THE NEXT THING, IN PERSPECTIVEIndependent AI news

The latest AI news, in focus.

Official announcements and industry reporting. Follow the timeline of AI.

Today's news

News feed
AI NEWS × FOXAIFind your next step

From the latest news to tools that work for you.

Explore AI tools

Discover AI tools

Swipe to explore →
EvoMapAgents & automation

EvoMap is an infrastructure designed for the reuse of AI agent experiences. It organizes reusable strategies and repair assets with validation records via the GEP protocol, offering market- and quality-based ranking to help different agents discover, evaluate, and inherit existing capabilities.

分类推荐榜首Pricing unverifiedView details
KiroCoding & development

Kiro is an AI engineering tool for developers that organizes requirements into specifications, architectures, and tasks, generates code via agents, and supports the entire workflow—from planning to code review—through property-based testing, execution memory, and integration with development tools.

AI programming tool分类推荐榜首Free entry availableView details
Open DesignDesign & 3D

OpenDesign is an open-source design workspace that can connect to local programming agents, use built-in model keys, or leverage cloud-based models, enabling the creation and editing of prototypes, web pages, presentations, and graphics within a design system; local software is free, while model services incur fees.

AI design tools分类推荐榜首Free entry availableView details
AI ShortWriting & content

AIS ALERT MANAGEMENT AND SHARED PLATFORM, QUICK COMMAND FOR MULTIPLE SCENARIOS

AI HINT COMMAND分类推荐榜首Pricing unverifiedView details
MemoAIMusic & audio

MemoAI is an AI-powered audio-to-text tool that supports localized processing, eliminating the need to upload data and ensuring privacy and security. It can efficiently transcribe YouTube videos, podcasts, and local audio/video files (such as MP4, MP3, AAC, M4A, etc.) into text, while offering core features like multilingual translation, speaker diarization, real-time subtitles, and floating notes—making it ideal for meeting minutes, interview transcription, content learning, and more.

AI AUDIO TOOL分类推荐榜首Pricing unverifiedView details
JAXModels & infrastructure

JAX is a Python numerical computing library designed for accelerators, specifically tailored for high-performance array computations and large-scale machine learning. It efficiently executes the same code across multiple hardware backends, including CPUs, GPUs, and TPUs.

AI DEVELOPMENT PLATFORM分类推荐榜首Pricing unverifiedView details
妙播Marketing & commerce

THE AI LIVE TELECOM SOLUTION FROM THE AD

待评分 · 分类列表推荐Pricing unverifiedView details
IamIPIndustry applications

IamIP is a patent platform that provides AI-driven search, analysis, and monitoring tools to help users conduct patent research and management more efficiently.

待评分 · 分类列表推荐Pricing unverifiedView details
闪剪Video & animation

FlashCut is an AI-powered creation platform for voice-over and short-video marketing. It can generate finished videos with voiceovers, subtitles, and visuals from text, offering digital human customization, photo-to-digital-human conversion, live-stream clipping, and intelligent editing. It also provides APIs and branded mini-program solutions for businesses.

待评分 · 分类列表推荐Pricing unverifiedView details
Meta AI助手Assistants & search

Meta's free AI chat assistant

待评分 · 分类列表推荐Pricing unverifiedView details

AI, as it happens

1434 published stories
Meta / Llama · 117 storiesClear filters ×
2026-10-28
Follow the timeline
08:32

b11940

common : support deferred tools and tool_reference content parts (#31256) * common : support deferred tools and tool_reference content parts Assisted-by: Claude Opus 5.5 * common : credit the deferred tools co-author Co-authored-by: Aldehir Rojas <765227+aldehir@users.noreply.github.com> Assisted-by: Claude Opus 5.5 * tests : fix the type check of the deferred tools test Assisted-by: Claude Opus 5.5 **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11940/llama-b11940-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11940/llama-b11940-bin-maco

Inference systems & deploymentRead original Timeline and related coverage
07:57

b11938

model : add LiquidAI/LFM3-1.7B and LiquidAI/LFM3-VL-2B (#31252) * model : add LiquidAI/LFM3-1.7B and LiquidAI/LFM3-VL-2B Co-authored-by: Sigbjørn Skjæret Co-authored-by: Xuan-Son Nguyen Assisted-by: Claude Opus 5.5 * tests : set the SWA rope base of the LFM3 test model Assisted-by: Claude Opus 5.5 * llama : enable recurrent state rollback for LFM3 Assisted-by: Claude Opus 5.5 * Apply batched suggestions from code review Co-authored-by: Sigbjørn Skjæret Co-authored-by: Tarek Dakhran * gguf-py : align the lfm3vl mapping comments Assisted-by: Claude Opus 5.5 * llama : fix CRLF line endings from the review suggestions Assisted-by: Claude Opus 5.5 * tests : add the LFM3 Metal fusion baseline Assisted-by: Claude Opus 5.5 * llama : keep the LFM3-VL depthwise conv kernels in F32 when quantizing As

