Vannarot Roeung
Recent posts by Vannarot Roeung
CityRAG Turns a Street Photo Into a Navigable City Walkthrough Video
Researchers from Google, Cornell, and Stanford built CityRAG, an AI that generates navigable city walkthrough videos from a single street photo matched to real-world geography.
AnyRecon Turns Sparse Photos into Full 3D Scenes in 105 Seconds
New research builds explorable 3D scenes from sparse phone captures in 105 seconds, with code and model weights already available.
Meta AI Is Changing Your Ad Creative Without Asking
Meta's Advantage+ Creative tools have been modifying ad images and generating new visual variants without advertiser consent, raising brand safety and legal concerns for creators and businesses.
fal Opens GPT-Image-2 API: No Waitlist, 4K, Commercial
fal launched a no-waitlist enterprise API for GPT-Image-2 on April 21, 2026. Developers get immediate 4K access with two endpoints, pay-per-use pricing from $0.01/image.
YouTube AI Deepfake Detection Now Covers All Celebrities
YouTube expanded its AI Likeness Detection system to all celebrities and entertainment professionals on April 21, 2026. No YouTube channel required -- talent agencies can enroll entire client rosters.
Latitude Voyage: Build AI RPG Worlds With No Code
Latitude, makers of AI Dungeon, launched Voyage on April 21, 2026 -- an AI-native RPG platform where creators build game worlds with natural language and every NPC thinks independently.
Claude Cowork Adds Live Artifacts: Always-On Dashboards
Anthropic added Live Artifacts to Claude Cowork on April 20, 2026: AI dashboards and trackers that connect to your apps and files and auto-refresh with current data every time you open them.
Epic Launches Conversations AI NPCs for Fortnite Creators
Epic Games gives Fortnite creators a new Conversations tool that turns NPCs into AI-powered characters with unscripted voice dialogue, using Gemini 3.1 Flash-Lite and ElevenLabs.
Adobe, NVIDIA, WPP Tune Firefly to Brand IP at Scale
Adobe expanded its NVIDIA partnership at Adobe Summit on April 20, 2026 with Firefly Foundry custom-tuned models, 3D digital twins on Omniverse, and CX Enterprise Coworker agents. Launch partners include WPP, CAA, WME, and UTA.
TwelveLabs Launches Rodeo, an AI Video Co-Pilot for Editors
TwelveLabs launched Rodeo at NAB 2026 -- an AI creative co-pilot that lets video editors find, assemble, and cut footage with natural language, cutting workflows from days to minutes.
Deezer: 44% of New Music Uploads Are AI-Generated
Deezer reports 44% of new daily uploads are AI-generated, ~75,000 tracks per day, though AI consumption stays at 1-3% of streams and 85% are fraudulent.
Artlist Launches Studio Platform and Cinematic Image Model for Pro Creators
Artlist launched Artlist Studio on April 20, 2026, an AI production platform that gives creators shot-by-shot directorial control, alongside Artlist Original 1.0, its first cinematic image model.
Makko AI launches 2D game studio for creators
Makko opens its AI 2D game studio out of beta with Collections for style consistency and a Code Studio for playable prototypes. 4,500 beta users generated 40,000+ assets and 3,500 games.
Amazon Adds $25B to Anthropic, Locks 5GW of Compute
Amazon committed up to $25B more to Anthropic on April 20, 2026, plus 5 gigawatts of new Trainium2 and Trainium3 compute. Anthropic pledged $100B+ to AWS over the next decade.
Voicebox: Free Open-Source Voice Studio with 7 TTS Engines
Jamie Pine's Voicebox brings seven TTS engines and voice cloning to your local machine. Free, open source, and entirely offline.
Google AI Studio Adds Nano Banana Pro for Paying Subscribers
Google AI Studio is now giving Pro and Ultra subscribers access to Nano Banana Pro, the company's most capable image generation model, along with higher usage limits across Gemini models.
Quiver SVG Generation Launches in ComfyUI Partner Nodes
ComfyUI added Quiver Arrow 1.1 as a Partner Node on April 20, 2026, bringing structured SVG generation to open-source visual workflows for the first time.
Qwen3.6-Max-Preview: Alibaba's New AI Tops Coding Benchmarks
Alibaba released Qwen3.6-Max-Preview on April 20, 2026, topping six programming benchmarks including SWE-benchPro. Here is what it means for creators building AI workflows.
Apple Names John Ternus CEO as Tim Cook Steps Down
Tim Cook will step down as Apple CEO on September 1, 2026, with hardware chief John Ternus taking over. Here is what it means for creative tools and Apple Intelligence.
Kimi K2.6 Deep Dive: Inside the Open-Weight Model That Tops Every Coding Benchmark
Moonshot AI's Kimi K2.6 scores 58.6 on SWE-Bench Pro, placing first above GPT-5.4 and Claude Opus 4.6. Full architecture breakdown, benchmark analysis, and what the open-weight license actually allows.
TRELLIS.2 Image-to-3D Now Runs on Apple Silicon Without Nvidia
A community port of Microsoft TRELLIS.2 brings image-to-3D mesh generation to Apple Silicon. No Nvidia GPU required, GLB output ready for Blender and Unity.
OpenAI GPT-Image-2 Rolls Out to Plus and Pro Users
OpenAI has quietly begun rolling out GPT-Image-2 to paid ChatGPT subscribers -- no announcement, no blog post, just a significantly more capable image model showing up in Plus and Pro accounts.
OpenAI Sora Has Shut Down: The Best Alternatives
OpenAI shut down Sora: the app closed April 26, 2026 and the API on September 24. These are the video generators worth switching to.
xAI Launches Grok Voice API: $0.10/hr STT, $4.20/M TTS
xAI launched the Grok Voice Agent API April 18 with standalone speech-to-text and text-to-speech endpoints priced at $0.10 per hour batch STT and $4.20 per million characters TTS.