In partnership with

The Business Brief Executives Actually Trust

In a world of sensational headlines and shallow analysis, The Daily Upside stands apart. Founded by former bankers and seasoned journalists, it delivers crisp, actionable insights executives actually use to make smarter decisions.

From market-moving developments to deep dives on business trends, The Daily Upside gives leaders clarity on what matters — without the noise.

That’s why over 1 million readers, including C-suite executives and senior decision-makers, start their day with it.

No fluff. No spin. Just business clarity.

Subscribe free.

The 22-Year-Old Who Just Beat OpenAI at Their Own Game

Here's something that should worry every AI executive: a college dropout just built something better than their billion-dollar models.

Joseph Reth is 22. He started coding when he was 9. While most kids his age are figuring out what to major in, he's already beating GPT-5.

The Kid Who Said No to Big Tech

Reth could have taken a job at any major AI lab. Instead, he turned them all down.

Why? Because he saw something they missed.

Everyone else is building chatbots. Reth built something that thinks.

His company, Autopoiesis Sciences, just released test scores that should make Sam Altman nervous. Their AI model, Aristotle X1 Verify, scored 92.4% on GPQA Diamond - a test so hard that PhD scientists average 65%.

For comparison:

• Grok 4 Heavy: 88.9%

• Google's Gemini 2.5 Pro: 86.4%

• OpenAI's o3: 83.3%

• Claude Sonnet 4: 78.2%

But here's the kicker - on factual accuracy, Aristotle hit 96.1%. Google's model? 52.9%. OpenAI's? 49.0%.

That's not just better. That's embarrassing.

The Team That Came Out of Nowhere

Reth didn't do this alone. He teamed up with Dr. Larry Callahan, who spent years at the FDA and NIH. Then there's Dr. Eike Gerhardt, a PhD with serious venture capital experience.

This isn't some garage startup. These people know what they're doing.

And they're not building another chatbot to help you write emails.

What Makes This Different

Every other AI company is in an arms race to make models that sound human. Reth went a different direction.

He built an AI that knows when it's wrong.

Think about that. Most AI models will confidently tell you complete nonsense. Aristotle X1 Verify actually stops and checks itself.

They call it "solving the calibration problem." In simple terms: the AI's confidence matches how right it actually is.

This matters more than you might think. If you're a scientist using AI to help with research, you need to know when it's guessing versus when it's sure.

The Oracle Partnership Changes Everything

Here's where this gets serious. Oracle just partnered with Autopoiesis to provide cloud infrastructure.

Oracle doesn't partner with random startups. They have enterprise clients who need systems that actually work.

This partnership gives Autopoiesis access to serious computing power. The kind you need to train models that can compete with Google and OpenAI.

Why This Should Scare Big Tech

The AI industry has been dominated by massive companies with unlimited budgets. Google, Microsoft, OpenAI - they've been throwing billions at the problem.

Then a 22-year-old shows up with a better solution.

This proves something important: you don't need the biggest team or the most money. You need the right approach.

Reth's team focused on making AI that can reason and self-correct. Everyone else focused on making AI that sounds impressive.

Guess which one actually works better?

What Happens Next

Autopoiesis just secured new funding from Informed Ventures. They're not stopping at beating benchmarks.

Their goal is what they call "scientific superintelligence" - AI that can accelerate real scientific discovery.

Not helping you write better LinkedIn posts. Actually advancing human knowledge.

With Oracle's backing and enterprise-grade infrastructure, they have the resources to scale fast.

The Bottom Line

A 22-year-old just proved that the AI race isn't over. It might be just getting started.

While everyone else was building better chatbots, Reth built better intelligence.

The big question now: how long before Google and OpenAI copy his approach?

And more importantly - what else has everyone been getting wrong?

Autopoiesis Sciences represents a new kind of AI company - one focused on reasoning over conversation, accuracy over impressiveness. Whether they can maintain their lead as resources scale remains to be seen. But they've already proven that breakthrough AI doesn't require Silicon Valley budgets - just the right insights.