The Science of Super-Intelligence (redux)
Details
Please join me in welcoming back Marcus Abundis. Since his 2023 talk, he has continued working toward practical Super-Intelligence. This comes at the same time as Yann LeCun’s exit from Meta and announced world-model centric startup.
In Marcus' own words:
"Thank you for your interest in today’s talk: The Science of Super-Intelligence, Part 1 of 3 – initial principles and primitives.
For background, this follows my prior Jan 15, 2023 talk: The Advent of Super-Intelligence An early version of today’s talk was given at Exeter University’s Sept. 2025 Super-Intelligence conference (SiC25), where I hope to offer further notes at SiC26.
My approach highlights human creativity (adaptive logic), alongside Nature’s own deep creativity (broad ‘causal Entropy’). I thus call this work Natural Informatics, akin to natural philosophy as the origin of all science. My project surpasses naive sci-fi fear-mongering and targets an informatic 'insight engine' – ‘Leo’, DaVinci Engine – to aid human innovation and discovery as a ‘structured/extensible scientific vehicle’, starting with bare ‘principles and primitives’.
To strengthen this work, I need your questions! Posing a ‘structured science’ beyond Wolfram’s "A New Kind of Science", is only the start. ‘Communicating that structure’ is a wholly different matter and where you can help! Thus, there are no bad questions as all help me understand how the material is received and understood (or not). There are no ‘experts’ here as we are all learning about this new area of research together. So please ask away!
If you have questions beforehand, you can email me at 55mrcs@gmail.com. Additional personal information on my background is found here.
And again, I thank you for your interest!"
Here is a paper by Marcus that was submitted to the ICLR 2026 Workshop on Foundation Models for Science: