AI Beat Humans at Persuasion (And That’s Not Even the Scary Part)

In partnership with

Hey

Ever feel like AI is not just getting smarter but also... more convincing?

You're not wrong.

I recently stumbled upon a fascinating, and frankly, a bit unsettling, piece of research. It pitted an LLM (Claude Sonnet 3.5, to be exact) against human experts in a persuasion game, a quiz where they had to influence participants' answers.

The thought to ponder on? The LLM was significantly better at it, whether guiding towards correct or incorrect answers.

Think about that for a second. AI is already outperforming paid human professionals in the art of persuasion.

This isn't just about acing a quiz; it's a glimpse into a future where AI could influence decisions in far more critical areas.

(I mean I’ll have to ask Claude to persuade you in joining our WhatsApp community now.)

And if we talk about future, AI isn't just taking our jobs here. It's taking our electricity too.

Tech Giants Going Nuclear (Literally)

Former Google CEO Eric Schmidt recently dropped an unsettling prediction: AI data centers could consume up to 99% of the world's electricity in the near future.

This isn't some far-off dystopian scenario. It's happening now.

Microsoft is restarting the shuttered Three Mile Island nuclear plant to power its AI data centers. Amazon is investing in small nuclear reactors, while Google is partnering with nuclear innovators like Kairos Power.

Why? Because training and running advanced AI models demands astronomical amounts of electricity. Some next-gen data centers will consume as much power as ONE MILLION American households.

But here's Schmidt's wild solution: launch AI data centers into space with solar power. Yes, you read that right, orbital data centers.

The Job Apocalypse No One Wants to Talk About

While we're figuring out how to power these AI systems, another crisis looms: massive job displacement.

Anthropic CEO Dario Amodei recently warned that AI could eliminate up to 50% of all entry-level white-collar jobs in the next 1-5 years, potentially pushing unemployment to 10-20%.

The industries most at risk? Finance, law, consulting, and technology—precisely those that have traditionally absorbed college graduates.

Amodei painted a particularly chilling picture: "Cancer is cured, the economy grows at 10 percent a year, the budget is balanced—and 20 percent of people don't have jobs."

The Democratization Countertrend

But amidst these concerning developments, there's a fascinating countertrend: the democratization of AI through local deployment.

Tools like Ollama, Vllm, LM Studio, Jan etc are making it possible to run sophisticated models on consumer-grade hardware. And to add to it, there were many new open source launches last month -

DeepSeek R1-0528  - it closed the gap with proprietary giants, nearly matching OpenAI and Google across benchmarks.

 MiMo-VL-7B by Xiaomi - a 7B reasoning model and VLM variant that's outperforming models 10-30 times larger.

Mistral Devstral - a 24B parameter coding model built with All Hands AI that outperforms much larger models while being small enough to run on your Mac or a single GPU.

The beauty of this moment is that you don't need to be a tech giant to participate in the AI ecosystem.

Whether you're experimenting with Ollama for local deployment or building applications that address the current and future needs of the target market, the barriers to entry have never been lower.

Which of these trends do you see having the biggest impact on your AI journey? Hit reply and let me know. I'd love to continue this conversation.

Until next time,

-AP

Learn AI in 5 minutes a day

This is the easiest way for a busy person wanting to learn AI in as little time as possible:

1. Sign up for The Rundown AI newsletter

2. They send you 5-minute email updates on the latest AI news and how to use it

3. You learn how to become 2x more productive by leveraging AI

Sign up to start learning.