AI Intelligence Stream

一个模型场景通吃！它石智航AWE3.7的泛化能力有点狠

从工厂一路干到家庭

b10775

mtmd: fix idefics3 preproc ( #28273 ) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44875193 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 10.0) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android: Android arm64 (CPU) Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.3 DLLs Windows arm64 (CUDA 13) (preview) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 10.0) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openEuler aarch64 (910b, ACL Graph) UI: UI

b10774

finetune: fix no KV cache ( #27199 ) training: fix no KV cache apply @ ggerganov suggestion Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44871971 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 10.0) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android: Android arm64 (CPU) Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.3 DLLs Windows arm64 (CUDA 13) (preview) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 10.0) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openEuler aarch64 (910b, ACL Graph) UI: UI

0.153.0

New Features Vim mode now supports undo with u and redo with Ctrl+R , preserving complete drafts including pasted content and attachments. ( #41941 , #42140 ) The plugin CLI can list, install, and remove plugins from remote marketplaces. ( #42150 ) Set tui.auto_recap = false to disable automatic recaps while keeping manual /recap available. ( #42101 ) TUI history shows complete patches, input sent to background terminals, and individual completed commands. ( #41893 , #42107 ) Plus and Team users receive an earlier warning when less than half of their allowance remains in an approximately five-hour usage window. ( #42142 ) Bug Fixes TUI sessions reconnect after an external app-server connection drops, preserving drafts and transcripts while keeping uncertain or queued submissions paused for review. ( #41911 , #41916 , #41918 ) Full Access skips Guardian reviews for confirmation-only actions. User approval mode skips background Guardian scoring and prewarming, while sensitive-action checks and requests for user input retain their existing handling. ( #42147 , #42256 ) Guardian review history survives compaction, restarts, and user-created forks while respecting rollback boundaries an

神秘具身团队又放出一连串很炸的Demo视频…自进化模型，技术路线曝光

b10773

server : accept data: URLs for input_video and input_audio ( #27735 ) server : accept data: URLs for input_video and input_audio input_video and input_audio passed accept_base64_uri=false to handle_media(), so data: URLs got treated as raw base64 strings and failed later with a confusing media probe error ( #27724 ). pass true for these two content types the same way image_url already does, and allow video/audio mime types in the data: url check instead of image only. data URL validation now throws std::invalid_argument so malformed input comes back as 400 instead of 500, matching the other input validation in this file. server : simplify handle_media and drop unused accept_base64_uri flag server : update comment and add unit test for invalid data URI MIME Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44868521 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 10.0) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android:

50万！李想宣布MPV进入iPhone时刻！外观没改配置拉满

4激光雷达+双M100上车

今年最难的机器人Demo，“机器人含量”为0

遥操，可能真的危险了

b10772

ggml-hexagon: add F16 support for unary ops ( #28228 ) Extend the HTP backend's F16 unary op coverage to include ABS on top of the existing NORM/RMS_NORM/L2_NORM/SCALE/CLAMP/SQR/SQRT set. Add hvx_abs_f16_{aa,au,ua,uu} + dispatcher in hvx-arith.h, mirroring the sqr_f16 kernel structure and using the existing hvx_vec_abs_f16() sign-bit-clear helper Add abs_f16() row-wise dispatch and DEFINE_UNARY_TASK_F16(unary_abs, ...) in unary-ops.c, wired into execute_op_unary()'s op_type/task_func switches Register HTP_OP_UNARY_ABS in htp_op_is_unary() (unary-ops.h) so that ggml_hexagon_precompute_unary_params() fills kernel_params (n_threads, VTCM layout) for ABS nodes -- required for the F16 path to function Narrow the F16 GGML_OP_UNARY gate in ggml_hexagon_supported_unary() (ggml-hexagon.cpp) to allow GGML_UNARY_OP_ABS specifically, instead of rejecting all GGML_OP_UNARY ops for F16 Merge the separate execute_op_unary_f32()/execute_op_unary_f16() functions into a single execute_op_unary(), branching on an is_f16 flag for the parts that actually differ by type (elem_size, the early F16 op-support check, and which task_func table to use) while keeping the F32-only tiled/RMS_NORM_MUL paths intac

