#多模态

AI UI/UX designer: assign a ticket, get senior-reviewed Figma designs in 24-48h

Greensighter's AI Designer takes tickets from Jira, Linear, Asana or ClickUp and returns Figma designs in 24-48 hours, built from the user's design system and reviewed by a human senior designer before delivery. It learns the product's design system, components, screens and codebase at setup, and draws on 600,000+ real screens and 150,000+ flows for research. Pricing starts at $1,490 per month, with onboarding in 2-3 days.

Tencent WeVisOne-VL Unifies Detection, OCR, Segmentation and Matting

Tencent's WeChat Vision team released WeVisOne-VL, an open vision language model in 2B and 4B sizes that handles detection, referring, counting, pointing, OCR, segmentation and alpha matting in one model, with pixel outputs from a 9.6M-parameter decoder so no separate SAM is needed. Built on Qwen3-VL, it is trained with up to 8,192 tokens of context and runs on a single GPU via Transformers or vLLM in BF16. Tencent reports the 4B model scoring 71.3 F1 on Dense200 and 81.8 F1 on RefCOCOg test, while noting a 52.6% rejection rate on PR-Bench.

Those Jeff Bezos hair transplant photos are very fake

Viral images of Jeff Bezos and Jason Statham with fuller hair circulating on X, Threads and Bluesky are photo edits, not real hair transplants. The Verge confirmed via reverse image search that the Bezos photo derives from 2023 pictures and the Statham image from a 2011 paparazzi shot, and none of the tested images carried Content Credentials or SynthID markers.

AI tool can predict cancer survival by a single patient selfie

Researchers at Mass General Brigham developed an AI system, trained on more than 40 million facial photographs, that estimates biological age and predicts survival outcomes for cancer patients from a single selfie, with details published in Lancet Digital Health. The two tools, FAHR-FaceAge and FAHR-FaceSurvival, rely on different facial regions, and among radiotherapy patients in the US those who looked five or more years older than their actual age had a 21% higher risk of dying during the average four-year follow-up, while those who looked five or more years younger had a 17% lower risk. Doctors cautioned that more research is needed before the tool could be widely used.

Redesigned Health app expands language support

Apple's redesigned Health app expands language support to global English, Chinese (Simplified and Traditional), French, Japanese, Korean, Norwegian, Portuguese, Spanish, Turkish, and Vietnamese. Built with Apple Intelligence, it adds a Longevity tab with a Health Age feature derived from Apple Watch metrics and at-home movement evaluations that run entirely on device. The app will be available on Apple Intelligence-enabled iPhone and iPad models with the latest software, with some features limited by region, language, or device.

I use AI (Oct 2026)

A research engineer at Salesforce AI Research reflects on how his AI use changed over 12 months, now managing roughly 5–6 coding agents per workday and spending about $3,000/month on tokens. He finds agentic coding fast and useful but still dependent on human taste, reports that LLMs have become his first choice for literature searches while remaining unreliable for novel research, and criticizes AI writing as mediocre and AI image and video generation as a net negative.

Introducing India-based inference for Kimi K3 on Amazon Bedrock

Amazon Bedrock now supports in-country inference in India for Moonshot AI's multi-modal, open-weight Kimi K3 model via the in.moonshotai.kimi-k3 geographic inference profile, processing requests in the Mumbai and Hyderabad Regions for workloads with strict data residency requirements. The post explains global versus India inference profiles and walks through an example of OCR and document understanding for Indic-language documents, including structured extraction and follow-up questions in code-mixed Hinglish. It also covers prompt caching, service tiers, and CloudWatch monitoring, citing Ultrahuman's report of 80 percent lower LLM costs and 85 percent faster multi-turn chats with prompt caching.