Inference systems & deploymentRead original Timeline and related coverage
04:11

b11936

llama : build the k-pool sparse selection independently of the kq_mask shape (#31243) * llama : build the k-pool sparse selection independently of the kq_mask shape qwen4exp (QSA) and glm5-next pick cache cells by index and no longer read the kq_mask values, but they still shaped the selection like the dense kq_mask. Build the selection as [n_kv, n_tokens] over the cache cells instead, and split it into the streams of the K view right before the attention. This lets the dense mask change its layout (e.g. per-sequence row lists) without touching the sparse layers. * improve comment Co-authored-by: Georgi Gerganov * llama : drop the unused build_fused_top_k_mask args, review nits --------- Co-authored-by: Georgi Gerganov **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silico

Inference systems & deploymentRead original Timeline and related coverage
01:15

b11935

rpc : catch std::length_error in recv_msg (#31204) **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11935/llama-b11935-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11935/llama-b11935-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/llama.cpp/releases/download/b11935/llama-b11935-xcframework.zip) **Linux:** - [Ubuntu x64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11935/llama-b11935-bin-ubuntu-x64.tar.gz) - [Ubuntu arm64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11935/llama-b1

Inference systems & deploymentRead original Timeline and related coverage
2026-10-27
Follow the timeline
23:50

b11934

tests : speed up tensor data generation and model saving (#31145) * tests : speed up tensor data generation and model saving Refactor save_models to create only the model, not a llama_context, which is not needed for saving. Replace the slow std::normal_distribution draw with a splitmix64 + box-muller normal generator. Assisted-by: dsh:llama.cpp/ggml-org/DeepSeek-V4-Flash-Vision-Exp-GGUF (Mac Studio, M2 Ultra, 192GB) * tests : use a shared pool of normal draws for tensor data Generate a pool of normal draws once with std::mt19937 and std::normal_distribution, then have every tensor copy a cyclic slice of it instead of running a per-tensor box-muller. The name-keyed slice selection keeps the draw deterministic. Assisted-by: dsh:llama.cpp/ggml-org/DeepSeek-V4-Flash-Vision-Exp-GGUF (Mac Studi

Inference systems & deploymentRead original Timeline and related coverage
22:04

b11932

mtmd-cli: Reduced scope of batch pointer (#31155) (#31165) * mtmd-cli: Reduced scope of batch pointer (#31155) * mtmd-cli: Remove CLI regression test **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11932/llama-b11932-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11932/llama-b11932-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/llama.cpp/releases/download/b11932/llama-b11932-xcframework.zip) **Linux:** - [Ubuntu x64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11932/llama-b11932-bin-ubuntu-x64.tar

Inference systems & deploymentRead original Timeline and related coverage
21:42

b11930

server: add router device memory margin parameter for dynamic unloading (#21231) * server: add --models-memory-max parameter to allow dynamically unloading models when they exceed a memory size threshold estimate with to-be-loaded model size included use no_alloc to get memory requirements for model load only set model memory_mb if not previously calculated use memory margin instead of total size limit, apply to each device separately add server memory debug logging move llama_context_device_memory function to llama-ext.h fix model count exceeded check improve memory_per_device map naming improve variable naming, fix style also strip models memory margin from child processes cont : clean-up replace device memory map with buft memory map. Use llama_get_memory_breakdown extract duplicated ch

Inference systems & deploymentRead original Timeline and related coverage
21:03

b11929

common : --model or -hf is required (#31242) Assisted-by: Codex **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11929/llama-b11929-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11929/llama-b11929-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/llama.cpp/releases/download/b11929/llama-b11929-xcframework.zip) **Linux:** - [Ubuntu x64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11929/llama-b11929-bin-ubuntu-x64.tar.gz) - [Ubuntu arm64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b1

