AI is taking evolution for a test drive

In partnership with

Hello readers,

Welcome to the AI For All newsletter! Today, we’re talking about how AI is speed running evolution in a computational sandbox, using AI to improve OTA updates, and more!

AI in Action: Where we’re going, you don’t need eyes to see

Building off the question of why animals evolved the visual systems they did, MIT researchers have created a computational “sandbox” where artificial agents evolve eyes over generations to solve survival-like tasks. Instead of rewinding time or relying on fossil records, this simulation environment allows embodied AI agents to develop visual systems — photoreceptors, lenses, and neural processors — tailored to their goals and constraints. The result is a dynamic tool for testing evolutionary hypotheses in code, not clay.

Depending on the task, agents evolved entirely different types of eyes: navigation challenges produced compound eye analogs (à la insects), while object recognition tasks led to forward-facing camera-style eyes with more acuity. The simulated evolution is governed by genetic encodings that mirror biology: optical genes for light interaction, neural genes for processing, and morphological genes for placement and perspective. Environmental constraints — like pixel budgets — force agents to make trade-offs, just like natural organisms.

The framework offers more than just theoretical insights. It could guide the design of task-specific vision systems for drones, robots, and wearable tech, balancing performance with constraints like energy and size. As the team looks to add LLMs for more intuitive “what-if” querying, they hope the tool will broaden how scientists experiment with complex systems — not by simulating atoms, but by simulating pressures.

This is evolutionary biology reimagined as software engineering — and a glimpse into how AI might one day help us reverse-engineer nature’s greatest inventions.

Your competitors are already automating. Here's the data.

Retail and ecommerce teams using AI for customer service are resolving 40-60% more tickets without more staff, cutting cost-per-ticket by 30%+, and handling seasonal spikes 3x faster.

But here's what separates winners from everyone else: they started with the data, not the hype.

Gladly handles the predictable volume, FAQs, routing, returns, order status, while your team focuses on customers who need a human touch. The result? Better experiences. Lower costs. Real competitive advantage. Ready to see what's possible for your business?

See it in action

🔥 Rapid Fire

• Investors seek protection from risk of AI debt bust

• Oracle’s $10 billion data center in limbo after Blue Owl pulls out

• Oracle delays multiple data centers for OpenAI to 2028

• Oracle added $150 billion of data center leases in November quarter

• Data center REIT Fermi drops after tenant terminates $150 million contract

• Commentary: busting the myths of generative AI

• OpenAI rolls back ChatGPT model router for most users

• OpenAI deal to license Disney characters is entirely in stock

• How to monitor tech debt that could fuel an AI bubble

• AI’s water and electricity use soars in 2025

• AI is not having the impact on jobs that some people think

• A billion-dollar question hangs over the AI search marketing industry

• Oracle's fall from grace has made it the poster child for the AI bubble

• YouTube shuts down channels using AI to create fake movie trailers

• Amazon pulls inaccurate AI recap of Fallout show

📖 What We’re Reading

How AI Improves OTA Updates