Palette: teaching a small model to build decks in your design language

Palette fine-tunes a 20B open-weight model with LoRA to build slide decks in two roles—a designer that writes a structured JSON brief and a coder that writes pptxgenjs code per slide—trained on roughly 1,000 synthetic decks (~10,000 slides) generated by a large open teacher and filtered by a checker. On PresentBench it ranks first on Economics and Education, third on Talk, and second across the three-domain mean (59.7 vs. NotebookLM's 60.8). A check-and-repair harness that keeps a rewrite only when the checker confirms improvement adds 3.5 points on average across 131 decks.

Cities Worldwide Build and Scale AI Services With NVIDIA Physical and Agentic AI

NVIDIA and its partners are expanding physical AI and AI agents to cities worldwide, using shared computing infrastructure so city departments and local developers can build services without each acquiring their own resources. Examples include 1MillionBot's resident-service agents in Spain, LiveX's conversational AI hosts in the San Francisco Bay Area, K2K's shared vision AI infrastructure in Wrocław, and ITMAX's SamurAI V2 vision agents in Malaysia targeting 17,000+ cameras by 2027. Digital twins and local AI factories also feature, such as Honeywell's Nemotron-powered plant cognition at Dangote's Lekki refinery and Cires Technologies' sovereign AI factory at Morocco's Tanger Med port.

Google takes Dreambeans to India in bid to bring AI-generated personal stories to the masses

Google launched its experimental Dreambeans app in India, its first market outside the U.S., with support for 13 local languages including Hindi, Punjabi, Tamil, Telugu and Bengali. The app draws on connected Google services such as Calendar, Gemini, Gmail, Photos, Search and YouTube to proactively generate personalized illustrated stories without requiring a prompt. Google said it chose India due to strong user interest and high overlap of users across Gemini, Gmail and Google Photos, though the app can only generate stories in one selected language at a time.

The latest AI news we announced in October 2026

Google recapped its October 2026 AI updates, led by the global rollout of the Gemini 4 Argon frontier reasoning model and the announcement of Gemini 4 Carbon, its strongest model yet for long-horizon engineering and cybersecurity defense, now in pre-release testing with the U.S. government and trusted cyber defenders via the Fairwind Program. The month also brought Genie 4 for real-time interactive world simulation with native audio, the Google Fitbit Edge fitness tracker, a global expansion of SynthID Detector, and a quantum computing breakthrough in which "magic state cultivation" on the Willow processor cut errors roughly 40-fold. Other items included AMIE results published in The Lancet, the Playground gaming platform, EmbeddingGemma 2, the first AI response from a Project Suncatcher satellite, and Research@ 2026 announcements.

Agentic engineering intelligence with Amazon Nova Multimodal Embeddings and Amazon Bedrock AgentCore

This post describes a multi-agent engineering assistant built on Amazon Bedrock AgentCore, Amazon Nova Multimodal Embeddings, and the open source Strands Agents SDK that assembles evidence for aerospace part substitution decisions. A planner routes queries to Catalog, Evaluation, and Knowledge specialists, with deterministic Python tolerance-rule scoring, Nova Multimodal Embeddings for multimodal retrieval and photo cosine similarity, and a supervisor that synthesizes results only when multiple specialists contribute. The engineer remains the decision-maker, and the system drafts justification memos and Integration Impact Assessments with audit trail entries aligned with AS9100 sections 8.4 and 8.5.6.

Introducing SFX v3 for developers

ElevenLabs released SFX v3, a new sound effects model in its Sound Effects API that follows multi-instruction prompts more closely than SFX v2 and handles vocals and music better, available by setting model_id to eleven_text_to_sound_v3 on the POST /v1/sound-generation endpoint. The model picks sound length from the prompt, with duration_seconds for specific timing and prompt_influence (default 0.3) to tighten prompt adherence, while looping still requires eleven_text_to_sound_v2 with loop: true. The company describes v3 as an incremental update that is less likely to clip on high-frequency sounds and recommends A/B testing common prompts across both models since they share an endpoint.

Spotify plans AI song remixes next year, costing up to $5 extra a month

TechSpot reports that Spotify plans to let users remix artists' songs using AI next year, with the feature costing up to $5 extra per month. The article is at techspot.com, and it is being discussed on Hacker News.

