We Watched Google's AI Conference So You Don't Have To - Promptlib #5
Also this week: Anthropic's 100,000 token limit, Meta brings AI tools to advertisers, and new releases in text-to-animation and text-to-3D.
On Thursday, during Google's annual I/O conference, the company announced a wave of new services fueled by artificial intelligence.
Image Credit: Bloomberg- Google CEO Sundar Pichai at the I/O Conference
With the public unveiling of Google Bard and the announcement of 20+ additional AI-driven apps and services, Google's vision of an AI-centric future is fast approaching.
But given the company's recent attempts to hop on the AI bandwagon, you might be wondering: is this just another publicity stunt?
To find out, we summarized every intriguing service that Google plans to power with its second-generation AI model, PALM 2.
In this edition
News & Insights: Which of Google’s generative AI services are actually worth trying?
Creative prompt: This is how to create consistent characters using Midjourney.
AI tool insights: Open AI released a text-to-3D model, text-to-animation is now compatible with Stable diffusion models, and Meta is bringing generative AI to advertising.
AI-Powered Tools 🛠
Before we get into the Google conference, here’s a quick rundown of all the game-changing generative AI releases we’ve seen this week.
Stability AI's new Stable Animation SDK, a versatile text-to-animation tool compatible with all Stable Diffusion models, promises to transform the animation and content creation landscapes.
Meta's AI Sandbox introduces generative AI tools for ad creation, featuring AI-driven text variations, automatic background generation, and optimized image cropping for Facebook and Instagram.
Image credit: Meta
Anthropic, an AI startup, has majorly upgraded Claude, their AI model, to handle 100,000 tokens from the previous 9,000 - a significant leap that outpaces OpenAI’s GPT-4. Currently, this feature is exclusive to those with API access.
News & Insights 🗞
Google Bard Releases to the public now powered by an improved language model— PaLM 2.
Yesterday's keynote revealed Google will drop the waitlist requirements for Bard, making it publicly available in 180 countries. Powered by a brand new second-generation language model, PaLM 2, Google Bard promises improved writing, coding, and debugging capabilities.
Image Credit: Google
Now the real question is: Can Google’s Bard hold a candle to the quality of ChatGPT? Test it yourself to find out.
Google’s Project Tailwind
The most exciting AI announcement was Project Tailwind. This tool, essentially an “AI-powered notebook”, can train a chatbot based on your personal documents and notes in Google Docs. While it's primarily marketed for students to streamline note-taking, its applicability extends much further. From organizing research data to enhancing workplace documentation.
Image credit: Google
It’s currently in beta and only available to select users, sign up for the waitlist here.
AI-Driven Google Search:
In a move with far-reaching implications, Google unveiled plans to revamp its Search results. Instead of the traditional list of blue links, Google is testing a search format that presents a more specific AI-generated text with a limited number of links.
Image Credit: An example of the new search results put forward by Google
Moreover, Google aims to use generative AI to curate shopping directly on the search page.
How Google will rank these AI responses remains unclear, but there's growing concern that this could jeopardize SEO traffic. If Google scrapes websites for information to provide AI-generated answers without directing traffic to the original sites, it could spell trouble for many— this article goes into depth on the issue.
Generative AI in Google Workspace
Workspace will now allow users to:
Draft, reply, summarize, and prioritize emails in Gmail.
Brainstorm, proofread, write, and rewrite documents in Docs.
Auto-generate images, audio, and video in Slides.
Auto-complete, generate formulas and categorize contextually in Sheets.
Generate new backgrounds and capture notes in Meet.
Enable efficient workflows in Chat.
AI assistant demo in Google Docs: Text & Formatting
Image credit: Google
AI assistant demo in Slides: Text & Image
Image Credit: Google
Further Reading 📖
An experimental AI tool that can turn text into music by Google called Music LM.
Google also unveiled Med-palm 2, a medical model fine-tuned to answer industry-specific questions.
Google will release a tool that’ll be able to differentiate AI-generated images in search results.
There are rising concerns that Google's yet-to-be-clarified ranking of AI responses could threaten SEO traffic, as AI-generated answers might bypass original sites.
Google is integrating AI into Google photos with Magic Editor - where users will be able to erase objects, color correct, and more using AI.
This is every single announcement google made at the keynote conference, summarized.
Community Updates 👩🎨
This week, I tried to get the hang of crafting consistent characters with Midjourney version 5 and found it a lot more difficult than I expected.
After trying outa bunch of different techniques, I wrote an easy tutorial sharing the tips and techniques that were useful when I created this set of consistent characters using Midjourney.
Thanks for reading!
We’ll see you next week with another Promptlib Newsletter.
For daily prompt templates & inspiration, follow us on Facebook or Instagram, and check out www.promptlib.com for a community feed of AI artwork, in-depth tutorials, and prompt engineering education.