Opening the Book

Giving Your Life Details to AI

Another week, another instance of the Internet freaking out about a new AI tool.

Now, it is Clawdbot, which is essentially a personal AI assistant. It can do everything from sending emails and clearing your inbox to checking you in for flights. You can think of it as an ever-present AI that is constantly working for you in the background.

Better yet, you don't need to access this through a CLI. You can do it through Telegram or WhatsApp.

This has become a pattern, but the hype around Clawdbot has reached a fever pitch. It's unclear how people are using this other than some claiming that it has been "mind-blowing."

At this very moment, I think this excitement over Clawdbot is mostly another hype-driven phenomenon that is primarily being used for engagement farming. That said, I'm more interested in the practical ramifications of giving an AI tool over a wide swath of your personal and professional life.

Yes, having a 24/7 assistant work on menial or repetitive tasks is helpful. It can eliminate some of the busy work that many of us hate. That said, there are some privacy concerns about giving Claude control over your passwords and local files. Speaking for myself, it's one of the main reasons that I haven't tried Clawdbot- even if it has the potential to make life easier.

It leads to the question of how much you want your AI to know about you. There are no easy answers here, but people are already voting with their feet.

Ultimately, there is a major divide in American society today. On one side, you have a hardcore group of AI-enthusiasts who are looking for any and all ways that AI can improve their lives. They are opening the book, so to speak, and letting AI run with it.

On the other hand, there are those who want nothing to do with this. They actively avoid AI-related content and resist incorporating it into their daily lives.

If I had to choose, I'd probably say I'm in the former camp. However, I recognize that we're living in a new world, where AI can have a more intimate view of you than your closest friend (or even spouse). And it can seem incredibly lifelike, even if it's just trying to provide the most optimal response per its training.

I've spoken before about how some people heavily rely on AI chatbots for companionship. But putting that aside, it'll be interesting to see how many people provide most (if not all) of their personal information to LLMs. This could be everything from haphazard ideas from a long walk to sensitive medical records. I'd expect many would.

In any event, the privacy angle of LLM-use intrigues me. If you've thought about this or have read anything interesting, don't hesitate to reach out!

Prompt of the Week

I'm a huge fan of Stan Druckenmiller. He's one of the greatest investors of all time. He often speaks about following liquidity over anything else (including valuations), yet it is hard to do in practice. So naturally, I asked ChatGPT. While this may not be directly relevant to you, it can show how you can get a different perspective on emulating people that you can admire.

"I would consider myself a value investor, yet I massively respect Stan Druckenmiller. He often talks about how the availability of liquidity is the biggest factor in his broader investment process. How can I incorporate this idea into my process as a small cap long-short value investor?"

Until next week, Adam