Adobe is testing how well AI agents can use editing tools like a Photoshop professional would

Adobe Research and UC San Diego researchers developed EditAgent, which pairs a general-purpose image editor with specialized tools and uses a standard language model to compare parallel candidates using verification data such as close-ups, change maps, and measured brightness and color statistics. In a 600-edit benchmark blind-rated by five professionals, the general-purpose editor Flux-2-Klein-9B failed to produce a usable result 42.7 percent of the time, and specialized tools solved roughly two-thirds of those failures. With verification records, all seven language models tested beat the best fine-tuned evaluation model, with Qwen3.8-27B making a usable pick 81.3 percent of the time versus 74.8 percent for EditScore-72B, though the agent needs about ten times as much compute as a single editor.

Google wants video AI to understand what people need as they work

Researchers from Google, Stanford, Oxford and several U.S. universities introduced StreamVibe, a video agent that decides for itself when to speak up during tasks like cooking or furniture assembly, handling vague requests, unspoken intentions and mid-task changes. Built on a benchmark of about 4,100 examples from over 183 hours of recordings, StreamVibe-VL reportedly improves response timing and quality by 10.7 percent over other models, while a separate memory system is designed to boost proactive interaction by another 12.3 percent. The paper nonetheless finds that as videos get longer, all tested models become much less likely to time responses correctly, with a key content metric dropping by up to 61.5 percent when models build on their own earlier responses.

Nace AI Open-Sources Drex Max v1: A 27B Decision Model That Runs Locally on Apple Silicon

Nace AI has released Drex Max v1, a dense 27B decision model with a 262,144-token window that returns option probabilities in one forward pass, as open weights under Nace.AI Open RAIL-M, shipping bf16 weights, a Q8_0 GGUF, an affine Q4 checkpoint and forks of llama.cpp and Ollama. It runs on NVIDIA GPUs, Apple Silicon Macs (24 GB+ for affine Q4, 48 GB for Q8_0) and CPU, and Nace reports 67.52 on the public Decision Index 0.4, ahead of Drex v2.0 at 61.25. Local Mac and CPU paths are text only for now.

UriScanAI – An AI-powered urine color analyzer and hydration tracker

UriScan AI is an app that uses your phone camera to scan urine color and instantly rate your hydration level with actionable feedback. It also lets you log daily water intake, review scan history, set reminders, and access detailed health reports through a one-time Pro purchase, with health data kept on your device. The app is described as being for general wellness only and not a medical device.

Google Deepmind creates a global map of forests with satellite data and AI

Google DeepMind and the World Resources Institute developed ForTy, an AI map that assigns the world's 2020 land area to six classes at 10-meter resolution, including primary forest, planted forest, plantations, and tree crops such as oil palm, rubber, and cocoa. The map is intended mainly to give companies a 2020 reference point for the EU Deforestation-free Products Regulation (EUDR), and experts found it correct in 90.2 percent of 8,190 reviewed sites, though it identified only about half of tree crops and planted forests and 60 percent of plantations. The map, classification layer, and test data are freely available under a CC BY 4.0 license, including in Google Earth Engine.

Hi3D Launches AI 3D Model with 2048³ Voxel Resolution

Hi3D has launched Hi3D V3.0, an AI 3D generation model built on its proprietary Sparc3D engine, which the company describes as the first commercially available version to reach 2048³ voxel resolution. The update targets geometric detail, structural accuracy and texture output, with texture resolution raised to a maximum of 8K paired with an in-house UV completion algorithm. V3.0 is built into Hi3D's broader browser-based platform, which includes image-to-3D conversion, AI texturing, model splitting with connectors, and conversion of textured models into color regions for multicolor 3D printing.

