@RitOnchain: Jane Street pays $750K/year for quants who master matrix calculations holistically that can be used to get alpha from s…
Summary
A free 57-minute resource by MIT's Applied Math team covers matrix calculations and automatic differentiation for quants and optimization, highlighting Jane Street's high compensation for such skills.
Similar Articles
@Phoenixyin13: If the full score is 10, I would honestly give this MIT paper's SMT idea and writing an 8. The paper proposes Supervised Memory Training, using Transformer as a super teacher to first distill in parallel the most important things to remember at each moment…
This paper proposes Supervised Memory Training (SMT), which uses Transformer as a super teacher to distill memory states in parallel, then trains RNN with one-step supervised learning, achieving fully parallel training and reducing gradient path from O(T) to O(1), significantly improving long-range dependency learning.
I didn't know it was possible to compile llamacpp to run cuda + vulkan at the same time..
The author discovered that compiling llama.cpp with both CUDA and Vulkan backends simultaneously is possible, yielding a ~10% improvement in tokens/sec for decoding. They plan to run further benchmarks to assess the benefits.
MIT’s Initiative for New Manufacturing builds momentum
MIT's Initiative for New Manufacturing (INM) celebrated its first anniversary with Manufacturing Week, attracting over 800 participants to discuss AI on factory floors, workforce solutions, and startup innovation, while launching programs to commercialize manufacturing technologies.
Making ast.walk 220x Faster
The Reflex team optimized Python's ast.walk by 220x for their AI code generation linter by removing generator overhead, inlining functions, and implementing a Rust binding.
@Gorden_Sun: https://x.com/Gorden_Sun/status/2066919099016630286
A long-term study involving 26,000 Chinese middle and high school students found that after students independently used AI, homework performance improved by 18%, but closed-book exam scores dropped by 20% within six months. Zhongkao and Gaokao scores dropped by 24% and 18% respectively, and 81% of students used AI to complete their homework.