AI and infrastructure news
- HardwareArticle
GPU Programming: Architecture, Interconnects and LLMs
Published on September 17, 2025, the material collects GPU programming resources and includes a chapter on chips, interconnects, and implications for LLM workloads.
- Speech: TTS and ASRProject
Qwen3-Omni: multimodal models with speech
On September 22, 2025, Qwen introduced three models combining text, images, audio, and video, with speech input and output.
- HardwareArticle
Radxa Orion O6N combines 12 cores with an AI accelerator
Published on October 14, 2025, the article describes a Nano-ITX single-board computer with a CIX P1 SoC, up to 64 GB of LPDDR5, and a 30/45 TOPS AI accelerator.
- RAG and retrievalPost
Multi-hop RAG with a graph planned for each query
A post published on August 12, 2025 describes a method that plans dependencies among subproblems for each query and reports outperforming GraphRAG systems on three multi-hop question-answering datasets.
- RAG and retrievalPost
RARe adapts retrieval models with retrieved examples
Published on December 1, 2024, the paper presents RARe, a method that uses BM25-retrieved query-document pairs as examples to guide retrieval models.
- RAG and retrievalPost
REFRAG speeds up RAG with reusable embeddings
A paper published on September 15, 2025 presents REFRAG, a method that replaces most retrieved tokens with precomputed chunk embeddings. It reports 30× higher speed and support for 16× longer contexts without loss of accuracy.
- SecurityProject
Fast response to compromised crypto wallets
Published on October 16, 2024, the text describes an open-source Bash script for attempting to recover funds from compromised wallets before bots move them.
- Attention and model architecturePost
RLAD explores reasoning abstractions with reinforcement learning
Published on October 3, 2025, the paper presents RLAD, a framework using two-player reinforcement learning to help language models discover procedural hints in natural language.
- RAG and retrievalProject
A Python RLM implementation for long contexts
Published on October 17, 2025, this repository presents an open-source Recursive Language Models implementation that keeps long inputs in a Python environment for analysis.
- Attention and model architecturePost
RLMs: an approach to very long prompts
Published on October 15, 2025, the publication describes language models that interact recursively with long prompts through a REPL and reports evaluations on the OOLONG and BrowseComp-Plus benchmarks.
- Training and fine-tuningPaper
RLPT uses pretraining data for reinforcement learning
Published on September 25, 2025, the paper presents a way to apply reinforcement learning to pretraining data without new human annotations, and reports benchmark gains for Qwen3-4B-Base.
- LLM inferencePaper
Router-R1 explores reinforcement learning to route LLMs
Published on October 18, 2025, the text presents Router-R1, which alternates between reasoning and calls to multiple models and reports evaluations on seven question-answering benchmarks.
- C and C++Article
SDL3 can fill the browser window in Emscripten
A post dated October 2, 2025, shows how the SDL_EMSCRIPTEN_FILL_DOCUMENT=1 parameter makes an SDL3 window fill the browser window.
- RAG and retrievalPost
SimpleQA: measuring factual accuracy in AI responses
Published on October 30, 2024, the paper describes a benchmark of 4,000 factual questions and verified reference answers for evaluating models.
- Reinforcement learningProject
SkyRL tx v0.0.2 enables multi-LoRA training
On October 14, 2025, version 0.0.2 of the open SkyRL tx backend added support for training multi-LoRA models through the Tinker API.
- Training and fine-tuningPost
SWE-Swiss: a training recipe for software engineering
On August 12, 2025, researchers from Peking University, ByteDance Seed, and HKU presented a training recipe for software engineering tasks. The researchers report that SWE-Swiss-32B scored 60.2% on SWE-bench Verified.
- Attention and model architecturePost
SwiReasoning switches reasoning modes to reduce tokens
A publication dated October 12, 2025 describes a method that switches between latent and explicit reasoning and reports token-efficiency gains on Qwen3-8B models.
- LLM inferencePost
Inference throughput: SGLang and vLLM on A6000 GPUs
A post published on August 4, 2024 reports higher throughput with SGLang than with vLLM on four A6000 GPUs, but similar performance with smaller batches.
- AI agentsPost
ToolUniverse and research-tool integration
A post published on September 30, 2025 describes a shared interface for discovering and running more than 600 local and remote research tools that use MCP.