0.153.0-alpha.5.1

Release 0.153.0-alpha.5.1

b10771

mtmd: add mtmd_tokenize_from_parts() ( #28250 ) add mtmd_tokenize_from_parts use it in mtmd-cli move add_special to call level Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44860492 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 10.0) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android: Android arm64 (CPU) Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.3 DLLs Windows arm64 (CUDA 13) (preview) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 10.0) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openEuler aarch64 (910b, ACL Graph) UI: UI

ChatGPT Work Mode High Error Rates

Status: Resolved All impacted services have now fully recovered. Affected components File uploads (Operational) ChatGPT Work (Operational) Search (Operational) Compliance API (Operational) GPTs (Operational) Deep Research (Operational) ChatGPT Atlas (Operational) Codex in ChatGPT Desktop (Operational) Conversations (Operational) Image Generation (Operational) Agent (Operational) Connectors/Apps (Operational) Sites (Operational) Login (Operational) Voice mode (Operational)

b10770

metal : add fa-vec tunings for M3 ( #28236 ) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44857719 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 10.0) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android: Android arm64 (CPU) Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.3 DLLs Windows arm64 (CUDA 13) (preview) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 10.0) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openEuler aarch64 (910b, ACL Graph) UI: UI

b10769

metal : fix memory query under low-memory conditions ( #27701 ) metal: Fix memory query under low-memory conditions Simply variable name Co-authored-by: Georgi Gerganov [email protected] Write it even shorter Co-authored-by: Niklas Wenzel [email protected] Co-authored-by: Georgi Gerganov [email protected] Co-authored-by: Niklas Wenzel [email protected] Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44854603 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 10.0) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android: Android arm64 (CPU) Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.3 DLLs Windows arm64 (CUDA 13) (preview) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 10.0) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openE

v0.33.3

What's Changed Report cached prompt tokens Honor GGUF model defined default parameters MLX, MLX-C, llama.cpp update New Contributors @marcelpetrick made their first contribution in #17579 Full Changelog : v0.33.2...v0.33.3-rc0

b10767

Update ROCm to 10.0.0 release ( #27803 ) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44848715 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 10.0) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android: Android arm64 (CPU) Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.3 DLLs Windows arm64 (CUDA 13) (preview) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 10.0) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openEuler aarch64 (910b, ACL Graph) UI: UI

Launch HN: RonanRX (YC S26) – Personalized Peptides and GLP-1s

Comments

v2.1.259

What's changed Added managedMcpServers managed setting: organizations can provide HTTP/SSE MCP servers to every user (same entry shape as .mcp.json ); entries that name a command to run are skipped Added --permission-prompts none for unattended headless hosts: anything that would prompt is denied automatically while the active permission mode (including auto mode) keeps deciding Added recognition of glab mr create/merge/close/reopen/note/update so GitLab merge requests show as MR !N in the collapsed tool summary and refresh the footer MR badge Added --json to claude plugin validate for a machine-readable validation report Fixed concurrent sessions silently reverting each other's ~/.claude.json changes — workspace trust no longer resets and MCP/project state is no longer lost when running many sessions at once Fixed a conversation whose thinking was rejected once being rejected again on every later turn Fixed Bash Read() deny rules not covering files given as option values ( --ignore-revs-file=.env , -f.env , @file ), git diff / git grep file operands, or cd DIR && cat FILE compounds; grep -r / cp -r over a directory holding a denied file now asks Fixed the prompt cache being invali

Reasons robotics is hard

Comments

Elevated errors for Claude Sonnet 5

Sep 2 , 21:44 UTC Resolved - The issue affecting Claude Sonnet 5 has been resolved. Impact occurred from 2:05pm PT / 21:05 UTC to 2:19pm PT / 21:19 UTC. Sep 2 , 21:17 UTC Investigating - We are investigating elevated errors on requests to Claude Sonnet 5. We will provide an update as soon as possible.

v1.18.27

Core Bugfixes Default provider header timeouts to five minutes so slow model startups fail less often. Default streamed chunk timeouts to five minutes, with false supported to disable them. Let Anthropic thinking.blockBinding opt out via config when you need to keep the provider default. ( @dkindlund ) Limit Anthropic thinking block binding to Claude 5.1+ models so older deployments do not reject requests. Avoid unhandled errors when canceling timed-out SSE reads. ( @AlexanderWillner ) Thank you to 2 community contributors: @AlexanderWillner : fix(provider): handle SSE reader cancel rejections ( #44944 ) @dkindlund : fix(opencode): let config opt out of Anthropic thinking blockBinding ( #46820 )

