datologyai

Grateful to Jensen and @nvidia team for their support. Together, we’re working to deploy at least 1GW of Vera Rubin systems, bringing adaptable collaborative AI to everyone. thinkingmachines.ai/nvidia-partner…

times are hard. had to teach an AI researcher how to use kubernetes today

things are gonna get weird. you must get commensurately weird.

@bilaltwovec i like to imagine myself carrying a node out of the burning dc when i do squats

@stephenroller damn i can only bench a single gb200..mo(r)e weight training needed

A fully-loaded nvl72 rack weighs 3,300lbs. Meaning the per-gpu weight is ~45lb, or one standard olympic plate.