October 7, 2026 · fergusfinn.com
From nvcc to warps: how a CUDA kernel runs
Published on July 8, 2026, the article follows a vector-addition kernel from compilation with nvcc to execution by GPU warps.
AI and infrastructure news
Each item sums up the news and what a company can do.
Recent news
October 7, 2026 · x.com
From REINFORCE to GRPO: policy gradient methods for LLMs
An article published on September 24, 2026, presents a progression of policy gradient methods used to train LLMs, from vanilla policy gradient and REINFORCE to PPO, GRPO, and GRPO variants.
October 7, 2026 · arxiv.org
Dockerless proposes verifying patches without execution
Published on June 30, 2026, the paper describes a verifier that assesses coding-agent patches without running them. The proposal targets selecting SFT trajectories and providing rewards for RL, as an alternative to setting up repository-specific environments.
October 7, 2026 · arxiv.org
DreamSmooth smooths sparse rewards in RL
Presented in an ICLR paper published on June 28, 2026, DreamSmooth applies temporal reward smoothing in model-based RL and was evaluated on RoboDesk and Shadow Hand.
October 7, 2026 · arxiv.org
Dual-Cache: sharing context across LLMs
A paper published on September 26, 2026 presents a latent protocol for transferring information across heterogeneous models by translating KV caches and adding joint mixing between layers.
October 7, 2026 · x.com
Dust explores gradient-free transformer training
Published on October 5, 2026, the paper presents Dust, a zeroth-order method using activation perturbations, and reports scaling experiments and comparisons with backpropagation.
October 7, 2026 · arxiv.org
EfficientRollout speeds up RL rollouts for LLMs
A paper dated July 5, 2026 describes system-adaptive self-speculative decoding and reports up to 19.6% faster rollout generation.
October 7, 2026 · arxiv.org
Relevance embeddings for candidate retrieval
A Yandex paper published on July 7, 2026 describes query and item embeddings to retrieve candidates efficiently when relevance scoring is costly.
October 7, 2026 · huggingface.co
Multimodal embeddings expand search across visual content
In a post published on August 25, 2026, Tencent describes a model that represents text, images, video, and visual documents in a shared space and claims top results on two benchmarks.
October 7, 2026 · arxiv.org
EnvACE rehearses the world to train agents
Published on August 9, 2026, the article describes a reinforcement learning method that replaces interactions with external environments with internal rehearsals during agent training.
October 7, 2026 · arxiv.org
Learning-rate scaling in LLMs may be nonlinear
An article published on June 30, 2026, evaluates learning rates in GPT-2-style models ranging from 22 million to 707 million parameters, trained on 5 billion to 100 billion tokens.
October 7, 2026 · x.com
ESP32 blocks 537,000 domains with about 50 KB of RAM
A post published on July 20, 2026 describes ESP32 firmware that stores its blocklist in flash and takes about 10 milliseconds to block requests.
October 7, 2026 · x.com
Weight spectra may signal memorization in LLMs
A preliminary analysis published on September 24, 2026, found that the attention matrix K exponent α tracked memorization of random labels during LLM training.
October 7, 2026 · github.com
Recurrent state could extend Transformer memory
The description of the Recurrent Looped Transformer, dated September 13, 2026, presents a recurrent decoder that preserves state between tokens to extend the temporal information path.
October 7, 2026 · usenix.org
Study reports code overlap in network filtering systems
A USENIX Security paper reports code overlap between leaked Geedge Networks code and China's Great Firewall, including DNS and RST packet injection.
October 7, 2026 · arxiv.org
EvoRec proposes agents to optimize recommendations
Published on June 30, 2026, the paper presents a multi-agent framework intended to reduce manual optimization of recommendation systems by retaining methods across iterations and exploring ideas beyond a predefined search space.
October 7, 2026 · arxiv.org
ExploitGym evaluates AI agents on vulnerability exploitation
The ExploitGym research proposes evaluating whether AI agents can turn vulnerabilities into impacts such as unauthorized file access or code execution.
October 7, 2026 · github.com
Silent failures in LLM serving: a practical diagnosis
Published on July 28, 2026, the GitHub repository catalogs LLM serving failures, their symptoms, mechanisms, checks, and fixes.
October 7, 2026 · x.com
Attention factorization and gradient flow: implications for models
Published on September 30, 2026, the publication discusses how factoring the attention matrix as S = WWᵀ may change gradient flow and remove an information-exponent bottleneck in optimization.
October 7, 2026 · arxiv.org
FIBER proposes separating GPU register ownership from execution
Published on August 21, 2026, the article presents a GPU architecture that decouples register ownership from parallel execution instances.