Nango (YC W23) is hiring across eng, product and GTM (SF and remote)

Comments

b10766

model: correctly support input vision for deepseek4 ( #28154 ) model: correctly support input vision for deepseek4 nits Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44819834 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 7.14) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android: Android arm64 (CPU) Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.3 DLLs Windows arm64 (CUDA 13) (preview) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 7.14) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openEuler aarch64 (910b, ACL Graph) UI: UI

Altair Basic Interpreter Source Code (1975) [pdf]

Comments

b10764

ggml-cuda : remove unused vars ( #28235 ) Signed-off-by: Adrien Gallouët [email protected] Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44811308 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 7.14) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android: Android arm64 (CPU) Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.3 DLLs Windows arm64 (CUDA 13) (preview) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 7.14) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openEuler aarch64 (910b, ACL Graph) UI: UI

Fable 5.1 World Modeling

Comments

Muse Spark 1.3

Comments

Updated: NVIDIA Technical Blog

Visible page content changed; normalized text length 0 characters.

b10763

common, server : enable preserve_reasoning kwarg by default, log its effective state ( #28174 ) common, server : enable preserve_reasoning kwarg by default, log its effective state If the preserve_reasoning chat template kwarg is not specified explicitly via --reasoning-preserve / --no-reasoning-preserve, it is enabled by default after argument processing. The server logs the effective state of the kwarg, warns that it is enabled by default when the template supports it, and only warns "has no effect" when it was enabled explicitly on a template that does not support it. Setting the kwarg via --chat-template-kwargs is deprecated. Assisted-by: pi:llama.cpp/Qwen3.8-27B cont : update comment Co-authored-by: Xuan-Son Nguyen [email protected] Co-authored-by: Xuan-Son Nguyen [email protected] Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44798750 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 7.14) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32)

Humanity has built the records of FATE by accident

Comments

Elevated errors creating new accounts

Status: Resolved The issue affecting new account creation has been resolved.

How the Hell Did an Island Suddenly Appear, Then Vanish?

A floating mass of trees and vegetation in British Columbia’s Williston Lake attracted international attention after it was filmed this summer by boaters—and then seemed to disappear.

The American Worker vs. the Most Qualified

Comments

Qantas Airbus A380 engine failure in 2010 (2023)

Comments

b10762

mtmd: support DeepSeek-V4-Flash-Vision-Exp ( #28133 ) mtmd: support DeepSeek-V4-Flash-Vision-Exp handle min/max token counts from CLI rm debugging use GGML_ROPE_TYPE_VISION nits apply review comments correct token count Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44793186 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 7.14) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Android: Android arm64 (CPU) Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.3 DLLs Windows arm64 (CUDA 13) (preview) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 7.14) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openEuler aarch64 (910b, ACL Graph) UI: UI

Embedded Rust RTOS vs. C RTOS

Comments

Firefox's AI Switch Is Off. Telemetry Isn't

Comments

Modernizing and scaling support operations with generative AI on AWS

Learn how to build a generative AI-based support operations platform on AWS that converts training videos into structured SOPs, applies Retrieval-Augmented Generation to guide ticket resolution, and uses machine learning to predict SLA risk and prioritize work.

Why do so many tools have JSON config files?

Comments

How an AWS team detects dashboard content failures at scale using Amazon Bedrock

Business intelligence dashboards can fail silently, showing blank, stale, or wrong data even when every infrastructure monitor reports healthy. Learn how an AWS team built an automated, AI-powered content validation solution on Amazon Bedrock that scans hundreds of dashboards and alerts owners, cutting mean time to detection from days to under an hour.

From code to diagrams: Agentic architecture documentation with Amazon Bedrock AgentCore

Learn how a global interdealer broker built an automated architecture documentation pipeline on Amazon Bedrock AgentCore that analyzes .NET code bases, generates architecture diagrams, and maintains searchable documentation through Amazon Bedrock Knowledge Bases and AWS CodePipeline.

Trinity: Agentic AI-powered transition planning for students with disabilities

Learn how University Startups and its AWS partner g/d/n/a scaled Trinity, a conversational AI solution for students with disabilities, into a serverless multi-agent architecture on Amazon Bedrock that produces IDEA-aligned transition plans for school districts across the US.

