AI in week 34.
37 new tools and models from 6 analysed videos; 13 confirmed available in the EU.
Videos analysed
New AI waifus, new Deepseek, realtime worlds, Happy Shrimp, tiny TTS: AI NEWSAIAI Search· 23 Aug 2026 · 19 tools
AI News: OpenAI Just Put the Brakes on AIMAMatt Wolfe· 21 Aug 2026 · 19 tools
11 Grok Bot Use Cases That Feel Like CheatingMAMatthew Berman· 20 Aug 2026 · 5 tools
The FASTEST local AI video generatorAIAI Search· 18 Aug 2026 · 5 tools
You NEED to try these 6 Open-Source Projects NOWMAMatthew Berman· 17 Aug 2026 · 6 tools
New #1 open source AI has reached FRONTIERAIAI Search· 17 Aug 2026 · 5 tools
Language Models
All Language Models →DeepSeek V4 Flash Vision Experimental
Model · DeepSeek
Experimental vision-enabled version of DeepSeek V4 Flash with image understanding; open weights under MIT, now retired on the API.
Ornith 1.5
Model
Open model family (9B dense, 35B MoE, 397B MoE) trained with a self-improvement loop for reasoning, coding and agentic tasks.
ChatGPT for Teens
App · OpenAI
OpenAI's age-appropriate ChatGPT experience for users aged 13–17, with learning tools and parental controls.
Google AI plan free for students
App · Google
Google promotion giving eligible college students a year of the Google AI plan at no cost, with Gemini study tools.
Image Generation
All Image Generation →Video Generation
All Video Generation →Bernini v2
Model · ByteDance
ByteDance's open-weights video generation and editing model pairing a Qwen2.5-VL planner with a Wan2.2 renderer.
Qwen Video Edit
Model · Alibaba
Research method for instruction-based video editing by repurposing Qwen-Image-Edit with Wan 2.1 video latents.
LTX Director
Library · WhatDreamsCost
ComfyUI custom node that works as a timeline video editor for LTX video generation, with keyframes, audio and IC-LoRA support.
LTX 2.5 Upscaler with MiniMax
App
Community ComfyUI workflow that generates video with MiniMax and upscales it using the LTX 2.5 spatial upscaler.
Voice & Audio
All Voice & Audio →Music
All Music →3D & Worlds
All 3D & Worlds →Evoke
Model
14B autoregressive world model that generates interactive, camera-controlled video worlds in 3 denoising steps.
4DAnyone
Model
Research method that reconstructs a 4D Gaussian Splatting human from a single casual monocular video.
GeoWeaver
Research
Research framework for globally consistent 3D reconstruction from long RGB videos using chunked priors and test-time adaptation.
Modly
App
Open-source desktop app that generates 3D meshes from images using AI models running locally on your GPU.
Agents & Automation
All Agents & Automation →AVO
Library · NVIDIA
NVIDIA agentic harness reported to lift Claude Opus 5 from 30% to 100% on the public ARC-AGI-3 set.
Diagram Design
Library
MIT-licensed agent skill that generates brand-matched diagrams as a single self-contained HTML + SVG file.
Ego Lite
App
Open-source desktop browser that lets AI agents like Claude Code or Codex automate tasks in isolated Spaces using your logins.
Coding
All Coding →Slack Code
App · Salesforce
Salesforce's Slack feature for dedicated code channels where teams and AI coding agents build and review code together.
Cursor Origin
App · Cursor
Cursor's built-in code hosting with repos, pull requests, code browsing and two-way GitHub sync, with agents in every repo.
Open Vone
App · Z.ai
Free Hugging Face Space from Z.ai that runs GLM-powered vulnerability scans on submitted open-source GitHub repositories.
Productivity
All Productivity →Fathom
App · Fathom
AI meeting notetaker that records, transcribes and summarizes Zoom, Google Meet and Teams calls, with CRM sync.
Meta AI Mac app
App · Meta
Desktop Mac app for Meta AI with scheduled tasks, artifacts, media generation and chat history.
Claude Gmail and Drive integration
App · Anthropic
Anthropic's Claude connectors for Gmail and Google Drive, letting Claude read, draft and act on mail and files with user approval.
ChatGPT Apple Messages plugin
App · OpenAI
Plugin for the ChatGPT Mac app that lets ChatGPT search, summarize and draft or send replies to Apple Messages.
Robotics
All Robotics →Unitree Superman
Hardware · Unitree
Unitree's reported humanoid robot with a ~2 m standing jump and 12.66 m/s top speed, shown in previews.
Elf Schwan 2.0
Hardware · Headform
Elf-themed bionic humanoid robot from Headform with an articulated body, shown at the World Robot Conference.
Gen 1.5
Model · Generalist AI
Robot foundation model from Generalist AI that learns new tasks in-context from a short demonstration.
PX Cap Pro
Hardware · Paxini Tech
Tactile data-collection glove from Paxini Tech for recording human hand data to train robot manipulation.
Adapt
Research
Research system that transfers real tennis players' motions onto a Unitree G1 humanoid for rallies and serves.
MX01
Hardware · Arc Shell Robotics
Transformable robot from Arc Shell Robotics that switches between biped, quadruped and drone-carried forms.
Qiji
Hardware · DAX AI
Rideable quadruped robot horse from DAX AI for rough terrain, reported at 300 kg payload and 40 km range.
Research
All Research →Platforms & Infra
All Platforms & Infra →LM Studio
App · LM Studio
Desktop app for downloading and running local LLMs, now with Bionic, an agent for work and code.
LTX 2.5 GGUF
Model
Community GGUF-quantized builds of the LTX 2.5 video model for running on low-VRAM GPUs, e.g. in ComfyUI.
Unsloth
App · Unsloth AI
Open-source desktop and web app for running, fine-tuning and serving open models locally, with agent and search tools.
slime
Library · Z.ai
Open-source LLM RL post-training framework connecting Megatron training with SGLang rollout, used for Z.ai's GLM models.