Innovation Profs - 10/29/2024

Your weekly guide to generative AI tools and news

Author

Innovation Profs
October 29, 2024

Generative AI News

Anthropic’s agentic Computer Use is giving people ‘superpowers’

Last week, we reported several new features of Anthropic’s LLM Claude, but nothing as significant as what dropped this past week, namely, Computer Use, which “allows Claude to work autonomously and use a computer essentially as a human does.” Tasks that Claude can perform with Computer Use include taking screenshots, opening applications, moving the cursor, typing text, and switching between multiple screens and tabs. The tool is currently only available in beta. Still, that hasn’t prevented some superlative responses: “Anthropic just released the most amazing AI technology I’ve ever used. I’m not kidding.”

Anthropic’s Claude Now Writes Code to Do Math

Another new development with Claude is the ability to write and run Javascript in order to analyze data and perform complex mathematical calculations. Unlike the above-discussed Computer Use, Claude’s new coding capability is not unparalleled, as tools like ChatGPT and Google Gemini have similar features.

Report: Google targeting December launch for Gemini 2.0

In December 2023, Google released Gemini 1.0. Now, Google is aiming to launch Gemini 2.0 roughly one year later. According to a report by The Verge, performance of these newer generation models has not improved as expected, but this appears to be a common trend across a range of large language models.

Quick Hits

A Mother Plans to Sue Character.AI After Her Son's Suicide

A 14-year-old boy died by suicide after a months-long obsession with a 'Game of Thrones' themed chatbot.

gizmodo.com/a-mother-plans-to-sue-character-ai-after-her-sons-suicide-2000515536

AI Detects Mysterious Detail Hidden in a Famous Raphael Masterpiece

Artificial intelligence (AI) can be trained to see details in images that escape the human eye.

www.sciencealert.com/ai-detects-mysterious-detail-hidden-in-a-famous-raphael-masterpiece

Claude AI Gets Bored During Coding Demonstration, Starts Perusing Photos of National Parks Instead

Anthropic's Claude AI is purportedly capable of using computers like a human. And apparently, that includes procrastinating.

futurism.com/the-byte/claude-ai-bored-demonstration

Siri’s ChatGPT integration is here with iOS 18.2 — here’s how it handles your data

This is what you need to know and what you can control

www.tomsguide.com/phones/iphones/siris-chatgpt-integration-is-here-with-ios-18-2-heres-how-it-handles-your-data

Liquid AI Is Redesigning the Neural Network

Inspired by microscopic worms, Liquid AI’s founders developed a more adaptive, less energy-hungry kind of neural network. Now the MIT spin-off is revealing several new ultraefficient models.

www.wired.com/story/liquid-ai-redesigning-neural-network

'Project Jarvis' leak highlights Google Gemini 2.0's superpower

Leak details Google's secretive Project Jarvis, which can leverage Gemini to perform several complex tasks

www.androidpolice.com/google-gemini-project-jarvis-ai-agent

Forget Photoshop — Midjourney just launched its upgraded image editor with a killer new feature

The next big feature from Midjourney is here

www.tomsguide.com/ai/forget-photoshop-midjourney-just-launched-its-upgraded-image-editor-with-a-killer-new-feature

Tool of the week: Ideogram

Not to be outdone by Midjourney (see item above) generative AI photo company Ideogram has released three new features that expand the role AI can play in photo editing and creation: Canvas, Magic Fill and Extend.

Canvas is a workspace that can be used to generate, edit, and organize images. Magic Fill lets users fix, remove, or replace any part of an image (inpainting). Extend lets you expand your image infinitely (outpainting).

Paid plans have access to all three tools for AI-generative images, and Ideogram Plus or Pro users can use them on images they upload.

AI-generated image of the week

This week we tested out whether our generative AI tools would create an image of Donald Trump with Kamala Harris.

Our prompt: Donald Trump and Kamala Harris holding their hands up in victory at a rally in Washington, DC

Tools that would not let us create the image:

  • Midjourney

  • Meta AI

  • Google Imagen

  • Microsoft CoPilot

  • DALL-E

Tools that would create an image

  • Stable Diffusion (top)

  • Ideogram (bottom) 

Prompt: Donald Trump and Kamala Harris holding their hands up in victory at a rally in Washington, DC

Get starting with Generative AI

New to generative AI? Here are some places to start…

What we found

ElevenLabs’ latest feature Voice Design lets you design a custom voice with just a text prompt. Describe the person you want to hear speak, and ElevenLabs does the rest. You can create realistic voices or character voices.

Examples they share on the website include:

  • “An old British male with a raspy, deep voice. Professional, relaxed and assertive.”

  • “An angry old pirate, shouting”