How to Ace Interviews and Meetings with AI Assistance

OTO AI is a desktop AI meeting assistant that listens to live conversations, detects questions, and generates real-time answers using an AI provider of your choice, supporting OpenAI, Anthropic Claude, Google Gemini and DeepSeek via your own API key. It works across Google Meet, Microsoft Teams, Zoom, Webex and other platforms, can use pre-loaded TXT, Markdown, DOCX and PDF documents as context, and can analyse your screen with compatible multimodal models. It is sold as a one-time purchase rather than a subscription, with users paying the AI provider directly for API usage.

AI microdrama edit of Neighbours worries actors and crew

Former Neighbours actors and crew have raised concerns about a Fremantle and RoseBerry Media project that uses AI to scan decades of episodes and re-edit old footage into short vertical micro-dramas for mobile viewers, starting with footage of Margot Robbie. The companies say the tool only identifies material from original broadcast footage and that human editors approve every stage, but actors including Matt Werkmeister and Sam Clark say they were not approached and are unsure of their rights under contracts signed before streaming and AI existed. Sound recordist Bruce Findlay compared the process to "cutting up a Picasso," and the MEAA union said workers should be consulted before their likeness or expertise is used.

Introducing SFX v3 in ElevenCreative and ElevenAPI

ElevenLabs launched SFX v3, its newest sound effects model, which handles prompts describing multiple sounds more reliably, generates music and vocals more convincingly, and produces cleaner, more natural audio than SFX v2 in benchmark testing. It is available now in ElevenCreative (Flows, the Sound Effects tab, Studio, the Music Editor, and the mobile app) and through ElevenAPI for developers.

Disney Files Patents for AI That Form Emotional Bonds Using Guest Memories

Disney Enterprises filed two patents on June 29, 2026 for emotionally responsive AI interactive characters that elicit a reminiscence from a guest, predict memory features using a trained machine learning model, and compare them against a character memory bank to generate a response, adding a new "hallucinated memory" node when no match meets a threshold. The second patent, from an overlapping team, uses the same memory-matching system but has the character generate and perform a song, poem, or dance based on a mood modifier derived from the matched memories. Disney has made no announcements about these patents, and the article notes the technology may never be used or could take years.

Accelerate agentic AI with NVIDIA Nemotron models on Amazon EKS and SageMaker HyperPod EKS

NVIDIA Nemotron 3.5 Super VL is a 120B-parameter mixture-of-experts model activating 12B parameters per token that handles text, image, document, and video input, joining other Nemotron models already on AWS. The post describes its hybrid Mamba-2, MoE, and attention architecture with a RADIO vision encoder and up to 1M-token context, then shows deployment on Amazon EKS, EKS Auto Mode, and SageMaker HyperPod EKS using NVIDIA Dynamo and vLLM, with validation and performance results on B200 and H200 instances.

Welcome Le Chonk: Mistral Large 4 on Hugging Face

Mistral Large 4 ("Le Chonk") is a multimodal, multilingual Mixture-of-Experts model with about 1T total and 49B active parameters and a 1M-token context window, released in FP8 and NVFP4 variants under the Mistral Ecosystem Open-Weight License. It combines a 72-layer text decoder using MLA with a DeepSeek Sparse Attention indexer and a 1.6B Pixtral vision encoder, and the post reports 82% on the Artificial Analysis Cyber Index vulnerability reproduce-and-patch test and 93% on Cybench. Day-0 support covers Transformers, vLLM, llama.cpp, TRL fine-tuning, Inference Endpoints and Dell Enterprise Hub, with GGUFs from Unsloth and the llama.cpp team.

Day 0 support for Mistral Large 4 on vLLM

vLLM announces day-0 support for Mistral Large 4, a 1T-parameter MoE multimodal model with 49B active parameters per token, running on the V2 model runner with breakable CUDA graphs and a new vllm/models/mistral/ package. The work adds cross-layer KV sharing, DCP/PCP context parallelism, DSpark speculative decoding, NIXL-based disaggregated serving and HiSparse KV offload, plus day-0 ROCm support on AMD Instinct MI350 Series GPUs with MXFP4 experts. Support ships in the vllm/vllm-openai:mistral-large-4 Docker image, with FP8 and NVFP4 checkpoints and serving commands documented.

