AI for ADHD

We have a special treat in this newsletter today: an excerpt from the new book AI for ADHD: A Practical Guide to Starting (and Actually Finishing) What Matters by Lindsay Scola.

I’ve read an early copy of the book, and I loved it. Scola does a great job breaking down complicated AI concepts into highly understandable ways - so every reader can use AI to its full potential. Whether or not you have ADHD, you’ll get a lot out of the book; I picked up so many strategies I could put to use right away to use AI to get more stuff done, faster.

Below is a part of the book I particularly loved explaining ChatGPT in a fun way.

Enjoy, and grab your copy of AI for ADHD for $0.99 here! - Molly Beck

“What Even Is ChatGPT? (And Why It's Not Here to Steal Your Soul)

Let’s start simple: ChatGPT is a very advanced autocomplete. Think of it like that time you started typing a text, and your phone aggressively tried to finish your sentence for you. The difference? Instead of just predicting a word or two, ChatGPT predicts entire sentences, paragraphs, even pages—based on patterns it’s learned from reading an absolutely ridiculous amount of internet text. (Plus it will never try to change anything to duck.)

Analogy Time: Imagine ChatGPT as an intern who has spent years cramming information from books, articles, blogs, and Reddit threads (some more reliable than others). It doesn’t actually understand anything—it just predicts what a smart-sounding response should look like based on the data it has seen.

Does this mean ChatGPT knows everything? Nope. Does it sometimes make things up with full confidence? Absolutely. (We’ll get to that.)

But here’s the thing: it doesn’t have opinions, emotions, or secret plans to take over the world—despite what some Reddit threads might lead you to believe.

The AI Landscape: ChatGPT Isn't the Only Player

While this book focuses on ChatGPT (it's the most widely used and accessible), it's worth knowing there are several other AI assistants out there:

• Claude (by Anthropic) – Known for longer context windows and nuanced responses

• Gemini (by Google) – Integrates well with Google services

• Mistral – An open-source option gaining popularity

• Specialized AI tools – Like Midjourney for images or Fathom for transcription

Throughout this book, we’ll use ChatGPT as our example—but the principles apply broadly. The good news? The skills you learn with one AI tool generally transfer to others. It’s like learning to drive a car: once you know how to drive a Toyota, you can probably figure out a Honda without too much trouble.

How ChatGPT Processes & Generates Text (aka The Magic Behind the Curtain)

Alright, for my fellow "But how does it actually work?" people—this part's for you. If you don't care and just want to use the damn thing, feel free to skim.

When you type something into ChatGPT—let's say: "Give me three simple strategies for staying productive while working." Here's what happens under the hood:

• Your words are turned into numbers. Not the scary ones from Severance (hopefully)—just in the "this is how computers process language" kind of way

• Those numbers run through a massive neural network (think of it as an AI brain with billions of digital "connections" that help it predict what words should come next)

• ChatGPT picks the next word… then the next… then the next. Based on probability, it chooses the most likely sequence of words that sound like a reasonable response

• The answer appears on your screen, word by word. And that's why sometimes it takes a second—it's literally generating the response on the fly

Important Note: Because ChatGPT doesn't think, it doesn't understand—it's simply predicting what a logical answer should look like based on what it's seen before. That's why its responses can sound fluent but still be completely wrong.

That's because AI isn't automatically searching the internet in real-time—it's working from its training data. Unless you're using a service with real-time browsing (which is often a paid feature on many AI tools), it's like a really well-read friend who knows a lot but doesn't always fact-check before speaking. (So, basically, a day that ends in y on Twitter.)” - excerpted from AI for ADHD: A Practical Guide to Starting (and Actually Finishing) What Matters by Lindsay Scola.

Thanks for stopping by Beck on Tech, Lindsay!

-Molly Beck