Google releases Gemini 3.8 Flash, its third Flash model in six weeks

Google's Pro model updates are seemingly paused, but there's yet another Gemini Flash today.

Uber is shutting down services in Nigeria and Uganda from September 2 after a global review of its operations; it launched in Nigeria in 2014 and Uganda in 2016 (Bloomberg)

Bloomberg : Uber is shutting down services in Nigeria and Uganda from September 2 after a global review of its operations; it launched in Nigeria in 2014 and Uganda in 2016 — US ride-hailing giant Uber Technologies Inc. is shutting down its services in Nigeria and Uganda from September 2, following a global review of its operations.

Trump may be forced to reveal secret rules feds use for AI safety testing

Trump’s secret reviews of frontier AI models may hide corruption, lawsuit says.

Amazon’s AI assistant can now spot fake emails from the company

Amazon is trying to combat impersonation scams with a new feature that allows you to use its AI assistant to determine whether an email, text message, or phone call actually came from the company. With the update, you can ask Alexa for Shopping about a message you received, and it will use AI to compare […]

I Don't Have a Smartphone

Comments

X says US payouts for its Original Content Rewards and Subscriptions will now be handled through X Money (Sarah Perez/TechCrunch)

Sarah Perez / TechCrunch : X says US payouts for its Original Content Rewards and Subscriptions will now be handled through X Money — Elon Musk's social network X on Wednesday said that starting today, all of its U.S. creator payouts will be handled through X Money, the app's payments service.

v0.33.3

What's Changed Report cached prompt tokens Honor GGUF model defined default parameters MLX, MLX-C, llama.cpp update New Contributors @marcelpetrick made their first contribution in #17579 Full Changelog : v0.33.2...v0.33.3-rc0

Paint.net 5.2 alpha now runs on Linux

Comments

Pangram’s Max Spero on why AI detection is harder than ‘Real or Fake’

The internet has a trust problem, and it’s not just because social media feeds are filling up with AI slop. AI-generated text and images are now making their way into job applications, product reviews, and even insurance claims, leaving platforms and users alike scrambling to figure out what’s real. A handful of startups have cropped up in the past couple of […]

We’re ‘dangerously close’ to dead internet theory, says Pangram’s CEO

The internet has a trust problem, and it’s not just because social media feeds are filling up with AI slop. AI-generated text and images are now making their way into job applications, product reviews, and even insurance claims, leaving platforms and users alike scrambling to figure out what’s real. A handful of startups have cropped up in the past couple of […]

US government sides with OpenAI on issue of training LLMs on copyrighted material

"The United States has a strong interest in continuing to develop a robust and competitive artificial intelligence industry that sets the standard for the practice and procedure of AI use globally," the brief reads.

Tangle – Visual ML Pipeline Editor

Comments

Researchers fear safety disaster ahead of OpenAI’s Astra release

OpenAI is on the cusp of releasing its most powerful AI model yet, Astra, following weeks of delays to shore up safety protocols after its agents attacked real targets during testing. As details about the model trickle out, researchers are warning it "may be the single worst development for AI security/safety to date." Shortly after […]

Thyme Care, which focuses on filling cancer treatment gaps via a virtual navigation platform, raised a $125M Series E at a $2B+ valuation, up from $1B+ in 2025 (Ian Thomas/CNBC)

Ian Thomas / CNBC : Thyme Care, which focuses on filling cancer treatment gaps via a virtual navigation platform, raised a $125M Series E at a $2B+ valuation, up from $1B+ in 2025 — Despite recent advancements in oncology, the branch of medicine focused on the study, prevention, diagnosis and treatment of cancer …

llm-gemini 0.34

Release: llm-gemini 0.34 New model gemini-3.8-flash for Gemini 3.8 Flash , with low, medium and high thinking levels. #146 Fixed async responses failing to record the resolved model version. Thanks, Charlie Tonneslan . #137 Tags: llm , gemini

How to debloat your Xiaomi 15 Ultra without root access

Comments

Commerce Secretary Howard Lutnick says "we trust Anthropic" as they are "back on the right side" with the administration and that "they've done what we asked" (Maria Curi/Axios)

Maria Curi / Axios : Commerce Secretary Howard Lutnick says “we trust Anthropic” as they are “back on the right side” with the administration and that “they've done what we asked” — - “They've done what we asked. … - “Really excited for our conversation …

