245 | 🙅You have no more excuses

In today’s Brainyacts:

1. No excuses

2. Can law students cheat doing this?

3. AI in Defense agencies and other AI model news

4. An event deepfake(?) and more news you can use

   👋 to all subscribers!

To read previous editions, click here.

Lead Memo

📧🍀 You have no more excuses for not sending good client emails.

I’ll admit it: I hate writing emails. For years, it’s been the knock against me—“non-responsive,” they’d say. But who can blame me? Emails are like Trojan horses, each one sneaking someone else’s work onto my plate. The constant pinging, the CCs, the sense that every unread email is a ticking time bomb—it’s exhausting. And yet, every ignored message comes with a cost.

What might be worse though? Never sending a useful email to a client.

A few weeks back, I found myself at this after-hours event. The topic was sharp, relevant—one of those rare talks that really gets you thinking. It was a mix of networking and insight-sharing, and I knew as I sat there that a few of my clients would eat this stuff up. Venture funding, family offices, law students partnering with outside firms—it was a snapshot of the future. And as I left, I knew I wanted to share it with my clients, to show them I’d been thinking of them and to pass along something useful.

But here’s the catch: I couldn’t, for the life of me, figure out how to broach it. How do you take an out-of-the-blue email, make it concise, interesting, and valuable without just cluttering an inbox? I didn’t want to send something that would annoy them or come off as empty schmoozing. I wanted it to feel intentional.

Now, I could have pulled out my phone right there, tried to hammer out some message, but honestly? I knew I’d spend the better part of an hour trying to get it just right. And maybe I’d scrap it halfway through. It was late, and I opted for a walk instead—the 20-minute stretch back to my hotel. But then I thought, “Wait, what if I let AI help with this?”

So, right there, walking down a quiet, moonlit street, I pulled out my phone, opened ChatGPT, and just started talking. “I just came from this event,” I said, and then rambled on, describing the topics, the clients I had in mind, the tone I wanted—friendly, casual, thoughtful. I asked ChatGPT to package it up in a way that would engage the managing partner and chairman of the firm, a way that said, “Hey, I saw this, thought of you, and wanted to make sure you were in the loop.”

I hit enter, went back to my room, opened my laptop, and there it was. The email, ready to go—thoughtful, concise, and just polished enough to feel intentional. I tweaked it, added a bit of myself, pasted it into Outlook, and hit send. Done.

Now, between you and me, that process probably saved me 45 minutes, maybe more. But the real kicker? Without AI, I might not have sent that email at all. And maybe you think, “So what? One email doesn’t make or break a relationship.” But the truth is, one email isn’t just one email. One missed check-in, one missed insight—it all adds up. Enough of those missed moments, and yes, it can absolutely impact your relationships. AI didn’t just save me time; it saved a small piece of connection. And in a business like ours, that connection is everything.

Here is the actual transcript of my talk with the OpenAI app. I am sharing this so you can see the robustness (and cluttered) way I prompted it. Its a mess but it has a lot of context and information buried in between the “ums” and “so’s”.

I continued to about double this length. You can see how I was just talking away.

And here is some of what I got back:

The point? You have no more excuse for not sending good client emails.

Spotlight

🚨⚖️ Balancing Innovation and Integrity in Legal Education

With the rise of accessible LLM platforms like LM Studio, it’s increasingly feasible for students (or practitioners) to run powerful AI models directly on their laptops, completely offline, without data traveling over the internet. This setup can support secure applications, allowing users to privately interact with sensitive documents or generate responses without the risk of data exposure.

However, this accessibility brings its own set of challenges. As I demonstrate in my video tutorial, local LLMs raise critical questions around academic integrity. Because local LLMs don’t rely on the cloud or external databases, they can circumvent the “guardrails” typically embedded in larger, cloud-based models like OpenAI’s GPT. For example, while some cloud LLMs detect and prevent use in certain testing or assessment contexts, local models—operated independently—lack this layer of oversight, potentially offering students a tool that could be used in ways instructors might not intend.

In the video, I walk through a demonstration of this exact scenario: using LM Studio to work through a torts exam question offline. It highlights how straightforward it is for users to get started with prompting models locally and potentially bypassing guardrails against unintended uses. For legal educators, understanding how students may use these tools is crucial for designing assessments that maintain integrity.

AI Model Notables

► AI courting US Defense agencies for big windfalls

• ► related: Chinese researchers build military AI using Meta’s open-source Llama model — ChatBIT allegedly performs at around 90% of the performance of OpenAI GPT-4 LLM.

• ► related: Meta says it’s making its Llama models available for US national security applications.

► Rare bees kill Meta’s nuclear-powered AI data center plans.

► Amazon’s new Alexa has reportedly slipped to 2025.

► OpenAI CEO Sam Altman says lack of compute capacity is delaying the company’s products.

► Chipmaking giant Nvidia in talks with Elon Musk over investing in xAI.

News You Can Use:

➭ A good recruiter can spot an AI-written application from a mile away.

➭ A state-funded radio station in Poland called Off Radio Krakow recently attempted a little experiment that quickly blew up in its face. They tried replacing all of their hosts and journalists with AI voices. To twist the knife further, the station then fabricated an interview with a deceased Nobel laureate using AI.

➭ A deepfaked event: Website written by AI sends thousands in Ireland to fake Halloween parade.

➭ Boston Dynamics releases video of Atlas robot that now needs zero help from humans to perform tasks.

Who is the author, Josh Kubicki?

Some of you know me. Others do not. Here is a short intro. I am a lawyer, entrepreneur, and teacher. I have transformed legal practices and built multi-million dollar businesses. Not a theorist, I am an applied researcher and former Chief Strategy Officer, recognized by Fast Company and Bloomberg Law for my unique work. Through this newsletter, I offer you pragmatic insights into leveraging AI to inform and improve your daily life in legal services.