PrO_RaZe 2.0

Meta: the company’s next 4 years of AR and VR hardware plans (till 2027) - PrO_RaZe Bookmarks #661

February 2023 AI Progress Updates (All Weeks)

Google Research: Monocular Depth Estimation using Diffusion Models - PrO_RaZe Bookmarks #659  

I think I predicted Elon Musk's new AI research lab's name lol - BasedAI - PrO_RaZe Bookmarks #658

Meta AI: Introducing CACTI — a framework for scalable multi-task multi-scene imitation learning - PrO_RaZe Bookmarks #657

Beating OpenAI CLIP with 100x less data and compute - PrO_RaZe Bookmarks #656

Talking About AI & Tech progress. Making Predictions (28 February 2023 Update)  Man I guess I will have to do updates like this every other week now, at least once a month. I mean things are progressing incredibly fast now. February is almost over here in India. And it's really crazy , 2023 so far already made 2022 look like a sleepy year. This morning I woke up with the news that Elon Musk is recruiting Top ML engineers and is going to start a new AI research lab to compete with OpenAI. This is what I wanted for real, I was literally wishing for it to happen weeks ago. My twitter feed is proof. If I have to guess, he may announce it officially with the company name on 1 March 2023 which is Tesla's Investor Day. Or maybe later. But it is great.  Xiaomi showed their new AR glasses prototype, things are improving in the VR/AR field too. It is just that it needs a lot of improvement to be mass adopted and manufactured.  I have hope that Apple's VR/AR glasses ...

Introducing Xiaomi Wireless AR Glass Discovery Edition - PrO_RaZe Bookmarks #655

Bing AI added to the Windows 11 taskbar in the latest OS update - PrO_RaZe Bookmarks #654

Introducing Typeface AI – a generative AI app to supercharge personalized content creation for businesses - PrO_RaZe Bookmarks #653

ELITE: Encoding Visual Concepts into Textual Embeddings for Customized Text-to-Image Generation (Similar to Dreambooth but trains in 0.05 seconds) - PrO_RaZe Bookmarks #652

Mark Zuckerberg has officially announced a team dedicated to building Generative AI for WhatsApp, Messenger and Instagram - PrO_RaZe Bookmarks #651

Microsoft: KOSMOS-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zero-shot) - PrO_RaZe Bookmarks #650

Elon Musk is starting a new AI company to fight OpenAI & Others - PrO_RaZe Bookmarks #649

Snapchat is introducing a chatbot powered by the latest version of OpenAI’s ChatGPT - PrO_RaZe Bookmarks #648

Chinese technology giant Tencent has revealed plans to develop its own large language model chatbot, “Hunyuan.” - PrO_RaZe Bookmarks #647

Check Your Facts and Try Again: Improving Large Language Models with External Knowledge and Automated Feedback (by Microsoft Research) - PrO_RaZe Bookmarks #646

You can now transform text into 360-degree worlds using a tool from Blockade Labs - PrO_RaZe Bookmarks #645

Composer is a large (5 billion parameters) controllable diffusion model trained on billions of (text, image) pairs - PrO_RaZe Bookmarks #644

Google AI: Spotlight, a vision-only model that achieves general user interface (UI) understanding from raw pixels - PrO_RaZe Bookmarks #643

OpenAI's Blog: Planning for AGI and beyond - PrO_RaZe Bookmarks #642

Meta AI announces a new SOTA large language model named LLaMA - PrO_RaZe Bookmarks #641

Nvidia predicts Al models one million times more powerful than ChatGPT within 10 years - PrO_RaZe Bookmarks #640

Introducing Raycast AI: The magic of AI, right on your Mac (Write smarter, code faster and answer questions quicker with ChatGPT in Raycast) - PrO_RaZe Bookmarks #639

Teaching CLIP to Count to Ten (by Google Research & More) - PrO_RaZe Bookmarks #638

Stable Diffusion running on an Android phone through full-stack AI optimization (by Qualcomm AI Research) - PrO_RaZe Bookmarks #637

Designing an Encoder for Fast Personalization of Text-to-Image Models (basically fast Dreambooth, only 5-15 steps to train using a single image) - PrO_RaZe Bookmarks #636

Google AI: Aligning Text-to-Image Models using Human Feedback - PrO_RaZe Bookmarks #635

MERF: Memory-Efficient Radiance Fields for Real-time View Synthesis in Unbounded Scenes (MERF that achieves real-time rendering of large-scale scenes in a browser) - PrO_RaZe Bookmarks #634

Luma AI introduces full volumetric photorealistic NeRF rendering on the web in realtime - PrO_RaZe Bookmarks #633

Google AI: Pre-training generalist agents using offline reinforcement learning - PrO_RaZe Bookmarks #632

MimicPlay - an imitation learning algorithm that uses "cheap human play data" (teaches robots to perform long-horizon tasks efficiently and robustly) - PrO_RaZe Bookmarks #631

Runway's Gen-1: The Next Step Forward for Generative Al (AI Videos) - PrO_RaZe Bookmarks #630

RoboNinja: Learning an Adaptive Cutting Policy for Multi-Material Objects (Robots using Knives) - PrO_RaZe Bookmarks #629

Clone announces that they are making synthetic humans, starting with a 1:1 copy of the human hand - PrO_RaZe Bookmarks #628

Your memories - how you remember them (Step inside your memories - on your phone, in VR, and in AR) - Wist: Immersive Memories - PrO_RaZe Bookmarks #627

CarperAI showcases how users can RLHF their own assistants with trlX using AnthropicAI's Helpful & Harmless (HH) dataset - PrO_RaZe Bookmarks #626

Uizard AI Autodesigner (basically ChatGPT for product design) - PrO_RaZe Bookmarks #625

ROSIE: Scaling RObot Learning with Semantically Imagined Experience (Text-to-image generative models, meet robotics) - PrO_RaZe Bookmarks #624

Runway announces the official launch of Runway Studios (for the next generation of AI storytellers) - PrO_RaZe Bookmarks #623  

Stanford Human Preferences Dataset (SHP) released: a collection of 385K *naturally occurring* *collective* human preferences over text (for training RLHF models) - PrO_RaZe Bookmarks #622

Google Research: Suppressing quantum errors by scaling a surface code logical qubit (towards scalable fault-tolerant quantum computing) - PrO_RaZe Bookmarks #621

Notion AI is now available to everyone (no waitlist or limited preview) - PrO_RaZe Bookmarks #620

T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffusion Models - PrO_RaZe Bookmarks #619

Spotify launches ‘DJ,’ a new feature offering personalized music with AI-powered commentary (using OpenAI's tech)  - PrO_RaZe Bookmarks #618

Amazon Web Services (AWS) is partnering with Hugging Face - PrO_RaZe Bookmarks #617

AGI within the next 5 years? From Conjecture AI - PrO_RaZe Bookmarks #616