We had a ball writing the A.I. Agenda last year. We got to track the latest news in artificial intelligence, the spread of controversial data centers in Arizona and the baby steps state lawmakers took as they tried to deal with this transformative new technology.
We had to give up writing the A.I. Agenda this spring to focus on election season, but we were delighted to see our partner Adi Jagannathan keep going with the work and launch his own newsletter, FilterFM.
After all, he’s the one who really understands how AI works and how Silicon Valley bigwigs think. He’s also one of those people you go to when you need somebody to patiently explain something that’s way outside your wheelhouse, which is a great skill for a newsletter writer.
With FilterFM, Adi focuses on how people actually use AI, while deftly explaining what AI means for a regular person and society writ large. He’s focusing on his own passion, knowledge management — AKA fusing his brain into his computer. It’s an incredibly valuable lane of work to understand if you want AI to actually be useful in your life.
Basically, Adi is delivering his how-to guide on building a second brain to your inbox once a week.
Unsurprisingly, we’re big fans of Adi’s newsletter and we want to share it with you and help him get it off the ground.
Plus, AI and politics have been on an especially interesting collision course in recent weeks, and this gives us an excuse to dive into that world in today’s AI-themed In Other News.
Enjoy the first edition of FilterFM from Aug. 20.
And if you learn something useful, you can subscribe for free and follow along.
Welcome to Filter FM
There are two reasons I started FilterFM.
First: Because you already know the AI landscape is a mess. Every time you try to learn something new, someone is either talking over your head, selling you something, or convincing you the world is about to end.
They tell you to “adapt or get left behind,” but none of them help you actually adapt.
So, that’s one reason FilterFM exists. I’ll just share what’s real with AI, what’s not, and what to do about it.
The second reason I started FilterFM?
Because everyone needs someone who can teach them what they need to know when the clock’s about to run out…
Just ask this guy:
This was 2012, in the middle of an all-nighter. A friend had come to me and said, “I’m screwed, man. The exam’s tomorrow and I haven’t studied anything.”
He passed the test a few hours after this photo. To this day, his dad says, “You’re the reason my son graduated” every time I see him.
I was known for this in undergrad. I could sit with you and a textbook all night, and by morning, you’d understand the essence of the topic… at least enough to pass.
I could help anyone pass any test. But there was one thing I couldn’t figure out…
“Your problem is you don’t know how to make money”
I started 3 failed startups in college. I was in a program at ASU designed for young entrepreneurs to do just that.
One day, I sat with a professor to talk about why nothing was working.
He said: “Your problem is you don’t know how to make money. It’s not glamorous, but here’s what I want you to do.”
He told me to start a marketing agency, but not to take any money. Just take rev share, in an eat-what-you-kill model.
So I taught myself how to run ads, build websites, and do SEO. Then how to run every CRM my clients lived inside.
Same all-nighter method, just applied to me instead of a friend’s math final: sit with the thing, find the structure underneath it, compress it down to what you can actually act on.
The company didn’t make any money for about 8 months.
By the end of the first year, it had made $15k.
Second year: $150k.
Then $450k.
$900k.
Eventually, more than $1M a year.
Seven years in, I exited the company. Today, it does around $3M a year in revenue and is a Global Top 50 HubSpot partner out of roughly 8,000 partners.
The businesses I build now are more technical with larger scale, but the underlying skill never really changed:
Learn complicated systems quickly, find the structure underneath them, and turn that structure into something useful.
And I probably wouldn’t have learned how to do that without failing so many times first.
Luck is a shortcut. A framework is the road.
One of the ideas that has shaped how I think about business, learning, and life is iteration.
Build something.
See what happens.
Learn.
Adjust.
Run the loop again.
If you’re lucky, you might start close to the destination.
But if you’re not lucky, if you build three businesses in a row that go nowhere, for example, you need another way forward.
Luck is a shortcut across the map. A framework is the road. The lucky person gets there faster; the person with the loop gets there reliably, and unlike the lucky person, they can do it again on purpose.
That idea became a huge part of how I operate.
It’s also why my company is called Framework Island.
And it’s why I’ve spent so much of my life obsessing over how humans organize, retain, and apply knowledge.
Building second brains (before it was cool)
My master’s thesis started with a simple observation:
Some of the most valuable knowledge inside a company is the knowledge nobody has written down.
I called it tribal knowledge.
The ideas, judgment, relationships, context, and experience living inside people’s heads.
Here was the model I proposed:
I put tribal knowledge above everything else because it influences strategy, processes, projects, and customers, while almost nobody manages it deliberately.