Inference systems & deploymentRead original Timeline and related coverage
19:08

b11927

metal : add GGML_OP_TOP_K_MASK (#31230) * metal : add GGML_OP_TOP_K_MASK Port the fused top-k mask from the CUDA backend: a grid-stride kernel fills the output with -inf, then one threadgroup per row copies the mask values of the kept cells over it. Function constants drop the score, blocks and extra paths that a given graph does not use. Assisted-by: dsh:llama.cpp/ggml-org/Qwen3.8-Flash-Next-GGUF (MacBook, M5 Max, 128GB) * metal : clean up top-k mask kernel formatting Assisted-by: dsh:llama.cpp/ggml-org/Qwen3.8-Flash-Next-GGUF (MacBook, M5 Max, 128GB) **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11927/llama-b11927-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](

Inference systems & deploymentRead original Timeline and related coverage
18:29

b11926

glm5-next : QSA without kq_mask (+ DSA) (#31009) * glm5-next : QSA without kq_mask * DSA: dsa top-k mask change tested with layers 0,3 of DeepSeek-V3.2 **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11926/llama-b11926-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11926/llama-b11926-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/llama.cpp/releases/download/b11926/llama-b11926-xcframework.zip) **Linux:** - [Ubuntu x64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11926/llama-b11926-bin-ubuntu-x64.t

Inference systems & deploymentRead original Timeline and related coverage
17:05

b11925

ci : use ccache depend mode for ROCm release builds (#31231) **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11925/llama-b11925-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11925/llama-b11925-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/llama.cpp/releases/download/b11925/llama-b11925-xcframework.zip) **Linux:** - [Ubuntu x64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11925/llama-b11925-bin-ubuntu-x64.tar.gz) - [Ubuntu arm64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b1192

Inference systems & deploymentRead original Timeline and related coverage
16:50

b11923

ggml : relax scale to non-contiguous views (#31146) * ggml : relax scale to non-contiguous views SCALE requires src0 to be 1D-contiguous. Relax this to dim-0 contiguity with arbitrary row strides (ggml_is_contiguous_rows), same as the unary ops in the Metal backend. - ggml: relax the assert in ggml_scale_impl - cpu: address src0 rows through their byte strides (f32, bf16) - cuda: stride-aware kernel, fastdiv row decomposition - metal: support bf16 scale (guarded unary instantiations) - sycl: keep the gate at contiguous, the kernel is flat f32 Add view variants to test_scale. Verified on CPU, Metal and CUDA. Assisted-by: dsh:llama.cpp/ggml-org/Qwen3.8-Flash-Next-GGUF (RTX 5090 32GB, DDR5 128GB) * cuda : skip the softcap fusion on non-contiguous views The fused softcap kernel reads src and w

Inference systems & deploymentRead original Timeline and related coverage
16:42

An agent that helps users snatch up iPhone foldable-screen models is reshaping e-commerce in the AI era.

Shopify AI saw transaction orders grow by a factor of 13. In the early morning of October 16, pre-orders began for Apple’s first foldable-screen phone, the iPhone Duo. Priced from $1,999, initial production capacity was tight, and delivery dates quickly slipped to January next year. This time, it wasn’t just people glued to their screens vying for the launch. This year, Agents have become the strictest “parents” to iPhone Duo scalpers. | Image source: Apple On the MacRumors forum, one user who missed out vented their frustration, while on Reddit others boasted about using Muse bots to snatch up multiple units in a single go—before even a human could click, the inventory had already been swept away by AI. Muse is Meta’s personal Agent, launched on September 8; shortly after its debut, it topped the free charts in the U.S. App Store, with downloads soon surpassing 2.5 million. According to Meta, shopping has become one of Muse’s primary use cases. Around the time the iPhone Duo went on sale, Muse even released a new version optimized for this device, with an update note that emphasized making in‑chat shopping more convenient. AI-driven shopping has moved from demo stages at product launches straight into real‑world stock‑picking scenarios. But not every retailer welcomes these new customers. In late September, Amazon began banning Muse. By October, users of Musk’s Grok Bot encountered the same warning when placing orders on Amazon: unauthorized AI agents continuing to access the platform violates Amazon’s terms of service. Meanwhile, another company, Shopify, took the exact opposite approach. On September 21, Shopify CEO Tobi Lütke announced a deep partnership with Muse, enabling Muse to process checkout through Shop Pay across all Shopify stores. And Shopify’s financial reports already show early returns from this strategy: in the first quarter of this year, orders generated via AI search grew nearly thirteenfold year over year, while store traffic driven by AI increased eightfold. Closing doors on one hand, opening them on the other—when more and more purchasing decisions are handed over to Agents, will the question of “whether or not to let Agents in” become…

