Research.
High-signal reports covering today’s emergent technologies
Explore research
The race to automate human labor
Intelligence is massively parallel tree search guided by memory and logic. From evolution's glacial updates to humanity's network-speed learning, we trace how the speed of belief updates determines intelligence levels — and why AGI will emerge from millions of specialized agents, not a single superintelligence.
Delphi recently hosted a live text-based AMA on Telegram with Teknium, Co-Founder and Head of Post-Training at Nous Research. The conversation was moderated by Tommy Shaughnessy and featured questions from the Ex-Machina Telegram Group Chat.
While GenAI has taken venture investing and the public markets by storm since late 2022, it is now also quietly infiltrating private equity. The typical rollup playbook is now being infused with AI – revamping operations in traditional businesses by leveraging the rapidly evolving capabilities. This model has been pioneered in the US but now appears poised to catch on in China.
In the first part of my "Experience Era" series, I covered the jagged intelligence patterns of foundation models and how model training is entering the experience era. With intelligence now abundant thanks to the "general recipe," the bottleneck for unleashing AI's potential has shifted to applications. Part II analyzes this transformation and its profound implications.
Perhaps the most comprehensive overview of today's AI frontier interweaving Data Foundries, Reinforcement Learning, Context Engineering, RL Environments, and World Models into an easily digestible whole
Multimodal medical foundation models are the future of healthcare. We need a DeepSeek-like company to build them — so we founded Sophont to make it a reality.
Browsers are emerging as the ultimate frontend for using AI. Explore how AI browsers like Perplexity Comet, Dia, and others are reshaping our digital experience, triggering a high-stakes battle for control of the web's most critical gateway.
A philosophical exploration of what intelligence truly is—across humans, animals, and machines—arguing that our failure to understand intelligence means we're building powerful AI systems without knowing what we're actually creating.