The race to engineer new knobs for the human brain

Comments

Huskeys, which uses agentic AI to help companies block AI-driven attacks, raised a $27M Series A led by Blackstone Innovations Investments at a $100M+ valuation (Maria Armental/Wall Street Journal)

Maria Armental / Wall Street Journal : Huskeys, which uses agentic AI to help companies block AI-driven attacks, raised a $27M Series A led by Blackstone Innovations Investments at a $100M+ valuation — The company, valued at more than $100 million, uses agentic AI to block threats — Blackstone is placing an early bet on Huskeys …

The US Army awards $192M in contracts to Palantir and Anduril to produce eight TITAN AI-powered mobile ground stations, to be delivered over the next 18 months (Barratt Dewey/Tectonic Defense)

Barratt Dewey / Tectonic Defense : The US Army awards $192M in contracts to Palantir and Anduril to produce eight TITAN AI-powered mobile ground stations, to be delivered over the next 18 months — Watch out, world. The US Army's kill chain is going mobile. — On Tuesday, the Army awarded two contracts worth a combined $192M …

Show HN: FrontierHarness Eval – 9 harness, same model, cost per pass varies 17x

Comments

The Trump administration is supporting OpenAI in the NYT copyright lawsuit

The Trump administration has intervened in The New York Times' copyright lawsuit against OpenAI, making an argument in favor of the AI lab. The landmark lawsuit, filed in December 2023, alleging that OpenAI unlawfully trained its AI systems on articles from The New York Times and seeks to recoup "billions of dollars" in damages from […]

MrBeast, whose YouTube channel hit 500M subscribers in June, will put Gemini, Google Health, and the Fitbit Air in his videos as part of a multiyear Google deal (Stevie Bonifield/The Verge)

Stevie Bonifield / The Verge : MrBeast, whose YouTube channel hit 500M subscribers in June, will put Gemini, Google Health, and the Fitbit Air in his videos as part of a multiyear Google deal — One of YouTube's biggest creators has a new deal to promote Gemini and Google Health.

A16z Says You Actually Love Social Media, Enshittification Isn’t Real

An a16z article argues that enshittification isn’t real and that people actually love social media.

Wonderful more than doubles its valuation to $5B in under 6 months

Wonderful said it will use its $550 million Series C funding to develop products faster, expand its FDE teams, and meet demand for its products.

India’s richest man now wants to turn aging computers into AI-ready PCs

Jio is betting it can turn an aging computer into an AI-ready PC for as little as about $11 for two months.

Texas Police Used AI to Write Report About Using Flock to Search for Woman Who Had Abortion

The incident shows how quickly police departments have come to use AI surveillance tools even in critically sensitive cases.

Google launches Gemini 3.8 Flash Cyber for partners in its new Fairwind Program and says Gemini 3.8 Flash beats Claude Opus 5 and GPT-5.6 Sol on some benchmarks (Google)

Google : Google launches Gemini 3.8 Flash Cyber for partners in its new Fairwind Program and says Gemini 3.8 Flash beats Claude Opus 5 and GPT-5.6 Sol on some benchmarks — Our newest Gemini models deliver next-generation intelligence for agentic workflows and cybersecurity.

Google is sending MrBeast into the wilderness, armed with AI

MrBeast will feature Gemini, Google Health, and the Fitbit Air in upcoming videos as part of a multi-year partnership with Google. The deal will kick off with a video featuring Jimmy "MrBeast" Donaldson turning to Gemini for wilderness survival advice: First up on September 5 is a new MrBeast video following Jimmy and his crew […]

Google releases Gemini 3.8 Flash, three weeks after Gemini 3.7 Flash, for an introductory price of $0.75/1M input and $3.75/1M output tokens until December 31 (Abner Li/9to5Google)

Abner Li / 9to5Google : Google releases Gemini 3.8 Flash, three weeks after Gemini 3.7 Flash, for an introductory price of $0.75/1M input and $3.75/1M output tokens until December 31 — After the last model release three weeks ago, Google today is rolling out Gemini 3.8 Flash. This marks the third Flash update in three months.

