btw Zai IPO'ed in Jan at HK$120 a share. when I first met @louszbd nobody really knew anyone using GLM's. now they have beat deepseek with the world's undisputed top open model and in some respects (see @ml_angelopoulos
) say top model period, and are returning to SF
MACHINE LEARNING
-
Zai IPOs at HK$120, GLM beats DeepSeek as world’s top open model
By
–
-
hf-claude works well with GLM 5.2, install via hf extensions
By
–
hf-claude works well with glm 5.2 hf extensions install hf-claude
-
Benchtalks Ep. 3: Continual Learning Bench with Pgasawa and Vincentsunnchen
By
–
Benchtalks Ep. 3 with @pgasawa (Continual Learning Bench): coming soon with @vincentsunnchen
-
Why Model Routers Fail for Coding: Information Deficit
By
–
What is actually limiting model routers for coding tasks? Most routers treat picking a model as a static, one-off classification. This paper identifies the real bottleneck as information deficit. Simply augmenting a vanilla LLM router with task-dimension-level performance
-
MBench tests memory in video world models
By
–
Can video world models truly remember what they see over time? Researchers from Tsinghua University, Tencent, and Peking University present MBench, a new benchmark that tests memory in video AI. Instead of just checking if videos look good, MBench measures whether models keep
-
Tip: Use Codex to update the global agents file
By
–
Codex tip: ask Codex to review your old PRs/sessions and update your global agents md file with your development workflow details: branch naming conventions, commit messages, attribution, test plan, and more
-
Not surprised about Grok, other models more reasonable now
By
–
Not surprised about Grok, but glad to see even other models to be much more reasonable now.
-
Street View Grounding Now Available for Google’s Offline Video Models
By
–
street view grounding now available for google’s offline video models. can’t wait till you can do “spatial RAG” to load in the right panoramas to reference – suddenly large scale real world locations become movie sets! https://t.co/AR3gLfj0p6
— Bilawal Sidhu (@bilawalsidhu) 23 juin 2026street view grounding now available for google’s offline video models. can’t wait till you can do “spatial RAG” to load in the right panoramas to reference – suddenly large scale real world locations become movie sets!
-
NVIDIA Nemotron-3.5-ASR: open-source streaming speech recognition model
By
–
🚨 @NVIDIA just quietly dropped an incredibly impressive speech recognition model that completely changes the math for local voice pipelines.
— Charly Wargnier (@DataChaz) 23 juin 2026
Nemotron-3.5-ASR is a 0.6B parameter, open-source model built specifically for real-time streaming.
What makes it so good:
→ 40+…@NVIDIA just quietly dropped an incredibly impressive speech recognition model that completely changes the math for local voice pipelines. Nemotron-3.5-ASR is a 0.6B parameter, open-source model built specifically for real-time streaming. What makes it so good:
→ 40+