ChatGPT for finance. Text-to-3D cinematic-quality digital humans. ChatGPT-like Language Model from Stability AI.
Greetings and Welcome to this week's AIBrews - your thoughtfully curated guide to the most promising AI products, learning resources and a concise roundup of the week's impactful news. The aim is to provide a balanced selection in the rapidly changing AI landscape, ensuring our readers stay informed without feeling overwhelmed. Please let us know how we can further optimize your experience and save you time.
Without further ado, let's dive in.
In today’s issue:
AI Pulse: News, Insights and Social Spotlight of the Week
AI Toolbox: Product Picks of the Week
AI Skillset: Learn & Build
🗞️ 🌏 AI Pulse: News, Insights and Social Spotlight of the Week
🔥 News & Insights
Stability AI released an open-source language model, StableLM that generates both code and text and is available in 3 billion and 7 billion parameters. The model is trained on a new dataset built on The Pile dataset, but three times larger with 1.5 trillion tokens. [Details | GitHub | HuggingFace Spaces].
Synthesis AI has developed a text-to-3D technology that generates realistic, cinematic-quality digital humans for gaming, virtual reality, film, 3D simulations, etc., using generative AI and visual effects pipelines [Details].
Nvidia presents Video Latent Diffusion Models (Video LDMs), for high-resolution text-to-video generation and having a total of 4.1B parameters [Details | video samples]
Adobe expands generative AI features of Firefly from images and text effects to video editing, audio, animation, and motion graphics design. [Details |Video].
OpenAI cofounder Greg Brockman on TED Talks: The Inside Story of ChatGPT’s Astonishing Potential [Link]
WebLLM: an open-source chatbot, built through collaboration between CMU, OctoML and SJTU, brings language models (LLMs) directly in web browsers. Can now run instruction fine-tuned LLaMA (Vicuna) models natively in browser via WebGPU with no server support [Details].
Raspberry Pi Foundation and DeepMind launched Experience AI: an educational program that provides teachers and students aged 11-14 with cutting-edge resources on artificial intelligence and machine learning [Details].
Atlassian launched ‘Atlassian Intelligence’ - an AI-driven ‘virtual teammate’ that combines their models with OpenAI's to create custom teamwork graphs showing the types of work being done and the relationship between them. It can create, summarise and extract information from content, automate support interactions right from within Slack and Microsoft Teams, generate insights using data from multiple sources in Atlassian Analytics and more [Details | Video]
Vercel introduced ‘AI Playground’, a tool to compare LLM prompt results from different providers like OpenAI and Anthropic [Detail]. Vercel also added a couple of new AI templates: AgentGPT with Langchain, Chatbot UI and more [Detail].
Chegg launched CheggMate, a GPT-4-based AI companion, offering tailored learning paths, custom quizzes, and guidance for students [Details].
Snap has made its AI chatbot, My AI, available to all users after initially launching it as a premium feature [Details].
Meta AI has developed and open-sourced DINOv2, a self-supervised computer vision model that doesn't require fine-tuning and is pre-trained on a dataset of 142 million images [Paper | Demo].
Google is working on a fresh AI-powered search engine and is simultaneously adding AI features to the current one under Project Magi [Details].
Microsoft is reportedly developing its own AI chips to train large language models, aiming to reduce dependency on Nvidia [Details].
Elon Musk plans to launch 'TruthGPT', a maximum truth-seeking AI that tries to understand the nature of the universe [Details].
🔦 Social Spotlight
A Mental Models iOS app built with the help of ChatGPT and launched on App Store in 3 weeks with zero prior coding experience [Link].
A dataset of every US Patent ever filed to be used in an AI system to advise on new patent ideas [Link].
HealthGPT, an open-source iOS app, that allows users to interact with their health data stored in the Apple Health app using natural language [Link].
AutoGPT has now 85+ stars on GitHub. A list of 5 tools that let you try AutoGPT in browser [Link].
🔍 🛠️ AI Toolbox: Product Picks of the Week
FinChat
FinChat is Chat GPT for Finance. It provides information on 750+ companies and 100+ super investors acting as powerful equity analyst assistant. It's trained on up to date financial data including years of transcripts, filings and reports from 750+ companies, 10,000+ company specific metrics, investing books, etc. FinChat is free to use for 10 prompts per day.
Doodle Tale
Doodle Tale is a delightful app for kids that brings their drawings to life with AI-narrated stories. Children can unleash their creativity by adding objects and characters to craft their unique coloring book pages. The AI then generates an engaging story based on their drawings. Offers stories in English, Spanish, and German with impressively realistic narrations. You can DM founder for free credits.
Klu
Klu is an internal search engine powered by AI to find, save, and organize information (from your gmail, google drive, notion, slack etc) in one centralized place. You can ask about your next meeting, flight, or launch date, and it will provide an answer based on your data.
📕 AI Skillset: Learn & Build
The Complete Beginners Guide To Autonomous Agents [Link]
25 Best Midjourney Prompts: Visual Guide [Link]
How to Build a Personalized PDF Chat Bot with Conversational Memory, powered by ChatGPT API, LangChainAI and Databutton [Link]
Blend Images with Reverse Engineering in Midjourney:
Thank you for reading AI Brews! If you have any thoughts, questions or just want to say hello, please don't hesitate to hit reply. Mariam
Thanks for another great round-up, Mariam. Doodle Tale sounds delightful, I'll try to give it a shot with my kids this weekend.
And I appreciate the mention in your Learn & Build section!