Hello and welcome to the Automated, your AI tour guide.

Imagine sinking billions into AI research, only to have a group of researchers pull off something just as impressive for a mere $50.

That’s exactly what Stanford and the University of Washington did with their reasoning AI model, s1, which rivals the performance of heavyweights like OpenAI’s o1 and DeepSeek’s R1

Here’s what we have for you today:

Turns out, just like DeepSeek showed us, you don’t need a mountain of cash to build a solid AI model—just $50, some ingenuity, and a clever trick.

A team of researchers from Stanford and the University of Washington pulled off something wild: they trained a reasoning AI model, called s1, that performs on par with OpenAI’s o1 and DeepSeek’s R1—but at a fraction of the cost.

And the best part? It's open source and completely free on GitHub.

Instead of burning millions training a model from scratch, the team took an off-the-shelf AI model from Alibaba’s Qwen and supercharged it using a technique called distillation.

Think of it like copying the smartest kid’s homework and still acing the test.

They trained s1 using answers from Google’s Gemini 2.0 Flash Thinking Experimental, a model Google offers free access to—though not exactly for this purpose.

No surprise—OpenAI and friends aren’t exactly thrilled.

They've already accused DeepSeek of improperly scraping their models for distillation, and while Google forbids reverse-engineering its AI for competing services, it's unclear if this project actually crossed the line.

Either way, you can bet big AI labs aren’t thrilled that high-end reasoning is suddenly looking this cheap.

Training s1 took just 30 minutes on 16 Nvidia H100 GPUs—the kind of hardware AI companies spend millions on.

A Stanford researcher involved in the project casually mentioned he could rent the compute today for about $20. Yikes.

Oh, and they discovered telling the AI to ‘wait’ before answering actually made it think harder and give more accurate responses. Turns out, patience isn’t just a virtue—it’s an AI cheat code.

This whole experiment raises some huge questions.

Of course, distillation isn’t a silver bullet—it lets you copy existing AI capabilities, but it doesn’t magically create something new and groundbreaking.

That’s why Meta, Google, and Microsoft are still pouring hundreds of billions into AI research.

Still, one thing’s clear: the future of AI might belong to whoever can innovate on a budget.

Click here to dive deeper into the research and its impact on the future of AI.

We’re excited to introduce The Autonomous Agency, the next evolution of The Automated.

After helping 16,000+ subscribers grow with AI-driven insights, we’re now opening up our playbook to help YOU build and scale your personal brand or business with the exact tools and strategies we’ve used.

Here’s how you can get early access and exclusive rewards:

✨ 1 Referral: Unlock exclusive Loom videos where we reveal how we built our processes for creating and managing content, automating workflows, and scaling efficiently.

✨ 3 Referrals: Earn a 50% discount on your first month with The Autonomous Agency. Let’s take your brand or business to the next level with powerful automation!

✨ 5 Referrals: Get a 30-minute one-on-one call where I’ll personally explain how the system works and how you can implement it for your brand or business.

📣 Let’s keep the momentum going—refer your friends and unlock these rewards today! The more you share, the more you gain

Google is stepping up its AI game with the official release of Gemini 2.0 Pro Experimental, and that’s not all!

The company is also integrating its reasoning-focused Gemini 2.0 Flash Thinking model into the Gemini app.

This is all to capture the attention of a market that is increasingly drawn to Chinese AI startup DeepSeek’s cost-effective models.

If you’ve been keeping up, then you know that DeepSeek’s AI models have been making waves for delivering top-tier reasoning capabilities at a fraction of the cost of American alternatives.

In fact, Google and DeepSeek released reasoning models in December, but DeepSeek’s R1 gained significantly more attention.

Now, Google is making a strategic push to showcase its own reasoning tech to a broader audience.

Google let the cat out of the bag early when it accidentally announced Gemini 2.0 Pro in the app’s changelog last week, but now, it’s official.

The company is rolling out an experimental version of the model on Wednesday, making it available through Vertex AI and Google AI Studio. Gemini Advanced subscribers can get their hands on it directly in the Gemini app.

And here’s some quick facts about it:

The rollout also includes the general availability of Gemini 2.0 Flash, first announced in December, and a new cost-efficient model, Gemini 2.0 Flash-Lite.

Google says Flash-Lite outperforms the previous Gemini 1.5 Flash model while maintaining the same speed and pricing, positioning it as a direct competitor to DeepSeek’s affordable AI options.

With these launches, Google is making it clear that it won’t let DeepSeek steal the spotlight without a fight.

Click here to dive deeper into Google’s latest AI advancements.

Analyzing a dataset helps uncover patterns and insights that might not be immediately obvious.

It ensures decisions are data-driven, making outcomes more accurate and reliable. Plus, it saves time by automating complex analysis, freeing you up for strategic work.

Here's a prompt that can help you with that.

Source

This should help streamline your interactions with ChatGPT and get the results you need more efficiently.

Best of all, It's free!

Click here to get it now.

That's all we've got for you today.

Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇

Your feedback means a lot to us and helps improve the quality of our newsletter.