INN023: Did this $500B AI Arms Race start a New Cold War? | Deepseek a Trojan?

Remember when high-end AI was a playground for just a few tech giants? Those days are over. From China's open-source powerhouse causing waves to AI that actually takes time to think (just like us), we're watching the future arrive faster than anyone predicted.

And what a week it's been. OpenAI's trying to turn ChatGPT into your personal assistant (with mixed results), China's DeepSeek just showed everyone how to run advanced AI on a fraction of the hardware, and Trump just bet $500 billion that America can win the AI race. Meanwhile, Perplexity's new mobile agent is showing Google how app integration should work, and Europe's still trying to figure out which way is up.

First stop: we'll explore why some AI models are starting to show signs of actual reasoning (and why that might be more important than raw computing power), dig into the real capabilities of these new "autonomous" agents, and uncover why a $500 billion bet on AI infrastructure might be the biggest technology play since the Space Race.

Why does it matter? 
Stick with us and find out. Let’s keep riding this tech wave together, and don’t forget to follow us on socials for more.

Cheers,
Patrick, Nikola & Aragorn 🚀

OpenAI's Operator: Not Quite Skynet (Yet)

OpenAI just dropped their Operator feature, and while everyone's calling it an autonomous agent, let's get real about what it can and can't do.

Think of it as an AI intern with strict boundaries – it can browse the web and handle basic tasks, but don't expect it to access your accounts or make purchases. Running in its own sandbox browser, Operator is more of a preview than a revolution.

Here's the catch: if you're in Europe, you can't even try it yet. While U.S. users experiment with this new toy, the rest of the world watches from the sidelines. But maybe that's not such a bad thing – France is already calling for Europe to pump the brakes on AI regulation and focus on innovation instead.

Why does it matter?
While it's a step toward autonomous AI agents, it's more a glimpse of the future than a fully realized solution. The feature highlights both the potential and current limitations of AI agents.

DeepSeek R1: China's Open Source Surprise

Talk about a plot twist – DeepSeek's R1 just showed up and started matching OpenAI's best work while running on just 10% of the hardware. This isn't just impressive; it's game-changing. The model offers:

• Web search that o1 doesn't have

• Way more sources per answer (try getting that from ChatGPT)

• Detailed reasoning chains that show its thinking

• Did we mention it's free?

Sure, it has some interesting quirks around certain topics (try asking about Chinese history), but the capabilities are undeniable. And while some worry about data privacy with Chinese AI, let's be honest – are Western tech giants really that different?

Why does it matter?
When high-end AI goes open source, everything changes. The pricing models that kept advanced AI exclusive are crumbling, and that's going to reshape who gets to innovate.

The AI Reasoning Revolution

Remember when everyone said AI was just a fancy pattern matcher? Well, the new models just scored 25% on tests that stump human experts. But here's the really wild part – it's not just about more data or bigger models. These AIs get better results when they have time to "think."

Just like humans perform better when we're not rushed, these models improve with processing time. We're watching AI develop something eerily similar to human reasoning, combining different types of intelligence just like our brains do with the limbic system and prefrontal cortex.

Why does it matter?
The shift to AI that takes time to "think" changes everything we thought we knew about AI development. Consider the implications:

• The focus on reasoning time over training data means smaller players can compete

• Models showing human-like improvement with thinking time suggests new approaches to AI development

• The combination of different AI models (reasoning, voice, language) mirrors biological intelligence

• Performance improvements of 1,150% in weeks show we're on an exponential curve

When AI starts thinking more like us (and better than us), every assumption about AI capabilities needs to be reconsidered.

Project Stargate: America's $500B AI Bet

While Twitter debates about who has more billions to spend (yes we mean Elon’s response under Altman’s post), something bigger is happening: a $500 billion investment in American AI infrastructure. Think Manhattan Project meets Space Race, but this time they're announcing it to the world.

Sure, they're promising job creation and economic growth. But let's be clear – this is about winning the AI arms race. In the short term, we'll see construction jobs and infrastructure growth. In the long term? We're watching the opening moves of the 21st century's defining technology race.

Why does it matter?
Just like the Space Race gave us everything from memory foam to water filters, this AI race will drive innovation across all sectors. In a nutshell this is America's Apollo moment for AI. Here's what's really at stake:

Short-term infrastructure investment creates immediate economic impact

• Competition with China drives innovation faster than market forces alone

• The scale of investment signals AI's strategic importance beyond tech

• Europe's position between two AI superpowers forces difficult choices

• The parallel with the Space Race suggests massive technological spin-offs

The question isn't whether this will transform technology – it's how prepared we are for the pace of change.

Book Corner: Living Longer, Thinking Deeper

Two must-reads this month:

• "Lifespan" by Dr. David Sinclair: Want to live to 500? It might not be as crazy as it sounds. Sinclair lays out the science of age reversal, though he might be underselling just how fast this could happen.

• "Co-Intelligence" by Ethan Mollick: Need to understand AI without the hype? This is your guide. When AI can generate 1,500 uses for a paperclip in seconds (while humans struggle to list 50), you know something interesting is happening. Read Patrick's full review here.

Why does it matter?

At a time when most AI books are outdated before they hit the shelves, these two works tackle timeless questions about human potential and our relationship with technology. Whether it's extending human lifespan or augmenting human intelligence, they help us understand not just the technology, but its implications for our future.

Newsletter full video recording