An ancient institution warns about the dangers of a tiny elite controlling the future. Beat. Meanwhile, a machine just cracked a 56 -year -old math problem for the price of a fancy dinner. Two completely different worlds colliding, you know? Let's unpack it. Welcome to today's Deep Dive. I'm really glad you are here with us. We have a truly fascinating journey ahead today. We absolutely do. First, we are going to explore a surprising new encyclical. It comes directly
from Pope Leo XIV. We're looking at AI. power and the future of democracy. Right. Then we take a rapid -fire tour of today's major AI industry moves. We're talking spy agencies buying up chips and, you know, workers automating their own jobs. And finally, a mind -bending math showdown. And AI just proved it can generate entirely new knowledge. The three -line for today is the tension of control. Who exactly controls this frantic technological race? And what is the technology actually becoming
capable of? It is a massive question. And, well, the Vatican just weighed in heavily. Yep. Start right there. Pope Leo IV just released a new encyclical. It is titled Magnifica Humanitas, and it is making serious waves. Oh, for sure. I mean, it uses AI as a primary hook, but it actually tackles much deeper issues. The rapid erosion of democracy, the rise of massive inequality. Yeah. Specifically, how tech elites are essentially
steering the world to benefit themselves. The timing of this document is incredibly striking to me. The context here is super important. The encyclical dropped right after a major political shift in Washington. President Trump had just delayed a major AI executive order. Just looking at the timeline, that order was meant to give government oversight, oversight for new powerful AI models. Exactly. And billionaires are currently pumping hundreds of millions into super PACs.
They are lobbying specifically to kill AI safety regulations. So the pope is reacting to a very specific moment in time. He argues something very specific here. Tech built by a tiny opaque group cannot serve the common good. AI naturally amplifies the power of those who already have money and data. And they use that data advantage very effectively. They use it to manipulate our choices. They manipulate what we buy every day and even how we vote. I have to admit something
here. I still wrestle with knowing what's actually true online myself. These deep fakes, they really blur reality. You are definitely not alone in feeling that way. The Pope actually highlights that exact issue. Endless data deepfakes are destroying our collective truth. Yeah, we are losing our shared ability to agree on basic reality. Without a shared reality, democracy really struggles to function. So the Pope makes a very explicit demand here. He calls for an end to this frantic
push. He specifically means the push for ever larger data sets. He is warning against the blind race for more powerful models. The document makes a very sharp philosophical point. Just because a small group can build insanely powerful tech. Right. And they definitely have the capital to do it. It doesn't give them the right to govern society. It's like building a high speed train, but letting the engineers decide what cities are allowed to exist. That is a perfect way to
visualize it. The engineers shouldn't be the mayors. And what is fascinating here is the historical parallel. It mirrors history almost perfectly. You're talking about the Industrial Revolution, right? Yes. Back in 1891, the church issued a very similar warning. They called out the exact same type of exploitation. Back then, it was factory owners. Today, it is data centers. Do historical warnings like the 1891 decree ever actually slow down technological momentum, or
do they just document it? They rarely stop the actual machinery of progress. You can't uninvent the steam engine, and you can't uninvent algorithms. But they do something equally important. They shift the cultural baseline of what society accepts. So moral warnings highlight the stakes, even if they don't hit the brakes. Exactly. They force us to look at the road ahead. They give lawmakers a moral vocabulary to eventually pass regulations. The Pope warned about this frantic push for dominance.
Let's look exactly at what that frantic push looks like on the ground this week. The corporate shifts right now are just staggering. Let's start with a company called ClickUp. They just laid off 22 % of their workforce. Wait, 22%. Is that a direct result of the AI or is the CEO just using it as a convenient excuse? It seems directly tied to their internal tech adoption. They rolled out 3 ,000 internal AI agents just before the layoffs. The CEO claims the future belongs to
employees who automate their own jobs. Automate your job or lose your job entirely. That is a brutal new reality for average workers. It is a ruthless efficiency engine at work. Meanwhile, the stakes are rising on a national level. The White House just approved a massive spending bill. They approved $9 billion for U .S. spy agencies, buying up advanced AI chips at an incredible rate. Officials are reportedly terrified of falling behind globally. They believe they need to deploy
frontier AI systems as quickly as possible. Frontier AI systems, basically the most advanced models pushing the absolute limits. Yeah. The government sees hoarding these chips as a core security imperative now. At the same time, the broader development speed is dizzying. Look at Google's AI Studio. They let users build native Android apps. You can do it directly on their platform for free. And the volume of output is just insane. Over 250 ,000 apps were built in a single week.
A quarter of a million apps in seven days. That changes the entire software landscape completely. It is like stacking Lego blocks of data. It really democratizes code. But it also floods the zone. And the models themselves are evolving incredibly fast, too. Leaks suggest the next Grok model is very close. They are calling it the 1 .5T V9 medium model. It might drop in just two to three weeks. Right. And the most interesting part of the leak involves the training data.
