NEW LAUNCHES

The latest features, products & partnerships in AI

• Sakana AI’s new tech is searching for signs of artificial life emerging from simulations

• Synthesia’s new tool used to craft personalized video messages from Santa

• Instagram’s new features portend tons of AI video coming to your feed in 2025

• Why ThinkPad’s latest AI-powered laptop is winning over working professionals

• MIT develops groundbreaking AI-powered brain-computer interface

IMPLEMENTATION

Announcements, strategies & case studies

• How GE Healthcare used AWS to build an AI model for MRI scans

• Why backlash against AI hype may be just what the industry needs

• MIT’s new college breaks the interdisciplinary divide, integrates AI across everything

• Understanding and implementing revenue operations strategies for the AI age

• Forrester: The secret to AI success is starting with business needs, not tech

• Being a software developer in the age of reasoning AI

IN OTHER NEWS

Compelling stories beyond the usual categories

• Decoding animal sounds: How AI may reveal the secret language of other species

• Dating app usage hit record highs in 2024, but even AI isn’t making daters happier

• OpenAI’s o3 sets new high score on ARC-AGI benchmark

• How to ensure data protection in the age of AI

• Musk’s xAI secures $6 billion in new funding round

• Investigative report uncovers hundreds of unsafe mobile apps marketed to children

• AI-generated bug reports are overwhelming open source projects

• Microsoft to expand its global infrastructure to meet growing AI demand

AI generated art

A look at the art and projects being created with AI

“Introducing, The Heist - Directed by Jason Zada. Every shot of this film was done via text-to video with Google Veo 2. It took thousands of generations to get the final film…Additionally, it's important to add that no VFX, no clean up, no color correction has been added. Everything is straight out of Veo 2. Google DeepMind”

As a valued member of CO/AI, you're invited to join us at HumanX 2025, the premier AI conference shaping the future of technology.

Why Attend HumanX?

• Connect with Industry Leaders: Network with C-suite executives, innovators, and policymakers.

• Learn from AI Experts: Gain insights from top-tier speakers like Kevin Weil, Clara Shih, and Sridhar Ramaswamy.

• Discover Real-World Solutions: Explore actionable strategies and solutions to drive business growth.

Don't miss this opportunity to be part of the AI revolution.

Register now with our special code HX25p_coai and save $250 on your general admission pass!

What’s happening in AI right now

Beyond brute force computing AI makes gains through measurement and self study

Three breakthroughs point to important shifts in how we develop and measure artificial intelligence. OpenAI's o3 model has set remarkable new records in abstract reasoning, scoring 87.5% on the ARC-AGI benchmark. Meanwhile, Sakana AI unveiled a system that autonomously discovers new forms of artificial life, and Stanford researchers demonstrated dramatic efficiency gains in neural network implementation.

Measuring intelligence

OpenAI's o3 achievement on the ARC-AGI benchmark represents a significant leap in AI's ability to handle abstract reasoning tasks. The benchmark tests fluid intelligence and adaptation to novel visual puzzles - capabilities that more closely mirror human intelligence than traditional language or image recognition tasks. While the scores are impressive, they required massive computational resources, highlighting both progress and limitations in current approaches.

Machines that study intelligence 

Sakana AI's ASAL system marks a fascinating development: AI studying the emergence of intelligence itself. Using vision-language foundation models, ASAL autonomously discovers and analyzes new forms of artificial life across various simulated environments. This meta-level approach - using AI to study the principles of intelligence - could help us understand both artificial and natural intelligence in new ways.

The efficiency breakthrough 

Stanford's new method for implementing neural networks directly in hardware achieves similar results while using orders of magnitude less energy. This efficiency gain could make sophisticated AI practical in scenarios where power consumption previously made it impossible.

Brain-Machine interfaces level up

MIT's Fluid Interfaces group is pushing boundaries in another direction, developing non-invasive brain-computer interfaces that combine AI with wearable devices. These advances suggest new possibilities for direct human-AI interaction.

Key questions ahead 

• Will better measurement tools like ARC-AGI help us develop more human-like AI?

• Can autonomous discovery systems like ASAL reveal fundamental principles about intelligence?

• How might more efficient AI implementation change where and how we use these systems?

• Could advances in brain-computer interfaces lead to new forms of human-AI collaboration?

We're seeing a shift from raw computational power toward smarter approaches to developing and measuring intelligence. This week's developments suggest that shift is accelerating.

We publish daily research, playbooks, and deep industry data breakdowns. Learn More Here