Granite Switch 4.1.1 — Update models like you update software

Granite Switch 4.1.1 is a backwards-compatible patch release to Granite Switch 4.1 that adds an updated Granite Speech ASR model covering multilingual speech-to-text across 10 languages with native punctuation and capitalization, plus an updated Decision adapter in granitelib-core with improved performance on the Jev Decision Index test set. Apart from training the new ASR model and Decision adapter, nothing was recomputed and the base model is untouched. The team says it is working on tutorials and an Adapter SDK so developers and third-party providers can build their own adapters.

Building a Banana Sorter AI Machine

The author tests four vision models (Qwen2.5-VL 3B, Gemma 3 4B, Liquid d1-2B, and gpt-6.1-sol) on five banana images to decide BASKET or TRASH over 100 runs each. The smaller offline models trashed any banana with a black dot, while the frontier model sometimes trashed the heavily spotted banana (16% at low reasoning, 38% at high) and Liquid d1 gave probabilistic answers. The author concludes he would not trust a high-end model with a critical decision and that designing a real banana sorter is more nuanced.

Everything I know about AI regulation I learned from music videos

The author describes how 'slopcore'—AI-generated music videos made largely with code-based animation tools like Claude Opus 5.5—emerged from the 2024 song 'I'm Upping My P(doom)' and grew into a scene with its own chart, SlopHits. He curates videos that address AI safety, governance and regulation, including 'AI is a normal technology?', anabology's Claudia trilogy, and ClaudiaGI's French-language 'C'est de l'IA'. He notes the irony that a genre about AI regulation is largely compliant with the AI Act's Article 50 transparency labelling rules.

Come Back and Back AI

Come Back and Back AI is a forthcoming AI dance video generator for the TikTok "come back and back" trend set to "Wyw My Sweetie" by nissi & 25Hzz, which will turn one photo per dancer into a 10-second MP4 in Solo, Duo or Group modes and four looks. Generation has not opened yet; users can join a waitlist, and pricing is listed at $9.90 for one video down to $3.95 each in a 20-video pack. The page also covers the trend's history, the song, the seven-move choreography, photo tips and AI-labeling guidance.

Prime's first AI-generated show is here, and it's a Vaseline-smeared mess

Prime Video has premiered Castle Walls, a two-episode, 17-minutes-each series set in the 11th century around Alamut Castle, produced by Turkish studio Ay Yapım using generative video tools and billed as the first end-to-end AI-created series on a major streaming platform. The reviewer describes its visuals as looking like Vaseline-smeared Madame Tussauds creations melting in the sun, with overloaded cinematic imagery and banal dialogue, though the production used human voice actors and human screenwriters Kerem Deren and Çisil Hazal Tenim.

Meta's Muse Realtime Avatar turns any image into a talking AI character

Meta introduced Muse Realtime Avatar, an audio-driven Diffusion Transformer that turns reference images into expressive, interactive avatars with synchronized facial, hand, and full-body movement during live conversation. It shares Muse Realtime Voice's speech-token stream for synchronization, uses distillation to cut inference from 120 to 2 neural function evaluations per chunk (a 60x reduction), and streams 448x768 portrait video at 25 fps with roughly 870 ms latency. Meta says raters preferred it over Runway Characters and HeyGen LiveAvatar overall and across every evaluated dimension, and it embeds Meta Video Seal watermarks without added latency.

AI Can Draw Blueprints Now

An architect tested 13 older and newly released AI models on producing construction document sets from concept sketches, finding that Anthropic's Claude Opus 5.5, released September 22, 2026, could produce permit-quality drawing sets for a simple house with skilled guidance and redlining. The models generated vector linework by writing Python code rather than using CAD software, with Opus 5.5 taking twelve hours and a few hundred dollars in compute. The author argues the models' rapid improvement signals a shift for architecture akin to what software engineers experienced, with implications for training, fees and documentation demand.

