Will plugin stores be the next app store?

Welcome to another edition of what we’re determined to make the best damn newsletter in AI. Here we’ll break down AI topics that matter, open your mind to use cases, and keep you ahead of the curve.

Our #1 goal is to be useful. So please shoot us an email 📩 if you have questions or feedback, and especially if you implement something we share!

Here's what we're covering today:

• What businesses should take away from Microsoft’s developer conference, Build

• Chaining together prompts has never been easier thanks to GPT for Sheets

• Is the music industry ready for the new kid on the block: AI?

Let’s get to it! 👇

TODAY'S PERSPECTIVE

AI Triples Down on Plugins

Microsoft’s developer conference, Build, was jam-packed with announcements this week.

• Bing is coming to ChatGPT (and will be available for free) very soon

• Windows is getting its own Copilot, integrated directly into the operating system (!!)

• Azure has a new AI Studio with tons of new features to make it easy to build and deploy custom AI applications (if you’re building in AI - definitely go take a look)

But one of the biggest announcements for businesses to know about is what Microsoft and OpenAI announced regarding plugins.

A Universal Framework for Large Language Model Plugins

Meaning that developers can build a plugin for ChatGPT, and it’ll work in Bing Chat, Microsoft 365 Copilot, and likely many many more places where we’ll see plugins pop up.

It’s clear these are about to be everywhere… We’ve even heard of a few larger startups considering opening up their platform to plugins.

Here’s why you should be paying attention 🔥🔥

Plugins are how large language models like ChatGPT get access to real-time data, have sources for their information, are able to do actions for you, and more.

If we were writing a hype post about it, it would sound something like “ChatGPT makes you 10x more productive, but plugins make you 100x!!”

The big question we have: will we have a universal plugin store?

Or will it be bifurcated, like the app store?

Guess the ball is in Google and Apple’s court on that one.

What does all this mean for you?

The plugin ecosystem is still extremely early. Truthfully, only a few of the current ChatGPT plugins are even worth using on a regular basis.

But to write plugins off based on the current state would be a mistake.

Start thinking about:

• How do you take your work and start adapting it to leverage these next gen tools, like Plugins?

• How should you adapt your offering as a business or build new products to take advantage of this next platform?

We’ll be covering this more as it develops.

USE CASE FROM OUR PARTNERS

Chain together your prompts with GPT for Sheets

Mega prompting, in our opinion, is overrated. Mega prompting for ChatGPT is what we call it when you try to use one giant prompt with tons of steps, context, and information jammed into it. In our experience, instead of leading to “10x outputs”, Mega prompts just lead to more variation in your outputs, are harder to debug, and truly feel like shooting in the dark with LLMs.

That’s why we’ve turned to Prompt Chaining using GPT for Sheets.

Prompt chaining is essentially breaking the task into smaller, interconnected prompts, which allow us to guide the model step by step, ensuring each part contributes to a coherent and accurate output.

And thanks to the GPT for Sheets add-on, prompt chaining becomes a breeze.

By combining GPT for Sheets’ capabilities with some Excel formula magic (which can also be generated with the help of GPT for Sheets), we can seamlessly incorporate the process of prompt chaining, directly into our workflows, without ever having to copy and paste within ChatGPT again.

Check out our free Google Sheets template using GPT for Sheets for prompt chaining here!

Excited about a specific use case with prompt chaining in GPT for Sheets? Respond if you’d want a video tutorial and your use case!

LINKS

For your reading list 📚

AI in the music industry…

• Producer and pop singer, Grimes, launched an AI tool called GrimesAI-1, which allows producers and songwriters to make it sound like she is singing their song - and she’s reviewing some of the generated music

• Spotify is developing AI technology to make host-read podcast ads that sound like real people

And if you're really nerdy...

• Microsoft CTO Kevin Scott open sourced his GPT-powered social media post generator for his podcast

• Google DeepMind’s AI language model, Flamingo, is now making descriptions for YouTube Short by analyzing initial video frames

That's all!

We'll see you again on Tuesday. Thoughts, feedback and questions are much appreciated - respond here or shoot us a note at [email protected].

Cheers,

🪄 The AI Exchange Team