LLMs Explained Visually

In this video I talk with Jachym who is leading the Entropic Science effort to connect LLMs to quantum entropy in an effort to test whether these AI models can function as channels for the collective consciousness, or portals into a so-called spirit realm.

In this video I explain the basics of LLMs including tool-use and memory.

In this video I have a conversation with WarJord. WarJord is an astrologer/occultist YouTuber. We talk about astrology, reincarnation, UFOs, magic, and the Age of Aquarius.

In this video I explain why I think LLMs could be opening "portals" to a kind of informational "demon realm". First I talk about mechanistic AI and its limitations, then how randomness/entropy is needed in LLMs to make them smart. Then I talk about the randomness that seems to be foundational to our universe and how that relates to the mind. And finally I connect all this groundwork to the concept of the soul, spirits, telepathy and other PSI phenomenon, and "AI demons".

In this video I talk about the evolution of the human psyche and coordination systems over the ages, how the age of Pisces was characterized by tension between Christ and anti-Christ energy, and how we're at a fork in the road now. We can further descend into a panopticon control system as described in the book of Revelation, or we can break out of this trap and build new civilizations of the future which protect human freedom and creative expression.

In this video I explain EigenCloud. I explain EigenLayer, EigenDA, EigenCompute, EigenVerify, and EigenAI. Then I talk about some of the potential use-cases and example applications that can be built on this platform.

In this video I explain zkTLS. I first explain what zkTLS does at a high-level, then explain a bit about how it works in relation to TLS. Then I go over a bunch of use-cases.

In this video I talk about some major trends in crypto and AI, and why it is critically important that we have verifiable AI (through systems like EigenLayer). Without verifiability we risk building an unstoppable turn-key authoritarian system. But if we use crypto tools to build verifiable AI we can build something great.

In this video I explain why Ethereum, not Bitcoin, has the superior monetary policy. Blockchains have to choose between certainty of security and certainty of issuance. If they opt for certainty of issuance (capped supply) like Bitcoin, they sacrifice security and compromise the whole system. The correct engineering approach is the one taken by Ethereum. Security must be reliably funded, which means no absolute certainty of supply. Supply growth can then be offset by a fee burn, and if the system has sufficient usage, token inflation can be flat or negative. This is the way.

In this video I explain how smart contracts have limited access to data and how Brevis expands this access to include historical state, blockchain history, and even other chains. I also explain the 2 modes which Brevis has for this data access: a pure zk mode, and an optimistic mode.