Knowledge catalogue
arXiv:2608.21208v1 Announce Type: cross Abstract: This paper investigates cross-agent specification portability using Oracle-to-PostgreSQL migration as a controlled software transformation task. The s
arXiv:2608.20477v1 Announce Type: new Abstract: Neural operators have emerged as efficient surrogates for time-dependent physical systems governed by partial differential equations (PDEs), but their f
arXiv:2608.20786v1 Announce Type: new Abstract: Multi-agent pipelines that author formal documents must both read a requester's forms and write against them. We report a deployed tender-response syste
arXiv:2608.20944v1 Announce Type: new Abstract: Multimodal object detection in remote sensing faces challenges due to semantic heterogeneity and modality-specific noise interference. To this end, we p
arXiv:2608.21035v1 Announce Type: new Abstract: Long-horizon manipulation tasks are often only partially ordered. For example, when assembling an electronic device, the battery and circuit board may b
arXiv:2608.20685v1 Announce Type: cross Abstract: Retrieval-augmented generation (RAG) has no model of time: when a fact changes across a coding session - a function is renamed, an endpoint moves, a d
This is one of the most useful writeups I have seen on keeping an LLM judge effective in production. (bookmark it) Netflix runs judges over hundreds of thousands of show-level recommendation explanati
arXiv:2608.20809v1 Announce Type: cross Abstract: Breast ultrasound diagnosis relies on clinically meaningful semantic concepts, yet most deep learning methods adopt end-to-end image-to-label paradigm
arXiv:2608.20777v1 Announce Type: new Abstract: As scientific literature grows and papers increasingly under-report limitations, multi-agent LLMs offer a promising approach to systematically uncover t
arXiv:2608.21008v1 Announce Type: new Abstract: Mobile AR frameworks attach a metric pose prior to every casual phone capture, and turning it into reconstruction-grade poses cheaply on CPU is the step
arXiv:2608.20627v1 Announce Type: cross Abstract: Agentic retrieval-augmented generation (RAG) interleaves retrieval, reasoning, and answer generation across multiple hops. A retrieval error at hop 1
arXiv:2608.20516v1 Announce Type: new Abstract: Joint-embedding predictive architectures are selected almost universally by linear probing and effective rank. We report a case where both read healthil
arXiv:2608.20400v1 Announce Type: new Abstract: Agentic memory under a fixed budget involves two stages: retention and retrieval. Existing retrieval-centered paradigms implicitly assume necessary evid
arXiv:2608.20558v1 Announce Type: new Abstract: Digital cameras embed device-specific artifacts into every acquired image through demosaicing, in-camera post-processing, and lossy compression. These t
A few weeks ago I shared Agent Quest, my open-source experiment that turns Claude Code and Codex sessions into heroes living inside a small 2D world. The original idea was mainly about making it easie
cuda : add POOL_1D support (#27573) cuda : add POOL_1D support fix: add missing trailing newline for editorconfig compliance Website: https://llama.app Attestations: https://github.com/ggml-org/llama.
vendor : update subprocess.h (#27409) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42402532 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (a
common : skip device_info loop if it's not going to be printed (#26692) The device_info loop iterates over the discovered devices and gets the available and total memory counts. With the CUDA backend
server : add LLAMA_SERVER_SLOTS_N_DIFF (#27600) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42423433 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple
test: move tools/parser to tests (#27548) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42442638 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silico
I guess children needlessly dying of a disease that has long been fully preventable is a small price to pay for Byron Donalds to become governor of Florida. Brennan: 'Nearly all of Florida is below th
What I ran: 8x B300 on Modal, 56.79 per hour, vLLM, tensor parallel 8, native MXFP4 Cold boot ~27 min (1.56 TB load, JIT, 51 CUDA graph captures) TTFT 0.92 to 1.02 s, decode 92 tok/s steady, 83 tok/s
If you maintain an AGENTS.md or a CLAUDE.md, this one is worth your time. (bookmark it) Researchers traced 94K development events across 557 agentic coding sessions, plus 690K file-level change record
Hi all Looking at a PowerColor R9700 for Qwen3.8-27B, Q4, llama.cpp/Vulkan. AMD's own blog quotes 51.8 tok/s but doesn't say what context length that's at, or whether MTP=2 was holding up. Separately
Hey everyone! Last year, Mozilla released Orbit, an AI-powered browser summarizer hosted on a GCP server. After people started digging into the extension, they discovered things like backend endpoints
Scientific terms should have precision. If we use the terms VLM, VLA, WAM in an indiscriminate fashion, as is becoming common in robotics, we are not helping clarity in communication. Let's keep the h
the difference in productivity between 0-1 mode vs. working within a big team at a large co seems universally true even pre-chatgpt. it's really easy to be 'productive' when (a) you're the only person
The latest RAG trend for the current agent harnesses (Codex, Cowork) is to do two passes of document processing to solve a knowledge work task over a data room of documents: 1️⃣ A fast and light pass,
The research on financial advice https://x.com/emollick/status/2085174123743842448?s=20 This paper by researchers from MIT and Stanford finds that most people would be financially better off if they f
They treat DC as a plaything for incompetent leaders. They treat their job as a grift and a joke. Vote them ALL out Mike Johnson says he wants to make the downtown DC Indycar race 'an annual event ...
