🔱  Sign up for a free demo today to see how the latest AI and digital advocacy tools can transform your work.

2.5 Minute Read

 The Most Controversial Term In AI (Part 1)

There is no more barbed wire topic in technology today than Artificial General Intelligence (AGI). Depending on who you listen to, whether and when AGI arrives might:

a) Cause the next Great Depression and destroy the world (not necessarily in that order);

b) Be the equivalent of discovering God, end poverty, cure disease, and solve all of the world's other problems; or

c) Be barely noticed except by matcha-sipping tech nerds in San Francisco

Given the current pace of AI development and this wide range of potential outcomes, it is time for us here at Artificially Intelligent to start talking about AGI. We will do our best to stick with the facts. When forced to speculate, we will do so only with the minimum level of responsible, irresponsible speculation. But the time for this conversation has come, because many aspects of AGI are here already, and the remaining elements may be close at hand.

And if that happens, it is highly unlikely that the impacts will be confined to Bay Area matcha bars.

AGI: The Definition That Keeps Changing

Ask ten AI researchers to define AGI and you'll get ten different answers. So with that proviso, here's our stab at a plain English definition:

AGI refers to AI that can understand, learn, and apply knowledge across a wide range of tasks at a human level or better—basically an AI that can do any intellectual task that a human can do. The key difference from current AI (like the chatbots you use today) is that AGI would be able to flexibly adapt to completely new situations and learn new skills without being specifically trained for them, just like humans do. Current AI systems are typically good at specific tasks they were designed for, but struggle when you ask them to do something truly novel or outside their training. A chess-playing AI is incredibly good at chess, but can't suddenly learn to cook or write poetry. A human can do all of those things. AGI would have that same kind of general-purpose intelligence—it could reason about any problem, learn from experience in one domain and apply it to another, and essentially think and adapt the way humans do.

So if AGI is simply recreating human-level thinking, why you might ask would it be so revolutionary?

Because humans are actually rather inefficient. We stop to sleep, eat, get sick, scroll TikTok, and take vacations. We forget things. We have bad days. We can only learn so much in a lifetime. And crucially, you can't copy-paste a human genius.

But AGI could theoretically replicate human-level genius across millions of algorithms—algorithms that never rest, never forget, and can be duplicated instantly. They could work 24/7/365 at a fraction of the cost of a human employee. Imagine 10 million Einsteins working in parallel in a single data center, sharing insights in real-time, each one capable of mastering physics, then pivoting to molecular biology, then teaching themselves Mandarin over matcha.

That's the vision of AGI: not just matching human intelligence, but deploying it at a scale and speed that's simply impossible with biological brains. One breakthrough multiplied by a million. One creative solution generated, tested, and refined in the time it takes you to read this sentence.

That's why even "just" achieving human-level intelligence would be transformative—it's not the level that matters, it's the scale.

And if you don't think that's a big deal, remember this:

JUST ONE EINSTEIN CREATED OUR ENTIRE MODERN WORLD.

So When Will AGI Arrive?

This is where it gets truly interesting. Last week we wrote about how AI itself is accelerating the pace of AI development. And there's a certain amount of moving the goal posts happening in defining AI. If you had told almost anyone ten years ago what today's AI systems do—hold nuanced conversations, write code, analyze images, pass the bar exam, compose music—they absolutely would have called what we already have AGI.

However, it's also true that we don't yet have that data center of Einsteins solving climate change or curing cancer. But here's what we do have: OpenAI, Anthropic, Microsoft, Google, Amazon, and Meta are spending hundreds of billions of dollars—with a B—building exactly those data centers. Massive facilities consuming huge amounts of power. Nuclear reactors being brought back online specifically to feed AI. The largest capital deployment in the history of technology, all betting on the same vision—AGI.

Those companies are not building those data centers to power animated videos of your pets. They're building the physical infrastructure for AGI, and they're doing it at breakneck speed. So if you think AGI is a pipe dream, ask yourself: what do you know that the world's richest companies don't?

One Way To Operationalize This Email

🔱 Read next week's newsletter to see how AGI might change public affairs. (Apologies for the goal post move there.)

🔱 If you're enjoying this content, please consider forwarding this email to a colleague or friend. 

🔱 If you're not already a subscriber, please sign up here to stay up to date on the latest developments in political technology.

🔱 If you have suggestions for future newsletters or any other feedback, we'd love to hear from you.