The total market size (TAM) for AI is a combination of two (2) primary components: The total cost of human workforces The amount of money that current and future companies will pay to start, 10x, or 1000x their business We're talking hundreds of trillions of dollars. Don't get distracted ..read more

Augmented vs. Natural Having used hundreds (and built dozens) of AI applications since late 2022, I've come to realize something crucial about good and bad use cases. Or, to phrase it as a question: When should we use AI for things vs. doing them manually? My first inclination was full speed ahead. Consume everything. Summarize everything. Blast away, basically. Why? Because I could. Because it wasn't possible before. Our first instinct with AI was to use it for everything, and it's time to start questioning that. But a few months later I realized there was a big problem with this that I hear ..read more

I clearly see the merits of both sides, similar to the end-to-end encryption discussion. The last few years have given me some clarity on where I stand on the topic of internet censorship. I was a bit wishy-washy on it for a while because the downsides of both are so bad, but for tech platforms like X and Facebook, I think I finally have an opinion. The platforms should make all legal speech available to anyone who wants to see it. They should also make it easy for users to shape the experience they want to have. In other words, the platforms shouldn't filter reality in any way other than l ..read more

The ideal number of human employees in any company is zero. This is a flaw of capitalism, not automation, yet in the next couple years there will be a growing tendency to blame AI for the loss of human jobs. The strange truth is that the entire phenomenon of people having jobs that pay them money comes from one simple dynamic: There are some people who want to make things that don’t exist yet, and they are unable to make those things without external help. When either of those two things reduces, or goes to zero, so goes much of our current economy. It’s the drive to make things, and the fact ..read more

UL Member Content Table of Contents Russia and Ukraine Russia and the US The Rise of Christian Nationalism US Goes Hardcore Pro-Israel China and Iran Become the Main Adversaries of the … The US Buddies Up Even Closer with India Axis and Allies The AI Crash-Takeoff Enterprise Continues to Roll Out AI AI Agents Start to Mature AGI in 2025? The Pro-business Environment Post-US-Election Combined Forces Accelerate the US Relative to the … Job Market Dynamics Getting Ready For All of This Predictions Recommendations Final Analysis As with every piece in this s ..read more

SECURITY | AI | PURPOSE UNSUPERVISED LEARNING is a newsletter about upgrading to thrive in a world full of AI. It’s original ideas, analysis, mental models, frameworks, and tooling to prepare you for the world that’s coming. TOC SECURITY AI / TECH HUMANS IDEAS DISCOVERY RECOMMENDATION OF THE WEEK APHORISM OF THE WEEK Hey there! Had a great conversation with Rob Allen from ThreatLocker about their Zero Trust approach: deny-by-default, dynamic ACLs, and blocking ransomware at every stage. A Conversation with Rob Allen from ThreatLocker The UL Black Friday Membership windo ..read more

The structure of the rate_ai_result Stitch Since early 2023 I’ve wanted a system that can assess how well AI does at a given task. And when I say "system", what I really mean is an AI system. Which means I want an AI system that rates AI systems. There are a bunch of these out there now, as well as a number of AI output eval frameworks that are somewhat useful. But I wanted a simpler architecture that uses high-quality prompting to do the work. In other words, what could I give a smart, Judging AI as instructions such that it can evaluate the sophistication of less smart, to-be-tested AI? So ..read more

SECURITY | AI | PURPOSE UNSUPERVISED LEARNING is a newsletter about upgrading to thrive in a world full of AI. It’s original ideas, analysis, mental models, frameworks, and tooling to prepare you for the world that’s coming. TOC SECURITY AI / TECH HUMANS IDEAS DISCOVERY RECOMMENDATION OF THE WEEK APHORISM OF THE WEEK Hey there! I created the first Fabric Stitch, called rate_ai_result (DIRECT LINK). It rates the quality of AI #1 using the judgement of AI #2! And the result is an assessment of how smart AI 1 was on the following scale: A rate_ai_result output example You ca ..read more

SECURITY | AI | PURPOSE UNSUPERVISED LEARNING is a newsletter about upgrading to thrive in a world full of AI. It’s original ideas, analysis, mental models, frameworks, and tooling to prepare you for the world that’s coming. TOC SECURITY AI / TECH HUMANS IDEAS DISCOVERY RECOMMENDATION OF THE WEEK APHORISM OF THE WEEK Hey there! ??Going forward I’ll be sending this newsletter from newsletter@unsupervised-learning.com instead of daniel@danielmiessler.com, so please add newsletter@unsupervised-learning.com to your contact list to avoid future newsletters going to spam. ?? My ..read more

I have some interesting thoughts going into the big day tomorrow, and I wanted to try to capture them concisely. The best overview debate (Harris vs. Shapiro) But first, here’s the best debate to listen to if you want the strongest arguments on each side. It’s Sam Harris against Ben Shapiro—with Bari Weiss moderating (and asking some extraordinary questions herself). Can’t imagine a better debate from better people. The power of Framing Anyway, one of my first thoughts is that they’re all—even Ben—missing something massive in all this. Vibes. Or in the language I’ve been using lately—Framing ..read more