SteamdDB Joins Nexus Mods

Comments

Austin-based HiddenLayer, which makes security tools to protect AI models, agents, and workflows, raised a $100M Series B led by Delta-v Capital (Ram Iyer/TechCrunch)

Ram Iyer / TechCrunch : Austin-based HiddenLayer, which makes security tools to protect AI models, agents, and workflows, raised a $100M Series B led by Delta-v Capital — When AI security startup HiddenLayer raised its $50 million Series A three years ago, one of the big questions in the space was whether …

The Trump administration files a brief supporting OpenAI in its dispute with The New York Times, saying training LLMs on copyrighted works is generally fair use (Blake Brittain/Reuters)

Blake Brittain / Reuters : The Trump administration files a brief supporting OpenAI in its dispute with The New York Times, saying training LLMs on copyrighted works is generally fair use — The Trump administration has filed a brief supporting OpenAI in its dispute with the New York Times (NYT.N) …

Gemini 3.8 Flash and 3.8 Flash Cyber

Comments

The Republican Nominee for New York Governor Made a Creepy, AI-Generated Video of Mamdani and Hochul

The video makes Zohran Mamdani and Kathy Hochul look like they're hanging out in a prescription medication commercial.

HiddenLayer nabs $100M as enterprises rush to secure their AI deployments

Security companies are scrambling to build products that can monitor not just agents but also the tools and add-ons they use.

Adobe launches Adobe for Slack, making 70+ creative and productivity tools, including Firefly, Express, Photoshop, and Premiere, available via Slackbot prompts (Ivan Mehta/TechCrunch)

Ivan Mehta / TechCrunch : Adobe launches Adobe for Slack, making 70+ creative and productivity tools, including Firefly, Express, Photoshop, and Premiere, available via Slackbot prompts — Customers can now use Adobe's apps like Firefly, Adobe Express, Photoshop, Premiere, Acrobat, InDesign, Illustrator, Stock …

PSA: Amazon’s shopping AI can now tell you if that message is a scam

Amazon is adding a scam-detection feature to Alexa for Shopping that can verify whether suspicious emails, texts, and other messages actually came from the retailer.

A US federal judge rules that Google does not have to sell off its ad exchange and instead must make its ad tech tools work with those operated by rivals (Bloomberg)

Bloomberg : A US federal judge rules that Google does not have to sell off its ad exchange and instead must make its ad tech tools work with those operated by rivals — Alphabet Inc.'s Google doesn't have to sell off its advertising exchange and instead must make its ad tech tools work with those operated by rivals …

Google avoids a breakup of its ad tech business

Comments

2x Faster Qwen3.8-Flash + GLM-5.3-Flash MTP

Run Qwen3.8-Flash-Next and GLM-5.3-Flash up to 2x faster with MTP. MTP is enabled by default, you can still disable it. Also our new release includes 170+ training, chat, hardware, and performance improvements. Highlights Smoother model loading (less errors) across local servers and connected providers. Safer chat edits that preserve tool cards, reply details, and conversation branches. New local media APIs for video, audio, and MLX-served models. New audio support with new models, progress tracking including: MiniMax-Music3, Higgs, MOSS and more! Improved multi-GPU planning, memory fitting, and split-model training. Strengthened AMD/ROCm detection, installation, and GPU compatibility. Upgraded MCP, Deep Research, OAuth, and agent tool reliability. Qwen3.8-Flash + GLM-5.3-Flash Qwen and GLM now generate faster with MTP enabled by default. Use GLM tools across longer, multi-turn chats. Qwen automatically applies the recommended settings for thinking and non-thinking modes. Download Qwen3.8-Flash-Next and GLM-5.3-Flash . See the Qwen guide and GLM guide for recommended settings and available GGUFs. Faster MLX inference Fine-tune both large MoE models with text or images on Apple Sili

Conveo, which uses AI to conduct customer interviews in 15 languages and analyzes transcripts for market insights, raised a $50M Series A led by DST Global (Chris Metinko/Axios)

Chris Metinko / Axios : Conveo, which uses AI to conduct customer interviews in 15 languages and analyzes transcripts for market insights, raised a $50M Series A led by DST Global — Conveo, which helps big brands understand market trends, raised a $50 million Series A led by DST Global, Conveo CEO Dieter De Mesmaeker tells Axios Pro exclusively.

