VERSES® is now trading on OTCQB:VRSSF and CBOE:VERS
In this month’s newsletter:
Welcome to our October newsletter.
Last week, Richard Sutton—the father of reinforcement learning—spoke with Dwarkesh and reignited debate around his “bitter lesson”: Real progress in AI comes from methods that scale with computation. Stop trying to encode knowledge by programming what you think you know. Build systems that learn on their own.
That principle gave us today’s large language models. They ditched hand-coded rules for massive datasets, but now they face a new bitter lesson. Once trained, they freeze and can’t learn from lived experience. They imitate, but they don’t adapt. Biological intelligence does.
At VERSES, we’ve focused on solving that problem. Our work on active inference shows how models can learn in real time, using far less data and computing power.
This month, at the International Workshop on Active Inference (IWAI), we are joining other leaders working toward the next generation of adaptive AI and showcasing some of the breakthroughs we’ve made.
Sutton’s reminder is timely: true intelligence won’t come from bottling what we know. It will come from building agents able to discover for themselves.
Gabriel René
Founder and CEO, VERSES
In a recent Financial Times feature, Karl Friston and Gabriel René share how VERSES is charting a new path for AI beyond brute-force scaling. Friston and René point out that, as generative models hit their limits, true intelligence demands active inference—AI that can predict, adapt, and learn from real-world experience.
Karl Friston appeared on the Machine Learning Street Talk podcast to discuss the Free Energy Principle—a unifying theory of life, mind, and intelligence. He explained how the brain continually predicts reality, and why understanding this process is key to building truly intelligent systems and advancing human well-being.
How Your Brain Predicts Reality - Top Neuroscientist
This month. Karl Friston breaks down some big ideas. Starting with the “bitter lesson,” he explains that while today’s AI models are experts at mimicking what we say, they don't truly understand the world. They lack a sense of cause and effect.
This is where our approach, active inference, changes the game. Instead of just chasing rewards, our AI systems build an internal model of their world to learn and act. This approach is paying off: a recent VERSES demo showed a 60% performance boost while using only 3% of the computing power of other leading AI methods. This efficiency is possible because VERSES AI's single objective is to reduce uncertainty, making it naturally curious and intelligent.
Karl also shared some insights into the future of the field, highlighting the significant growth of the International Workshop on Active Inference (IWAI). He noted the strong community feel, which he likens to an “extended family.” The next major breakthrough, he predicts, will be in understanding human reasoning, particularly what’s called System 2 thinking—our ability to reason deliberately. He'll be giving a keynote at IWAI offering a retrospective on active inference and hosting a fireside chat with renowned AI researcher Gary Marcus to discuss these breakthroughs.
Ready to dive deeper? Watch the full video to hear Karl's thoughts on the future of AI and the groundbreaking work happening at VERSES.
You can also catch up on everything we are doing on our subreddit.
This month, we are highlighting IWAI 2025—the 6th International Workshop on Active Inference, taking place October 15–17 in Montreal, Canada.
It marks the first time the event will be held in North America, and VERSES is proud to be a sponsor. IWAI was founded by our own Tim Verbelen, and this year’s co-chair is VERSES’ researcher/leader Mahault Albarracin.
We spoke with Tim Verbelen, Director of the Intelligent Systems Lab at VERSES, to discuss the international workshop he launched five years ago to advance the science of active inference. You can find the interview on our website here:
“This year will be incredible. It's the first and only time in North America; Karl will be there in person. We have made sure to have multiple focuses, including Industrial applications, clinical applications, robotics, ML and theoretical advances.”
Mahault Albarracin Ph.D., Director of Research Strategy and Product Integration at VERSES and General Chair of IWAI 2025
Certain information included in this newsletter contains statements that are forward-looking, such as statements related to our flagship product, Genius, the commercial application of Genius and our research developments, as well as developments in the AI sector. Such forward looking information involves important risks and uncertainties that could significantly affect anticipated results in the future and accordingly, such results may differ materially from those expressed in any forward looking statements made by or on behalf of VERSES.