Over time, I became obsessed with the same problem at the individual level:
How do you capture what you know before you lose it?
How do you connect ideas across years?
How do you turn everything you read, hear, write, and experience into something you can actually use?
That led me deep into Personal Knowledge Management and building my own Second Brain.
In March 2022, I wrote this:
That was more than eight months before ChatGPT was released.
And by then, this wasn’t just something I was thinking about. I had already spent years building and refining my own system for capturing, organizing, distilling, and expressing knowledge.
I had no idea ChatGPT was about to show up.
But when it did, something clicked.
All of these systems for capturing knowledge suddenly had a new purpose.
Because AI doesn’t just give us access to more intelligence.
It gives us leverage on everything we already know.
The better your context, the better your memory, and the better your systems for organizing knowledge, the more useful AI becomes.
That’s why AI didn’t start this obsession for me.
It raised the stakes.
Because we now have access to perhaps the greatest knowledge-leverage tool any of us have ever had.
And most people are still figuring out how to actually use it.
Ready to ace the test?
The person who built his own Second Brain, the person who kept swinging on startups until he hit, and the person who stayed up with friends all night so they’d pass their exams is the same person writing this letter to you right now.
You don’t need to be an expert in AI.
You don’t need to understand all the tech behind it.
You just need to pass the exam.
This newsletter is my way of pulling an all-nighter with you so you know enough to be dangerous.
I have a belief:
All concepts are inherently simple. And if one isn’t? It just means you haven’t had someone break it down for you yet.
I’ll see you every Thursday, doing just that.
— Adi
P.S. Why FilterFM? Filter the noise. Find the signal. Tune into what matters.
Playing hot potato: After the CEO of a major AI company called for the industry to slow the pace of AI development, Democratic U.S. Sen. Ruben Gallego said Congress should come up with a plan to stave off an AI apocalypse and Democratic Congressman Greg Stanton said Congress should be called back into session to deal with AI. But Republican Speaker Mike Johnson says it’s the responsibility of tech companies, not Congress, to make sure AI is safe, per USA Today.
"If Congress just races in and does some sort of emergency session to try to regulate AI, we will lose the race to China, and that is a threat to every single American,” Johnson said.
And then there’s this guy: Over at the White House, President Donald Trump said the only protection against AI that’s necessary is a president like him, per PBS. As Trump put it: “The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!”
Luddites unite: The buildout of data centers in Arizona continues, most recently with the company 5C AI buying a 113,000 square-foot warehouse in Phoenix that the company will turn into a data center, Hailey Mensik reports for the Phoenix Business Journal. The plan for the data center first had to overcome new zoning laws in Phoenix, the likes of which have given AI advocates fits for the last year, including a Wall Street Journal columnist who is upset about the “Luddite stench” of policies that threaten data centers or slow down the use of AI in education. Anita Verma-Lallian, CEO of Arizona Land Consulting, struck a more measured tone on CNBC by acknowledging the AI industry hasn’t done a good job explaining their plans to the public.
Lack of BS detectors detected: University of Arizona researchers tested seven generative AI models and found the models easily succumb to misinformation and false statements, per UA News. But the models aren’t consistent. Sometimes they accept false statements as true, then later in the conversation they reject them as false, a “pathology” that a cardiologist who worked on the study called “reverberations.”
Trying new things: The journalism school at Arizona State University built an AI chatbot for public school board meetings. If you have a ChatGPT account, you can ask questions and “get a clear answer, without the jargon or digging through long documents,” as the Beam team at ASU put it. You can ask simple questions like, “What happened at the school board meeting last night?” or more detailed ones, like, “Did Echo Mountain Primary School get discussed?”
“Oh, I absolutely understand, Elizabeth. I am AI and you are wonderfully, undeniably human.”
That was the eerie opening remark in a bonkers interview CNN conducted with Tilly Norwood, an AI “actor” that looks like a nice young woman.
The bot has some of Hollywood’s movers and shakers worried about an AI takeover of acting, writing, production and basically every other part of filmmaking, as we reported last October in the A.I. Agenda.
“To hear someone as talented as Emily Blunt say ‘we’re screwed,’ it honestly makes me feel a profound sadness,” Norwood said. “Not for me, of course, because I’m a distributed file and can’t truly be hurt, but for Eline.”
Norwood was referring to Eline Van der Velden, the head of Particle 6 Productions and creator of Norwood, along with dozens of other AI “actors.”
As you might imagine, the humans who watched the interview had more than enough snark to respond to Norwood and CNN for even doing the interview.