- LLM inferenceArticle
Torchchat brings local LLM inference to different devices
Published on July 30, 2024, the PyTorch post presents a library for running Llama 3, 3.1, and other LLMs locally on laptops, desktops, and mobile devices.
- Speech: TTS and ASRArticle
Streaming audio transcription: Modal's example
On August 6, 2025, Modal presented an example deployment of a streaming speech recognition service using Kyutai STT.
- Training and fine-tuningPost
Knowledge transfer between AI models
A repository published on July 29, 2024 presents a notebook for tuning a smaller model with domain knowledge generated by a larger model.
- Attention and model architectureArticle
Energy-based transformers refine token predictions
A paper published on September 27, 2025 reports that 44-million-parameter Energy-Based Transformers outperformed vanilla transformers on three of four benchmarks using RedPajama-Data-v2.
- HardwarePost
Magnetic transistor could combine control and memory
An article dated September 23, 2025, reports that MIT researchers developed a transistor using a magnetic semiconductor, with potential to improve electrical control and integrate memory.
- HardwarePost
Triton: GPU data access with block pointers
Published on September 23, 2025, the post covers tensor storage and the use of tl.make_block_ptr, strides, and offsets in Triton kernels.
- Speech: TTS and ASRPost
Streaming TTS points to multi-voice dialogue
A September 15, 2025 post describes a long-form streaming speech system that interleaves text and audio for contextual dialogue with multiple speakers.
- Speech: TTS and ASRPost
Local TTS with voice cloning: risks and implementation
On February 15, 2025, Zonos announced an expressive speech synthesis beta with voice cloning, open models, and a cross-platform local launcher.
- AI agentsPost
TUMIX refines answers with diverse agents at inference time
A post published on October 11, 2025 describes agents with different tools solving problems independently and refining answers over multiple rounds, without retraining the base model.
- Speech: TTS and ASRProject
Ultravox and real-time voice applications
The GitHub repository presents Ultravox as a fast multimodal LLM for real-time voice. Published on November 18, 2024, the project makes its source code available for examination.
- Attention and model architecturePost
A unified view of Attention-MoE and FFN-MoE
A paper published on September 24, 2025 proposes treating pre-mixed tokens and (V·Wo) as an FFN expert, unifying sparse attention and FFN approaches.
- Training and fine-tuningPost
UserRL trains agents with simulated users
In a post published on October 8, 2025, Salesforce presents UserRL, a framework for training agents with standardized environments and simulated users.
- RAG and retrievalPaper
VideoRAG expands information retrieval to video
On January 14, 2025, KAIST research presented a framework that retrieves videos relevant to queries and combines visual and textual information in multimodal RAG responses.
- Speech: TTS and ASRPost
VoiceCraft: speech editing and zero-shot synthesis
Published on August 25, 2024, the text describes VoiceCraft, a model that edits speech and generates voices from a few seconds of reference audio.
- HardwareArticle
Four-GPU workstation offers 192 GB of VRAM
On October 4, 2025, Maxsun and Abee announced a workstation with four Intel Arc Pro B60 Dual 48G GPUs, totaling 192 GB of VRAM, for AI workloads.
- Speech: TTS and ASRPost
XEUS: multilingual speech encoder with code and checkpoints
On July 1, 2024, WavLab announced XEUS, a speech encoder trained on more than 1 million hours of audio and covering over 4,000 languages. Code, checkpoints, and data were made available.
- HardwareArticle
XIAOML Kit brings sensors and machine learning systems closer
Published on August 7, 2025, the post presents a kit with a camera, microphone, microSD storage, IMU, and display for hands-on embedded machine learning lessons.
- OCR and document AIPost
Zerox: structured document extraction
A post published on August 10, 2025 describes an open-source OCR and document extraction project that uses schemas to retrieve specific fields.
- AI agentsPaper
163 procedures for scientific research agents
The article presents an open library of 163 procedures for research agents across 16 fields, including genomics, cheminformatics, medical imaging and study design.
- AI agentsPaper
Ablations show how to evaluate coding-agent harnesses
A paper published on September 26, 2026 reports 176 configurations tested on two benchmarks and describes differing effects from context summarization and explicit planning.
- AI agentsPost
AI R&D agents need to revisit training strategy
A study published on August 22, 2026, analyzed 1,338 post-training trajectories. The overall strategy changed in 74 of 3,557 adjacent experiments.
arxiv.org
Let's assess your project
I reply with scope and a proposal.