Welcome to another edition of the best damn newsletter in AI.

As a reminder, we love Reader Questions. Our mission is to make using AI accessible to every business, and a huge part of that is making sure you’re not stuck and that we continue answering diverse questions. So, submit your question here. We’ll review every single one and if we believe i’ll be helpful, we’ll cover answers here in the newsletter.

Here's what we're covering today:

• The funny thing about AI agents

• What’s in the news

Let’s get to it! 👇

TODAY’S PERSPECTIVE

The funny thing about AI agents…

If you’re subscribed to the AI Exchange, chances are — you’ve heard about AI agents.

They’re all the rage right now. But in case you haven’t, they are basically AIs that can reason, make decisions and take actions… autonomously.

One of the most popular AI agents to be released recently was Anthropic’s “computer use” capabilities within Claude — Read the Announcement Here.

Basically, you can now tell Claude to go search stuff on a web browser, download all of the information, open up excel, put it in a new excel spreadsheet, write data analysis code to analyze it, and run the program…

Yes, you read that right.

But how good is the tech really? That’s what we’re going to talk about today.

The potential: We describe what we want done, AI does it

AI agents are incredibly sexy for this exact reason.

It’s crazy how much time is spent on mundane, repetitive work that if we could just describe what needs to be done - we could get it off our plates.

As AI gets to be more powerful (controlling browsers, or even computers), the possibilities of the types of tasks you can delegate explodes… while the difficulty of delegating those tasks goes to zero.

This is coming. And hot.

The limitation: With great power comes… distraction?

This all sounds great… so why aren’t we all using it.

Well, the major problem right now with agents is how autonomous do we want them to be really?

If we need to define every single step, some of the magic is lost.

But if we give it an open-ended goal, currently - the AI often loses track and rabbit holes, never even completing the task.

Some even describe AI’s current capabilities in completing long-term tasks like the AI has learned to have ADHD. Maybe AI just needs adderall…

Regardless —

AI agents are coming, and it’s worth paying attention if you aren’t already.

Our philosophy is always: limitations today are some super smart AI engineering team’s #1 priority… and will be gone tomorrow.

LINKS

For your reading list 📚

• OpenAI plans to release its next big model by December

• An AI chatbot played an ethically questionable role in this teen’s suicide. This is an important story to keep watching.

• Meanwhile, Meta is working with the government on their AI plans

• CNET discusses how AI agents are coming for your apps

Have a burning AI question? Or stuck trying to help your team adopt AI?

We read every single question submitted and use these to inform the topics we go deep on. 👇️ Submit your question below!

That's all!

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

Cheers,

🪄 The AI Exchange Team