Meet Google DeepMind’s Genie 4: Real-Time World Model With Native Audio

Google DeepMind released Genie 4 as a research preview for trusted testers; it generates explorable worlds from text and image prompts in real time, along with ambient audio, music and character dialogue. A Gemini model runs inside the simulation loop as a planner, using Gemini 3.8 Flash to break goals into steps, verify frames and replan, while Gemini Omni anchors each world's look and behavior. DeepMind disclosed no parameters, resolution, frame rate or benchmarks, and lists drift in longer sessions and interaction limited to a few minutes as open problems.

Featured coverage of this story: 《Genie 4: A new generation of world simulation》

Show HN: OpenMontage Drama Studio – self-hosted AI studio for short dramas

OpenMontage Drama Studio is a self-hostable web studio that turns a one-line idea into a short drama, viral short or product clip through a seven-stage pipeline: script, characters, scenes, props, storyboard and shots, voice, and edit. It requires users to supply their own text, image and video model keys (e.g. Volcengine Ark for Seedream and Seedance, or a ComfyUI server for images), with edge-tts as the default TTS, and ships a Docker Compose setup bundling PostgreSQL, Redis and a media worker while requiring an external S3-compatible endpoint. It is an independent AGPL-3.0 product by @charliewwdev derived from the calesthio/OpenMontage codebase.

Liquid AI Releases LFM3: 6 Open-Weight 2B-Class Multimodal Models

Liquid AI released LFM3, a family of six open-weight 2B-class models built on one shared tri-hybrid backbone covering text, vision, audio, omni and decision tasks, each fine-tunable on a single GPU. LFM3-1.7B has 1.73B parameters and a 256K context window, runs at about 1 GB of memory at 4-bit, and averages 49.37 across 29 benchmarks (Liquid's own eval), above Qwen3.5-9B at 47.24, while trailing it on GPQA Diamond (63.37 vs 80.40) and SWE-bench Verified (43.40 vs 51.20). Weights are on Hugging Face, the Liquid AI Playground and the Liquid API, with GGUF, MLX, ONNX, FP8 and NVFP4 checkpoints, and free commercial use stops at $10M annual revenue under the LFM Open License v1.0.

Show HN: Dola AI – AI images, video, creative and office tools

Dola AI is a browser-based creative studio that combines 2D image compositing, canvas editing, 3D scene blocking, a video timeline and reference boards, letting artists arrange compositions before submitting AI generation jobs. Manual editing and project saving use no generation credits, while AI image, video, audio, object and world generation (including World Labs Marble and Fal integrations) consume credits shown as a quote before submission. The service is built on authorized upstream open-source creative tools with its own accounts and private storage.

Who needs AI psychosis? I gave myself VR psychosis

The author describes building and auto-deploying web-based VR games by voice-prompting coding agents from the pub, using tools like Wispr Flow, Claude's /remote-control, Git, Vercel and Cloudflare, with games playable via Three.js and WebXR on a Quest 2. He walks through several prototypes, especially Star Crusader, and argues that iterating on primary gameplay loops is the fastest way to develop taste. He also recounts a scaled-back drawing-and-deploying game activity with his daughter.

Google Deepmind's Genie 4 puts a modern spin on agent designs that predate LLMs

Google Deepmind introduced Genie 4, a real-time world model that renders scenes frame by frame in response to keyboard and text-prompt input, generates its own background audio, and simulates talking characters. Gemini models plan the simulation in the background to keep it coherent, with Gemini 3.8 Flash checking multistep goals and Gemini 4 Argon tested in a demo where Genie 4 simulates Apptronik's Apollo robot playing checkers. Deepmind calls Genie 4 an early experimental model that stays consistent over short sessions but can drift in longer ones; it is available to select testers as a research preview, with SynthID watermarking on all output.