https://huggingface.co/jojohai/Qwen3.8-27B-MTP-graft Tested on Vulkan, the grafting saves RAM compared to using an external file. What I don't guarantee however is the quality of answers. The model is
It performs at full capacity without any issues anywhere, but only the Ollama DeepSeek seems particularly trashy. Why on earth has the Ollama DeepSeek unusable trash that goes into over reasoning and
If anyone still remembers GLM-4.5-Air from last year, you can now get a nice speedup by enabling MTP in llama.cpp. It is a 106B MoE with only 12B active parameters, which makes it interesting for mach
A couple of days ago, I did a quick explainer on Claude’s new watermarking process and implementation. Since it’s such a popular topic and sparked such a lively discussion, I thought it might be inter
mtmd: support webp via ffmpeg (#27520) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42289035 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (
sycl : add Q2_K reordered MMVQ and ESIMD kernels (again) (#27490) Revert 'Revert 'sycl : add Q2_K reordered MMVQ and ESIMD kernels (#26336)' (#…' This reverts commit 7a0e42f. add gate params Website:
common : fix draft-mtp with embeddings (#26352, #27299) (#27400) common: fix draft-mtp with embeddings (#26352) --whitespace Co-authored-by: Sigbjørn Skjæret sigbjorn.skjaeret@huggingface.co Website:
ggml: optimize concat op by replacing per-element memcpy with row-level memcpy (#24575) ggml: optimize concat op by replacing per-element memcpy with row-level memcpy ggml: fix concat offsets for row-
mtmd: support dots3-note vision+audio (#27524) text: conversion init impl mtmd: conversion impl mtmd cpp Update gguf-py/gguf/tensor_mapping.py Co-authored-by: Sigbjørn Skjæret sigbjorn.skjaeret@huggin
ci : Restore ROCm job for Ubuntu (#27399) Revert 'ci : disable ubuntu-rocm (#26969)' This reverts commit 9558fa4. ci: set ccache compiler_check=content for ROCm build The ROCm toolchain is pip-install
common: add json.h abstraction (#27511) add common/json migrate common adapt jinja migrate server big wip migrate tests wip revert some excessive changes wip wip 2 revert redundant changes fix server
mtmd: use ggml_rope_set_offset (#27521) mtmd: use ggml_rope_set_offset add comment Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42334609 macOS/iOS: macOS
vulkan : added the PAD_REFLECT_1D operation (#26586) vulkan : added PAD_REFLECT_1D operation Implemented the GGML_OP_PAD_REFLECT_1D operation for the Vulkan backend Changes: pad_reflect_1d.comp: imple
common: json.h: fix clang lto (#27575) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42368523 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (
The artificial intelligence industry wants enterprises to measure progress in tokens, model calls and usage. But those are largely vendor-revenue metrics — not enterprise-value metrics. The Canva exam
Has anyone else experimented with ubatch size when running GLM-5.2 locally? I was testing the 226 GiB GLM-5.2-UD-IQ2_XXS GGUF on 3x RTX PRO 6000 Blackwell GPUs and got a pretty interesting result. Wit
Hi there. I have a Lenovo Yoga laptop (16GB LPDDR5, AMD Ryzen 7 8840HS, integrated GPU). I have downloaded Ollama and installed Qwen 2.5coder:3b llm. I have tried OpenClaw, and Qwen Code but it seems
Using llama.cpp I seem to be unable to get my to GPUs working tougether correclty, so I need help somehow. Setup: 96GB RAM, one Blackwell 5000 (48GB) and one 3090 (24GB). I am trying to run the UD-Q3_
One of the interesting properties we’ve observed around schema-guided, complex document extraction tasks 📑 is that coding agent harnesses are good baselines (in terms of cost/accuracy) for longer docu
Own your harness. I really wish the Claude Code harness was open source. I often think about this, especially now that custom harnesses are foundational to AI-native companies. I am now building mostl
The preprint can be accessed via the following link: https://arxiv.org/abs/2608.12408 (q-bio.NC / cs.LG). And for the code: https://github.com/nilsleut/evaluation-resolution-rsa The following assertio
Released today: version 1.3 of energygraph Zero build dependencies, lightweight tool for live views of the power-consumption. Version 1.3 adds support for dGPUs from nvidia, intel, amd. Depending on v
Where did THESE things go for ChatGPT? Claude has them. I can access older versions of my conversations easily. But with ChatGPT, they just DISAPPEARED. I can no longer regenerate answers or following
arXiv:2608.20181v1 Announce Type: cross Abstract: Studies of machine-learning-based power-system protection increasingly report near-perfect scores, yet the meaning of those scores depends strongly on
arXiv:2608.19487v1 Announce Type: cross Abstract: Python is widely used in scientific research because it enables rapid development and provides rich ecosystems for data analysis, artificial intellige
Claude: Anthropic says Mythos 5 is now in public beta in Claude Security for Enterprise users, and it is working with providers to embed Mythos 5 in defensive tools — No items found. — Reading time —
OpenAI announced on August 21, 2026 that it will reduce API and credit pricing for GPT‑5.6 Sol by more than 20% for the next three months. The price cut is part of the company’s broader effort to push
Revert 'tensor-split meta backend fixes (#26502)' (#27433) This reverts commit d59d455. Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42012577 macOS/iOS:
common : gracefully fallback on unsupported regex patterns in JSON schema (#26939) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42018860 macOS/iOS: macOS
CUDA: adding switch points per HW and quant type to tune the mvq->MMQ decode crossover (#26079) CUDA: runtime GGML_CUDA_MMVQ_MAX to tune the mvq->MMQ decode crossover Add a runtime override of the mul