Hey friend,

Don’t mind me. I’m just going to show up in your inbox like I haven’t been gone for 2 years. :)

Where have I been? Well, in my office overthinking things, obviously.

At this point, I feel like I’ve been on that sandy island where Odysseus was stranded for 7 years, being fed lotus flowers by a mysterious witch so he’d forget what he left behind and how to get back home. (If you haven’t seen The Odyssey on an Imax screen yet, I highly recommend it. But bring ear plugs.)

Over the years since I last sent you a letter, many of you have written in, or told me in person when I’ve bumped into you live, asking if you’ve missed any issues, or “is there any chance you’ll start writing SITN again, because I love it so much,” or “your newsletter was one of only three newsletters to make it through my regular mass unsubscribe events. It was you, Seth Godin, and James Clear.”

I was always so touched by comments like that, and, joking aside, they made me cry a bit inside. Why? Because sharing thoughts and ideas about business things that can help you breathe a little easier in life feels like what I was meant to be doing with the time I have left in this life.

What has stopped me from regularly doing that is things like the mortgage and groceries and health insurance.

Where I got stuck is in trying to make this newsletter the product. The main source of revenue. Doing that is possible, and it’s still the goal, but the journey to that destination feels as long as Odysseus’s journey back home.

In the meantime, I think I’ve found a way I can write to you, regularly, without needing the newsletter to be my sole source of revenue, and without needing to rob a bank.

That path involves a shift in focus, a narrowing of the area I point my laser, a defined area where I will allow myself to go down rabbit holes. (Yup, I still like metaphors.)

I have a large number of interests in life and business, and I’m constantly learning in many domains. But there’s one common thread of my interests over the last 20 years: using technology to help make people’s businesses easier, more efficient, and more profitable.

And there’s one area of technology that has emerged over the last 3 years that is having a larger impact, and greater opportunities to help us, than any other technology I’ve ever used. Yes. It’s artificial intelligence.

If you hate AI, before you head off to another email, hear me out. And I promise I’m not turning into an AI hustle bro. Seriously. I’m like the opposite of an AI hustle bro.

The thing is, there’s good uses of AI and bad uses. Personally, I can’t stand when people use AI to create images and videos that most people can’t tell are fake, and then post them on social media as if they’re real. I’m strongly against that use case of AI and hate it along with millions of other people.

But there are so many uses of AI that can make our lives and businesses better and more profitable. Like automating tedious tasks in your business that you or your employees really don’t need to be doing. That’s what I’m talking about.

Automating boring workflows that AI can easily and quickly handle so you and your team can focus on things that can grow your business…things that only humans can do.

It’s about using AI tools and agents to free up time for your team and bring a sense of calm into your business.

It’s about helping you and your team members become AI augmented superheroes.

That’ll be the focus for this next season of Signals In The Noise (The Reboot):

Simplifying the best tools and tactics from the very noisy, buzzy, FOMO laden, complicated world of AI so you can safely integrate practical workflows of automated tech into your business.

I’m talking about efficiencies and automations that will make a noticeable difference in the quality of life for you and your team, while adding some zeros to the bottom line of your P&L. More zeros to the right are good. We like zeros.

If you’re down to go on this journey with me, I promise there will be no giant cyclopses, no ship destroying whirlpools, and absolutely no witches that will turn us all into pigs.

This newsletter is getting long already, but maybe you’re remembering that I have a bit of trouble writing short? I’ll wrap things up, but let me get to a few AI things that could help you in your daily work life, shall we?

1
Writer’s Talk

Let’s get right to the low hanging fruit. This is one of the easiest ways to get back several hours of time per week. But you’ll have to let go of your romantic love of typing with your hands. If you can do that, Wispr Flow will give you back literally hours of time each week. For decades I’ve tried various speech-to-text apps and could never get past all the errors they made. I always, always went back to typing with my hands after a month or two. Wispr Flow is the first one that blew my mind with how accurate it is. Look, I love writing with my hands, but the speed you can write emails and other things with Wispr Flow is stunning. For me, it’s 2 to 3 times faster than writing with my hands, and I can type pretty fast. 

2
Hit Me With Your Best Bot

I’m a heavy Claude user. “Claudie,” as my wife calls him, is great for a lot of things, but like other frontier AI models, he still gets things wrong. The more complicated and multi-layered the topic is, the greater chance he’ll get things wrong, while still acting as if he knows exactly what he’s talking about. (I’m pretty sure he’s an enneagram 5.)

The best way I’ve found to keep him, and other AI models, on track is this:

  1. Do whatever you’re going to do with Claude, or chatGPT, or Gemini, such as creating a deep dive research report on a topic.

  2. Take the artifact that the model you’re using generates, like a markdown or pdf file of a report, and give it to another model. In this example, let’s say you have Claude do a deep dive research project and then create a report. Take that report and give it to chatGPT/Codex and give it context about what the thing is you uploaded. Then, ask chatGPT/Codex to grade the report and suggest ways it could be improved…and then have it to improve it.

  3. Take the second improved version from chatGPT/Codex and upload it back to Claude (which you should sometimes pronounce as “Claudé,” like the way the singer Sade pronounces her name), and ask Claude to grade chatGPT’s version and make suggestions to improve it. Then ask it to create an improved version 3.

That gives you the best of both worlds. Two frontier AI models work on it and correct each other’s mistakes. You’ll end up with a much better report, or whatever artifact you’re creating, and you’ll have a much higher level of confidence that the output isn’t full of errors.

3
Get The Grok Outta Here

Lest you think I’m an AI shill that will recommend any AI tool no matter what, here’s a head’t up about something to avoid. Grok. Musk’s AI model. My recommendation: just…don’t use it. It can do a few neat things, but a) Musk is morally reprehensible, and b) he had no problem releasing features in it that allowed people to create nude images from a photo of any clothed woman you gave it, and c) it’s nowhere as good as the other frontier AI models.

I find Claude to be the best at writing (currently, the Opus 4.8 Claude model is doing a bit better than Opus 5.) I still use Claude Code for coding, but I use Codex for code review. Many people do the reverse, with Codex as their main coding tool and Claude Code as their reviewer. And Gemini is great to pull in for deep dive research projects.

Alright. That’s enough for today. If you’ve read this far, thank you for spending a little time with me today. I hope this was helpful.

It’s simply lovely to be back in your inbox again.

Let’s see how this all unfolds.

I hope you’ll stick around with me on this journey.

Your friend,
~ Forest Linden
The AI guy