Teahose.
SIGN IN
NEW HERE — WHAT TEAHOSE DOES
We read the entire AI & tech firehose — so you don't have to.
PODPodcastsAll-In, No Priors, Acquired…
NEWNewslettersStratechery, Newcomer…
PAPPapersPhysical AI research
PHProduct Huntdaily launches
VCInvestor ScoutSequoia, a16z, Benchmark…
CLAUDE DISTILLS →
7 reads, 30 sec each — free, 6 AM ET.
+ a live graph of the companies, people & themes underneath.
HOME/PEOPLE/SONGLIN YANG
// PERSON

Songlin Yang

ROLE CORE AUTHOR OF DELTANETMENTIONS 2LAST SEEN AUGUST 3, 2026
// BIO

Songlin Yang is a Member of Technical Staff at Thinking Machines Lab, focused on machine learning and language model architectures. She earned her PhD from MIT CSAIL, where she was advised by Professor Yoon Kim and specialized in hardware-aware algorithms for efficient sequence modeling. She is best known for developing Gated Linear Attention transformers, creating the Flash Linear Attention library for hardware-efficient attention kernels, and co-developing DeltaNet, which parallelizes linear transformer training with the delta rule across sequence lengths.

// RECENT MENTIONS
// SIGNALS
2 SIGNALS
01
mention·晚点聊 LateTalk·AUGUST 3, 2026

KDA — from the Kimi Mini Linear paper to the 2.8T mainstream model — took less than one year. So if your genius idea is right, you don't need to wait for a new paradigm. Someone will come along and adopt it.

Source
02
mention·张小珺Jùn|商业访谈录·JULY 28, 2026

Last year, Mandos' Jichao — that's Peak — also wrote a blog about how to design an infra-friendly harness framework, with the most important part being fully utilizing prefix caching.

Source

AI-extracted from podcast / newsletter / paper summaries. May contain errors.