Prompt engineering is dead?

AutoGPT and other AI agents; consider using a privacy-aware ChatGPT

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:

  • Why is everyone up in arms about AutoGPT?

  • If you're looking for a privacy-aware ChatGPT, you'll want to check out this use case deep dive

  • And go further down the AutoGPT rabbit hole with these links

... and if someone forwarded this email to you, thank them 😉, and subscribe here!

Let’s get to it! 👇


Is prompt engineering dead? Meet AutoGPT.

Sam Altman, the CEO of OpenAI has long hinted that prompt engineering is a short-term thing.

We think that’s an oversimplification, but innovations like AutoGPT do get us closer.

AutoGPT is a new open-source software program that can prompt itself to complete an objective. Yes, you read that right. It prompts itself.

Anyone can run and use it (with a little bit of coding knowledge). Here’s the github.

It works by asking a user for the role it should play, and what it should try to accomplish. Then it “thinks”, “reasons”, “creates a plan” and even “critiques” that plan before suggesting an action to the user for it to take.

AutoGPT isn’t alone. There are other AI agent systems emerging that connect ChatGPT to other machine learning models to help it complete different tasks and of course, ChatGPT plugins, which give ChatGPT the power to interact with other tools.

Big things are coming.

Premium subscribers, read our full chapter on AI Agents.


Privacy-aware ChatGPT

As we’ve covered, any data you put into ChatGPT the app is part of Open AI’s research preview and could be used in future model training. We recommend a simple heuristic… The Reddit Rule: if you wouldn’t put it anonymously on Reddit, don’t put it in ChatGPT.

But if you’re looking for a better solution, consider using a privacy-aware ChatGPT.

Damien Tanner, one of The AI Exchange members, runs a company that built its own version of ChatGPT specifically focused on team collaboration and keeping data private. It uses your own OpenAI API key, which means your inputs and outputs stay isolated to your account.

Check out 🙌


For your reading list 📚

If you want to go further down the AutoGPT rabbit hole...

And if you're really nerdy...

That's all!

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

... and if someone forwarded this email to you, thank them 😉, and subscribe here!


🪄 The AI Exchange Team