Amazon's new tools for building with Generative AI. Doodles to Animations. Autonomous agents and more.
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: Weekly News and Innovations at a Glance
AI Toolbox: Product Picks of the Week
AI Skillset: Learn & Build
🗞️ AI Pulse: Weekly News and Innovations at a Glance
Amazon announces:
Amazon Bedrock, a new service that makes foundation models (FMs) from AI21 Labs, Anthropic, Stability AI, and Amazon accessible via an API [Link]
Amazon’s new Titan FMs: The first is a generative LLM for tasks such as summarization, text generation, classification, open-ended Q&A, and information extraction. The second is an embeddings LLM that translates text inputs into numerical representations (known as embeddings) that contain the semantic meaning of the text [Link].
the general availability of Amazon CodeWhisperer, the AI coding companion, free for individual developers. It has built-in security scanning for finding and suggesting remediations for hard-to-detect vulnerabilities, such as those in the top ten Open Worldwide Application Security Project (OWASP), those that don’t meet crypto library best practices, and others. [Link].
Meta has released Animated Drawings - an open-source project that turns doodles into animations [Link]
Stability AI announced Stable Diffusion XL (SDXL) - the latest image generation model, now available through their API, excels at photorealism & adds many cool features like enhanced face generation, minimal prompts & legible text. SDXL also has functionality that extends beyond just text-to-image prompting, including image-to-image prompting (inputing one image to get variations of that image), inpainting (reconstructing missing parts of an image) and outpainting (constructing a seamless extension of an existing image) [Link].
Google introduced Med-PaLM 2, expert-level medical LLM that consistently performed at an “expert” doctor level on medical exam questions, scoring 85%. This is an 18% improvement from Med-PaLM’s previous performance and far surpasses similar AI models [Link].
Databricks announced Dolly 2.0 - the first open-source, instruction-following LLM (12B parameter) that’s available for commercial use [Link].
Poe, Quora's AI chatbot app, now features the ability for users to create custom bots using just prompts, with options such as Claude Instant or ChatGPT as a base. Quora plans to cover large language model fees, making it free for users at the moment [Link].
Zapier added new AI features in its ‘Interfaces’ no-code tool which lets users create interactive pages and app. Now, one can create customized ChatGPT-powered bots, embed them anywhere, and trigger automations based on chat responses [Link]
Demo projects from a ChatGPT hackathon, held last week and sponsored by OpenAI, Replit and others [Link].
CAMEL (Communicative Agents for “Mind” Exploration of LLM Society) - AI agents interacting with each other and collaborating. For e.g., two ChatGPT agents playing roles as a python programmer and a stock trader collaborating on developing a trading bot for stock market. [ Colab of the demo | Project website]
Open AI introduces ‘Consistency Models’ as an alternate to Diffusion based models (used by tools like Stable Diffusion, Midjourney etc.) that can generate a complete image in just one step. [Link to Paper | Link to TechCrunch article].
Stanford and Google researchers developed a virtual town populated by 25 ChatGPT agents to test machine learning models in creating realistic, adaptive generative agents simulating human behavior. In a Sims-inspired environment, agents store experiences, synthesize memories, and plan behavior in natural language. They engaged in complex actions such as organizing a Valentine's Day party, and their actions were rated as more human-like than humans roleplaying! [Demo Link | Link to Paper].
LangChain announced support for running LangChain.js in browsers, Cloudflare Workers, Vercel/Next.js, Deno, Supabase Edge Functions, alongside existing support for Node.js ESM and CJS [Link].
Artifact, the recently launched personalized news app from Instagram’s founders adds a social discussions feature [Link].
Open AI announced a bug bounty program with rewards ranging from $200 for low-severity findings to up to $20,000 for exceptional discoveries [Link].
Boston researchers have developed an AI tool called Sybil, which can detect early signs of lung cancer years before doctors would find it on a CT scan [Link]
Alibaba Cloud unveiled Tongyi Qianwen, a ChatGPT-like AI with bilingual capabilities, to be integrated into its business applications, including DingTalk and Tmall Genie [Link].
Hubspot introduced several improvements for its generative AI tool ChatSpot [Link]
🔍 🛠️ AI Toolbox: Product Picks of the Week
AgentGPT
AgentGPT brings Auto-GPT-like functionality directly to the browser (learn more about Auto-GPT in today's learning section). This tool allows you to set a goal for your AI agent, watch it develop an execution plan, and take action. It works by using models to generate a task list and then progressively performs tasks while evaluating if they are finished or require further sub-actions.
Leonardo.Ai
A game-changing AI tool for creating production-ready game assets. Users can opt for existing models or train their own AL models easily. With user-directed prompts like futuristic, medieval, fantasy, sci-fi, or synthwave etc., the tool generates variations that maintain a consistent finish. The process is iterative, enabling upscaling and refinement.
WisdomAI
WisdomAI is a GPT-powered chat tool for audio and video content. Simply import podcasts, YouTube channels, Zoom recordings, Facebook Lives, and more to ask questions and get insightful answers. Customized chatbot can be embedded in your website.
Chatbot UI
Chatbot UI is an open-source alternative to OpenAI's ChatGPT interface with added features. It enables search within conversations, chat import/export, prompt templates, and folders for organizing chats and prompts and more.
📕 AI Skillset: Learn & Build
A comprehensive list of Stable Diffusion tools and resources [Link].
Auto-GPT and BabyAGI: How ‘autonomous agents’ are bringing generative AI to the masses [Link]
An introduction and setup guide for the trending Auto-GPT project:
A tutorial on building a conversational retail shopping assistant, using Redis, LangChain, and OpenAI, to help customers find items of interest that are buried in a product catalog [Link].
Shaan Puri and Sam Parr chat with Hubspot's Co-founder and CTO, Dharmesh Shah, about AI, its myriad business opportunities, and why Dharmesh believes AI's impact will surpass that of the internet.
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 yet another jam-packed post. I'm actually thinking of including Meta's animation demo in my next post, as it's a stripped-down but beginner-friendly version that applies a few predetermined animations to any photo you upload. Really fun to play with: https://sketch.metademolab.com/canvas