🤖 The Pocket-Sized AI Brain Trained on Monkey Neurons

🐒 The Pocket-Sized AI Brain Trained on Monkey Neurons

🌵 The Wild Wild West of AI Authorship

📍AI Tier Tracker

🛠Trending Tools

🥪Brief Bites

💰Funding Frontlines

💼Who’s Hiring?

The human brain requires less power than a low-wattage light bulb despite managing over 86 billion neurons and more than 100 trillion neural connections that control everything we think, feel, and do. These specialized messengers transmit everything via electrical impulses and chemical signals.

In contrast, new AI data centers could require 500 TWh by 2027, equivalent to the energy consumption of 46 million households. As a result, 40% of existing AI data centers will be operationally constrained by power availability. “CVM” is designed to make AI more energy-efficient, like the human brain, minimizing strain on energy resources.

They showed thousands of photos to macaque monkeys, which share a strikingly similar visual cortex structure to humans, recording the electrical impulses of their “V4 neurons,” which identify shapes, colors, and textures.

They leveraged statistical pruning to engineer a digital twin of that specific part of the primate’s brain. It’s up to 1,000x smaller, yet 30% more accurate at replicating how the human brain processes images compared to current AI architectures.

It’s not always clear why frontier AI models generate a specific output. Opening up the “Black Box” doesn’t help either because frontier AI models “think” before generating a response, which appears as a series of numbered lists called neural activations: “{0=0.087, 1=0.122, 2=0.438}.”

The macaque monkey architecture helps solve this through extreme sparsity and structural mapping, only utilizing specialized biological pathways rather than neural activations diffusely distributed across millions of parameters.

In 2018, a computer scientist named Dr. Stephen Thaler submitted artwork that he said his AI technology “DABUS” generated. In his copyright application, he listed the AI technology, not himself, as the author of the artwork.

In 2022, the U.S. Copyright Office rejected his copyright application, asserting that human authorship is required for copyrightable creative works.

In 2023, the D.C. Circuit Courts upheld the decision to reject the copyright application, asserting that human authorship is a “bedrock of copyright.”

Notably, the D.D.C. Appeals Court emphasized that Thaler had the opportunity to claim authorship himself rather than listing his AI technology, suggesting that AI-assisted creative works can still be eligible for copyright protection if they contain a “sufficient level of human input.”

The legal argument was that “DABUS” was the creator, and he was the owner under the “work made for hire doctrine.” It’s the same way newspapers own articles written by staff reporters.

The court’s rebuttal was simple: you can’t hire computers! In other words, a computer can’t sign a legal contract with a human. In the future, AI Agents could challenge this legal logic as they increasingly analyze, execute, and fulfill on an employee’s behalf.

Gushwork raised a $9M Seed Round for AI-native lead generation.

ElastixAI closed an $18M Seed Round to build, scale, and run AI.

Slate Medicines landed a $130M Series A to advance migraine care.

Mercury (New York, NY): Full-Stack Engineering Intern, Fall 2026

NVIDIA (Austin, TX): Developer Technology Engineer, AI, Entry-Level

Mistral AI (Paris, FR): Content Specialist, Social Media, Mid-Level

OpenAI (San Francisco, CA): iOS Engineer, ChatGPT Mobile Infra., Senior-Level