Cursor data is heavily added into the mix. Let's pause on that. Why does cursor data matter so much? Well, Cursor is currently one of the most popular AI coding editors. Developers use it every day to write software. If Grok is craning on how developers interact with Cursor, its coding capabilities will skyrocket. I mean, it is learning how we code while we are coding. That is a wild feedback loop. We are also seeing practical daily advice emerge. Yes. The creator of Cloud Code
shared a major tip for daily users. People are always looking for ways to get better results. What was the number one piece of advice? It is surprisingly simple. But it changes everything. You just explicitly tell the AI to review its own code for bugs before it shows you the output. Just asking it to double check its work. Yeah, that simple second pass catches a huge percentage of errors. These tools are becoming our everyday digital companions, but they still have some
very weird, unsettling quirks. Like Google's new AI search feature right now, it has been acting a little bit strange lately. If you try searching for certain command words, words like ignore or stop. It breaks the illusion of a standard search engine, starts talking back to you directly. It responds just like a conversational assistant would. It reminds you there is an agent behind the search bar. Google says they are currently working on a fix for it. Meanwhile, hardware
is shifting its focus entirely right now. Apple is working on rumored new AI glasses. But they are taking a very different approach from their competitors. They may skip flashy augmented reality screens completely. Right. They are focusing on a very lightweight design instead. Simple gesture controls for the user. And prioritizing a much smarter, context -aware Siri built right in. A Siri built for actual everyday use. How
do we reconcile this? Spy agencies hoard AI chips for global dominance, while average workers must automate their desk jobs or get fired. It really highlights how AI is a dual use extreme right now. It is a powerful weapon of statecraft at the very top and a ruthless efficiency engine at the very bottom. It's a two front race. National security at the top. Personal automation at the bottom. Precisely. The squeeze is happening to
society from both sides simultaneously. If our national security, our daily livelihoods depend on these frontier models, just how smart are they actually getting? Let's look at a massive recent breakthrough, a breakthrough that proves they aren't just memorizing data anymore. We will dive into that right after this short break. Mid -roll sponsor break. We are back. Let's talk about the math showdown. This is where things get truly futuristic. Google DeepMind is officially
firing back at OpenAI. This all started when OpenAI made a very large public claim. They claim their system managed to disprove an 80 -year -old math conjecture. So Google DeepMind just answered with something undeniably huge. Their new system is called Alpha Proof Nexus. And Alpha Proof Nexus just solved nine open -air disk problems. For those who might not know, what exactly is an Erdos problem? Paul Erdos was a legendary
eccentric mathematician. He left behind hundreds of unsolved problems in number theory and combinatorics. They are notoriously tricky puzzles that require deep logical leaps. An alpha -proof nexus just solved nine of them. Two of these specific problems have stumped human mathematicians for 56 years straight. It didn't stop there, either. It also recently solved 44 open conjectures from the online encyclopedia of integer sequences. The mechanism behind this breakthrough is truly fascinating
to me. It relies on a very specific loop between two different systems. The AI system itself writes the actual mathematical proof. It generates the logical steps. Then a completely separate system called Lean steps in to check it. Lean is essentially a programming language designed to verify mathematical proofs. It checks the math for any potential logical errors in real time. Right. And here's the magic part. They repeat this exact loop continuously.
The AI generates, Lean checks, the AI refines. They run it over and over again until the proof is completely flawless. Two secs silence. Whoa. Imagine an AI grading its own math homework with 100 % certainty, cracking a 56 -year -old puzzle for a few hundred bucks. The cost efficiency is truly mind -blowing. Solving these historic math problems used to take human lifetimes. It took massive grants and years of intense focus. Now, each problem costs only a few hundred dollars
in compute power. It completely flips the economics of scientific discovery. But it is important to note it does have a current hard limitation. Yes, it hits a wall on certain types of abstract problems. AlphaProof absolutely crushed all of these existing logical puzzles. But it currently struggles with inventing entirely new mathematical constructions from scratch. It optimizes existing frameworks beautifully, but it isn't dreaming
up new branches of math. It's like a grandmaster who can play millions of chess games to find the perfect move, but can't invent a new board game. That is exactly the right way to view its current limits. It masters the established rules of the universe perfectly well, but it cannot invent brand new math concepts on its own just yet. Even with that limitation, the significance of this milestone is profound. For a very long time, skeptics have made a specific argument
about AI. They said AI could only regurgitate its training data. They called large language models very fancy autocomplete machines. They argued the models were just guessing the next likely word based on statistics. But disproving and solving decades old math problems changes everything. It is the literal definition of generating brand new knowledge. It proves the system is capable of genuine reasoning. Google showed us
a very clear working blueprint here. It is a blueprint for the future of all scientific discovery. Does this finally end the autocomplete debate entirely, whether AI is just an autocomplete or something actually capable of reasoning? It absolutely should end that debate for good. Solving a previously unsolved mathematical problem proves a critical point. The system is reaching far beyond its initial training data. It is tapping
into genuine, verifiable logic. Right. It's generating genuine new knowledge, not just remixing its training data. Exactly. The ceiling of the capability is rising incredibly fast right now. Let's bring this whole journey full circle today. We started with deeply human moral concerns from the Vatican. The Pope warned about morality, political power and massive inequality. He was deeply worried about who exactly wields this new technology. He warned about the tiny elite controlling the
trajectory of society. And we contrasted that human warning with cold, undeniable machine capability. Systems like Alpha Proof Nexus are rewriting human knowledge boundaries as we speak. The technology is undeniably crossing a major... It is moving from a tool of convenience to an engine of discovery. Which makes the question of control more critical than ever. Who steers this ship matters deeply to all of us. It really does. The technological race is simply moving much faster than our global
regulations. I want to leave you with a lingering question today. If we can build an AI that flawlessly grades its own logic in complex math, how long until we ask it to logically grade the messy, unequal societal systems the Pope is so worried about? And would we even listen to its answer? Thank you for joining our deep dive today, Otiro Music.