Engineering ·
Qwen 3.5 and Nemotron 3 use different hybrid architectures
Published on July 8, 2026, the post compares two hybrid architectural approaches and relates them to the convergent evolution of linear attention and state-space models.
Radar
Daily bookmarks: relevant news, facts and discoveries.
Articles 1219
Engineering ·
Ridgeline: a roofline profiler for LLM inference
Ridgeline is a small profiler built with PyTorch to help engineers analyze LLM inference workloads using a roofline model.
Engineering ·
SearchEyes trains search agents in simulated worlds
Published on July 8, 2026, the summary presents SearchEyes, an approach that uses a typed knowledge graph as the basis for a simulated search world to train multimodal agents.
Engineering ·
BrowserBC turns browsing traces into skills for agents
The approach converts human browsing traces into natural-language skill cards that agents can retrieve and combine to navigate unfamiliar websites.
Engineering ·
Relevance-Based Embeddings for Candidate Retrieval
A paper describes query and item embeddings for efficiently retrieving candidates when relevance computation is expensive. The reported approach uses relevance to selected support items or queries.
Engineering ·
Harness engineering and AI self-improvement
Published on July 7, 2026, Lilian Weng’s article examines recursive self-improvement and presents designing and optimizing a model harness as a practical starting point.
Engineering ·
HiLS learns chunk selection for sparse attention
Published on July 7, 2026, the material presents Hierarchical Landmark Sparse (HiLS) Attention, a method that learns chunk selection end to end using the language-modeling loss.
Engineering ·
JetSpec speeds up speculative decoding with parallel trees
A post published on July 7, 2026 describes JetSpec, which uses a single forward pass with causal conditioning to create coherent draft trees. It reports speedups of up to 9.64× on MATH-500 and 4.58× in open-ended conversations.
Engineering ·
Live stream presents new chapters on CUDA programming
Published on July 7, 2026, the notice describes a live stream about new chapters of Programming Massively Parallel Processors and an overview of the material.
Engineering ·
Study traces training data linked to LLM circuits
Published on July 7, 2026, the work presents a method for linking training samples to interpretable attention heads and reports effects of interventions on those circuits.
Engineering ·
Rejection sampling for on-policy LLM alignment
Published on July 7, 2026, the summary presents a proposal to convert off-policy tokens into tokens treated as on-policy through rejection sampling, addressing variance in LLM alignment.
Engineering ·
Privileged-information distillation may affect long reasoning
A study evaluates self-distillation with privileged information in reasoning models and reports degradation in long trajectories across five models tested on AIME.
Engineering ·
A taxonomy of agentic recommender systems
A survey published on July 7, 2026 proposes classifying recommender systems with LLM-based agents by their level of autonomy.
Engineering ·
Autodata generates training data with agent loops
Published on July 6, 2026, this summary presents Autodata, a method that uses agents to create training and evaluation data by selecting questions that distinguish weak and strong solvers.
Engineering ·
AutoMem trains agents to manage their own memory
A post about a paper describes agents that write, search and organize notes, learning when memory is useful. According to the post, this improved an open 32B model's performance by 2x to 4x in long games.
Engineering ·
Floating-point arithmetic from scratch
Published on July 6, 2026, the article presents building floating-point arithmetic from scratch as a high-difficulty project.
Engineering ·
An Introduction to Compilers and Language Design
A 2021 book by Douglas Thain introduces compilers and language design for engineers interested in building these technologies.
Engineering ·
Functional Attention proposes linear-cost attention
Published on July 6, 2026, the paper summary describes learned bases and a compact matrix for representing attention, reports linear cost in the number of points, and cites results in three evaluation areas.
Engineering ·
NVIDIA CUDA C++ guide updated in June 2026
A bookmark published on July 6, 2026 reports that NVIDIA's CUDA C++ Programming Guide was updated in June 2026 and describes it as a current reference for engineers working with CUDA C++.
Engineering ·
HOLA combines recurrent state with a limited KV cache
In a post published July 6, 2026, HOLA combines recurrent state with a small KV cache for exact memory. The post reports a 16.1% reduction in perplexity and robust retrieval at 32k tokens.