Watch the full livestream replay to hear the conversations shaping the AI stack today and where it’s headed next:
RESEARCH
-
Quantization Impact on Model Quality and Expert Reduction
By
–
How confident are you with respect to the output quality given the 2-bit quantization and reducing experts from 10 to 4? Did you have a mechanism for measuring that?
-
Running 397B MoE Model on M3 Mac with Efficient Weight Streaming
By
–
Dan says he's got Qwen 3.5 397B-A17B – a 209GB on disk MoE model – running on an M3 Mac at ~5.7 tokens per second using only 5.5 GB of active memory (!) by quantizing and then streaming weights from SSD (at ~17GB/s), since MoE models only use a small subset of their weights for
-
Industry Challenges in Solving AI Development Problems
By
–
You are asking for things the industry as a whole has yet to figure out.
-
Recruiting for GenMedia Track – Lead Engineers and Researchers Wanted
By
–
im recruiting for our genmedia track! who was the lead eng/researchers on this!
-
MiniMax-M2.7 model hits SWE-Pro SOTA at 56.22%
By
–
MiniMax-M2.7 is here. → matching Sonnet 4.6 as an agent
→ recovering live incidents in just 3 minutes
→ hitting SOTA in SWE-Pro (56.22%) It even edits your Office files -
Failed AI pioneers’ paradoxical success
By
–
Geoff Hinton set out to figure out how the brain works and failed.
Andrew Ng set out to build a complete robot and failed.
Demis Hassabis set out to achieve AGI using deep RL and failed.
Yet they all succeeded. -
Opinion on Sam Altman’s CEO Role at OpenAI
By
–
19% more deaths from Covid in the US than counted thru 2021 https://
science.org/doi/10.1126/sc
iadv.aef5697
… -
Jensen Huang and the Prescience of Deep Learning in 2015
By
–
The signature is alluding to NVIDIA GTC 2015, where Jensen excitedly told an audience of, at the time, mostly gamers and scientific computing professionals that Deep Learning is The Next Big Thing, citing among other examples my PhD thesis (one of the first image captioning systems that coupled image recognition ConvNet to an autoregressive RNN language model, trained end to end). This was back when most people were still unaware and somewhat skeptical but of course – Jensen was 1000% correct, highly prescient and locked in very early. [Translated from EN to English]
