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/YANG SONGLIN
// PERSON

Yang Songlin

ROLE RESEARCHER/ARCHITECTAT MINIMAXMENTIONS 1LAST SEEN APRIL 30, 2026
// BIO

Songlin Yang is a researcher and architect at MiniMax who specializes in efficient attention mechanisms for large language models. He is the creator of the Flash Linear Attention library and has been involved in the architectural development of MiniMax models including the M1 and M2. His work focuses on hybrid attention designs that balance the computational efficiency of linear attention with the performance of full softmax attention.

// RECENT MENTIONS
// SIGNALS
1 SIGNAL
01
mention·晚点聊 LateTalk·APRIL 30, 2026

Yang Songlin from MiniMax — I had previously heard that they internally favor absorption over linear attention. I personally think absorption is a more controllable engineering solution.

Source

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

Yang Songlin · Researcher/Architect at MiniMax — 1 mention on Teahose