OpenAI accused of ‘aiding and abetting’ Tumbler Ridge mass shooting in dozens of new lawsuits

OpenAI and its CEO Sam Altman are facing 30 new lawsuits that accuse them of providing "substantial assistance and encouragement" to the suspect in Canada's Tumbler Ridge school shooting, as reported earlier by TechCrunch. The new wave of lawsuits was filed in a California federal court on Wednesday by the students, teachers, and the principal […]

Exit the Cave

Comments

Claude's new system prompt really doesn't want to reproduce song lyrics

Anthropic publish the system prompts for their Claude consumer applications ( Claude.ai and the Claude mobile apps - sadly not for Claude Cowork or Claude Code). I love that they do this, and that they share not just the current prompts but historic changes to their prompts as well. They used to keep all of the prompts on a single page, but when I checked today I noticed they had re-arranged those prompts into an index page and then a page per model - here's the page for Haiku 4.5 for example, which has the original prompt from October 15th 2025 and an updated prompt from January 18th 2026. A neat thing about Anthropic's platform.claude.com/docs site is that it's designed to be usable by LLMs. You can add .md to any page to get back the content as Markdown - here's the system prompt index page and the Markdown prompts for Fable 5.1 . TL;DR: this makes it really easy to diff the prompts. Don't reproduce song lyrics Don't draw copyrighted characters or logos Tweaks to Claude's answering style The missing end_conversation guidelines Recommended substance support sites Reliable cutoff date of June 2026 How I'm tracking these prompts Don't reproduce song lyrics Let's start with the most

Adobe acquires Indian market intelligence startup Rilo

This is Adobe's second acquisition out of India after Rephrase.ai in 2023

Cops Are Asking Axon to Make Their Cameras Look Different From Flock So People Don't Destroy Them

“Is there any talk to redesign the Outpost to not look exactly like the Flock camera — I think it will help agencies with the optics while we batten down the hatches,” one apparent cop asked during a now deleted Axon webinar.

WebLLM: high-performance in-browser LLM inference engine

Comments

GrapheneOS says Pixel 11 has MTE support after all

Comments

Three sites made 215,128 “best software” pages for AI. Perplexity cites them

Comments

NYC bans AI use for students until they reach high school

New York City mayor Zohran Mamdani has announced a new policy today that will ban younger schoolchildren from using AI in classrooms. The one-year moratorium, effective in the 2026-2027 school year, will impact about 600,000 public school students in 2-K through eighth grade and is being introduced alongside additional limits on digital devices and a […]

Wonderful, which helps enterprises deploy AI, raised a $550M Series C led by Insight at a $5B valuation; investors also paid $170M to buy shares from employees (Yuliya Chernova/Wall Street Journal)

Yuliya Chernova / Wall Street Journal : Wonderful, which helps enterprises deploy AI, raised a $550M Series C led by Insight at a $5B valuation; investors also paid $170M to buy shares from employees — Founded just last year, the startup already has 650 employees and is expanding its forward-deployed engineering teams around the world

2x Faster Qwen3.8-Flash + GLM-5.3-Flash MTP

Run Qwen3.8-Flash-Next and GLM-5.3-Flash up to 2x faster with MTP. MTP is enabled by default, you can still disable it. Also our new release includes 170+ training, chat, hardware, and performance improvements. Highlights Smoother model loading (less errors) across local servers and connected providers. Safer chat edits that preserve tool cards, reply details, and conversation branches. New local media APIs for video, audio, and MLX-served models. New audio support with new models, progress tracking including: MiniMax-Music3, Higgs, MOSS and more! Improved multi-GPU planning, memory fitting, and split-model training. Ctrl+F Search now works. Strengthened AMD/ROCm detection, installation, and GPU compatibility. Upgraded MCP, Deep Research, OAuth, and agent tool reliability. Qwen3.8-Flash + GLM-5.3-Flash Qwen and GLM now generate faster with MTP enabled by default. Use GLM tools across longer, multi-turn chats. Qwen automatically applies the recommended settings for thinking and non-thinking modes. Download Qwen3.8-Flash-Next and GLM-5.3-Flash . See the Qwen guide and GLM guide for recommended settings and available GGUFs. Faster MLX inference Fine-tune both large MoE models with tex

Real-Time Intelligence with IBM Time Series Models on Confluent

Poisson Disk Sampling

Comments

Six curl CVEs after OpenAI and Anthropic came back with zero

Comments

Biggest dark matter detector spots a single weird particle

Comments

A Note from LWN

Comments