July 19, 2024

Windows Copilot Newsletter #33 - GPT-4o mini; Copilot reads handwriting; Google scans private files...

Windows Copilot Newsletter #33

OpenAI replaces GPT-3.5 with the faster, cheaper and smarter GPT-4o mini; Microsoft OneNote can now read your handwriting; Google caught out reading files into Gemini...

G'day and welcome to the thirty-third edition of the Windows Copilot Newsletter, where we collect, curate and share all of the latest news about AI chatbots. A flurry of announcements this week, so let's dive right in...

Top News

OpenAI launches 'GPT-4o mini': The free tier of ChatGPT used the nearly obsolete GPT-3.5 - until this week, when OpenAI released a faster, smarter and cheaper-to-run model based on its recently released multimodal GPT-4o. This 'GPT-4o mini' isn't quite as smart or capable as its bigger brother, but hey, it'll be free, and it will analyse images - which should lead to some unexpected applications...

OneNote gets an AI upgrade to read handwriting: Microsoft OneNote can now use Copilot to 'read' handwritten notes uploaded into it. It's a straightforward integration of the multimodal capabilities now common to AI chatbots - but it will also be a very useful feature. Read more here.

Gemini creates video presentations: It won't be putting Hollywood out of business, but Gemini has launched a new Workspace app - Vids - to generate video presentations. It's still in testing in Google's Workspace labs, but will be rolling out over the next few months. Read about it.

Spreadsheets get the LLM they need: We can't feed spreadsheets directly into any AI chatbots because, well, the chatbots simply don't know what to do with it. Microsoft Research unveiled 'SpreadsheetLLM', a set of tools designed to help chatbots ingest and operate on spreadsheet data. They didn't release any code, though, which means this innovation is likely years away from introduction.

Top Tips

How to get the iOS 18 beta: If you use an iDevice, you're likely keen to get the new public betas of iOS 18 - which will eventually support Apple's new 'Apple Intelligence' AI features. Instructions are here.

Turn off Copilot: Sick of Microsoft's AI assistant? Follow these instructions to shut it off.

Safely and Wisely

Google caught scanning files into Gemini: Is Google scanning user files into Gemini? Google says no, but a user says yes. Read the story, and make up your own mind.

YouTube will remove deepfakes of you: That is, if you ask nicely, and if you're the one doing the asking. That way it becomes a privacy violation - and they have policy to deal with that. Read more here.

Longreads

What is AI? Technology Review addresses the eternal question here.

Can ChatGPT write a beach novel? A writer for the New York Times took the chatbot to task - read the story here.

‘De-Risking AI’ white paper - now out

AI offers organisations powerful new capabilities to automate workflows, amplify productivity, and redefine business practices. These same tools open the door to risks that few organisations have encountered before.

Wisely AI’s latest white paper, ‘De-Risking AI’, lays a foundation for understanding and mitigating those risks. It's part of our core mission to "help organisations use AI safely and wisely". Read it here.

If you learned something useful in this newsletter, please consider forwarding it along to a colleague.

We'll be back next week with more of the most interesting stories about AI chatbots!

Mark Pesce
mark@safelyandwisely.ai // www.safelyandwisely.ai