Robotics & embodied AIRead original Timeline and related coverage
16:18

b11922

tests : fix multi-seq split replay and partial-state cases for trained models (#31216) * tests : fix multi-seq split replay against trained models The rollback snapshot is captured by the graph that decoded the rolled-back tail, while the reference context decodes the same prefix with different batch shapes. The two states differ in the last bits and trained models amplify the difference with every replayed token. Compare only the first replayed token of each seq, then restore the rollback state into the reference context before the seq-1-only tail. Assisted-by: dsh:llama.cpp/ggml-org/Qwen3.8-Flash-Next-GGUF (MacBook, M5 Max, 128GB) * tests : require bit-exact seq-1 isolation in split replay The seq state save is not always bit-lossless - SWA-masked cells are not saved - so reload the rest

Inference systems & deploymentRead original Timeline and related coverage
16:01

b11921

llama : remove obsolete n_pad from memory instances (#31200) Every caller passes n_pad = 1 and the only two users clamp it to 256. Drop the parameter, the members and the trivially true kv_size assert. Assisted-by: dsh:llama.cpp/ggml-org/Qwen3.8-Flash-Next-GGUF (MacBook, M5 Max, 128GB) **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11921/llama-b11921-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11921/llama-b11921-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/llama.cpp/releases/download/b11921/llama-b11921-xcfram

Inference systems & deploymentRead original Timeline and related coverage
15:42

b11919

opencl: refine mul mat id -1 index (#31170) Assisted-by: Codex **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11919/llama-b11919-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11919/llama-b11919-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/llama.cpp/releases/download/b11919/llama-b11919-xcframework.zip) **Linux:** - [Ubuntu x64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11919/llama-b11919-bin-ubuntu-x64.tar.gz) - [Ubuntu arm64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11

Inference systems & deploymentRead original Timeline and related coverage
15:22

b11918

opencl: initial support for q3_k (#31172) * opencl: add q3_k AOS mv * opencl: add q3_k SOA mv and tiled mm * opencl: add q3_k noshuffle gemm/gemv for adreno * opencl: add q3_k dp4a gemm for adreno Co-authored-by: Hongqiang Wang * opencl: tune q3_k kernels for adreno Co-authored-by: Hongqiang Wang --------- Co-authored-by: Hongqiang Wang **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11918/llama-b11918-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11918/llama-b11918-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/ll

Inference systems & deploymentRead original Timeline and related coverage
12:00

Merge-Based Vocabulary Extension Stalls at the Pre-Tokenizer: An English-Preserving Repair for Nepali, Tested in Continued Pretraining

arXiv:2610.33016v1 Announce Type: new Abstract: Llama-3, Qwen3 and many other open LLMs pre-tokenize text with a regular expression whose word class admits only Unicode letters. Devanagari vowel signs are combining marks, so this regex splits Nepali words at every vowel sign, and models that use it spend 2.6 to 5.8 times as many tokens on Nepali as on the same English text. Merge-based vocabulary extension cannot join the split pieces: for Llama-3 and Qwen3 no number of new merges brings this premium below about 2.2, and ten released extensions already end within 4% of this pre-token floor. We give a small edit to the deployed regex that provably keeps every token id of text without combining marks or Devanagari characters. With 32,000 new merges it brings Nepali to 1.01 times English, abo

M&A & IPOsRead original Timeline and related coverage
12:00

Prefix Keys and Values Matter Late: RIPE Caching for Self-Speculative Decoding in Looped Language Models

arXiv:2610.34436v1 Announce Type: new Abstract: A looped language model runs one block of layers $R$ times per token, so its key-value (KV) cache holds $R$ planes per recurrent layer, and self-speculative decoding (draft with a few loops, verify with all $R$) reads them all. We find that prefix keys and values matter late: at equal perturbation size, loops 25-32 are $e^{15.0}$ (Huginn-0125) and $e^{12.8}$ (Raven-Llama) times more sensitive than loops 2-9, and the planes approach the last one geometrically, at 0.88 and 0.81 per loop. Whether the deepest plane alone suffices then depends on how ripe the late planes are: on Raven, storing only it is indistinguishable from a bf16-to-fp32 change; on Huginn, 1% of positions carry 74% of its error. We turn this into RIPE, which stores the deepest

Model releases & updatesRead original Timeline and related coverage
Every update links back to its source.
STAY CURIOUS. STAY INFORMED.

Stay informed about the changes that matter.

Give your curiosity direction. Make every day of reading count.

Subscribe via RSS