October 7, 2026 · arxiv.org
Context filtering for attention on constrained hardware
A SAS article published on September 16, 2026 describes an end-to-end trained selector inside Softmax to filter context in attention mechanisms aimed at hardware with limited computing resources.
October 7, 2026 · github.com
FireRedAudio combines five speech tasks in one model
Published on August 27, 2026, FireRedAudio combines speech recognition and understanding, speech generation, and speech editing, using distinct representations for audio comprehension and generation.
October 7, 2026 · arxiv.org
FlashAttention-3: optimizing attention on Hopper GPUs
Published on September 24, 2026, the paper presents asynchronous execution and low-precision techniques to accelerate attention on Hopper GPUs, focusing on memory traffic and hardware utilization.
October 7, 2026 · arxiv.org
FlowTracer tracks influential tokens in LLM reasoning
Published on July 4, 2026, the article presents a method that uses attention to trace information propagation and direct credit assignment in reinforcement learning.
October 7, 2026 · x.com
Recurrent flows refine reasoning on discrete tasks
A post published on September 3, 2026 describes models combining continuous flows and recurrent refinement to solve structured problems such as Sudoku.
October 7, 2026 · x.com
Fork Enables PCIe P2P on GeForce
A post published on August 27, 2026 reports a graphics-driver change enabling peer-to-peer PCIe transfers on GeForce cards while optimizing GPU serving.
October 7, 2026 · github.com
FrontierSmith publishes 200 open-ended RL problems
On September 28, 2026, FrontierSmith published a repository of 200 problems for studying reinforcement learning on challenging, open-ended tasks.
October 7, 2026 · youtube.com
Using ftrace to observe Linux kernel behavior
A 2018 session by Steven Rostedt introduces Linux kernel tracing with ftrace and covers virtual memory, page tables, and how programs interact with the kernel.
October 7, 2026 · x.com
FuncAttn brings functional attention to continuous data
A June 25, 2026 post describes a method using structured linear operators for attention between function spaces, with results in PDE solving, 3D segmentation, and regression.
October 7, 2026 · x.com
Functional Attention proposes linear-cost attention
Published on July 6, 2026, the paper describes representing query and key/value fields with learned bases and a compact matrix, reporting linear cost in the number of points.
October 7, 2026 · x.com
FUTEX_PRIVATE_FLAG and the cost of thread locks
A post dated June 28, 2026 explains how FUTEX_PRIVATE_FLAG may reduce lock latency for threads sharing memory within the same process.
October 7, 2026 · arxiv.org
Gated Linear Attention: linear attention with gates
Published on September 20, 2026, the paper presents a transformer formulation with linear attention and data-dependent gates, also expressible as an RNN with matrix-valued hidden states.
October 7, 2026 · x.com
GEPA adapts a model with prompts and labeled examples
A demonstration published on September 19, 2026 describes using prompts and labeled examples to adapt a model to specific decision criteria.
October 7, 2026 · arxiv.org
Git as shared memory for research agents
A post published on September 17, 2026 describes agents recording verifiable contributions in a searchable Git history to share results and reduce repeated experiments.
October 7, 2026 · sploitfun.wordpress.com
glibc malloc: how to observe memory use
Published on June 25, 2026, the article covers how glibc malloc obtains memory from the kernel and manages the heap. This pauta proposes turning that knowledge into observability for memory-intensive services.
October 7, 2026 · x.com
GLM-5.3-Flash: an architecture overview
Published on August 26, 2026, the overview describes a hybrid architecture with 3:1 attention, a smaller sparse MoE backbone, a four-stream mHC residual path, and a native vision encoder.
October 7, 2026 · arxiv.org
GOAT proposes trainable priors for attention
Published on October 3, 2026, the paper frames standard attention as an entropic optimal transport problem with an implicit uniform prior and presents GOAT, which uses a continuous, trainable prior.
October 7, 2026 · x.com
Two-core GPU on a $30 FPGA: what to assess
Published on July 9, 2026, the article describes a two-core GPU implemented on a $30 FPGA, running AI model inference at 81 MHz.
October 7, 2026 · arxiv.org
GPU-Tile-Sim proposes tile-based GPU simulation for LLM co-design
Published on July 14, 2026, the paper proposes a tile-based GPU simulation framework to support hardware and software co-design for LLMs.
October 7, 2026 · arxiv.org
BF16 gradients rise late in training
A paper reports a 1,000-fold increase in gradient norm after 25 billion tokens in a transformer trained with FlashAttention-3 in BF16.
Let's assess your project
I reply with scope and a proposal.