Jon
96 posts
@Musingsaremine
Searching for missionaries on their life's work.
Very aligned. A few comments: 1. I think repaving and reinventing have a big Venn diagram overlap. I have gravitated to the word repaving to mean that the workflows that exist today in software will be reimagined via a first principals as AI stitched other with software as glue between the inputs and outputs of models. 2. The other part of this is data. Information and knowledge are both data problems. Human and systems data need to be integrated. Systems get connected with a semantic / metadata layer (aka not moving data, but integrating it so agents can access it). Here is what we did internally (openai.com/index/inside-o…) and those learnings + what our FDE teams learned are directly informing the “business context” layer of our Frontier platform (openai.com/business/front…). As for human data, that’s where the strength of the models + top down executive buy in are unlocking Enterprise customers is allowing for the accelerated mapping of things. Whether it is tribal knowledge, undocumented processes or legacy systems — today’s models (audio, visual understanding, computer use, etc) can be used as tools to ingest, break down and “understand” that human data and bring that unstructured data also into the context layer. Then you hill climb. With data connected, just like with “memory” in ChatGPT, there will be “organizational memory” for co mpanies. That will start to have a compounding effect. While we don’t have recursive systems yet, I think you start to get something that feels like. 3. Not doing this or some version of this is not an option. Some will lead and some will follow, but I think this new world we are going into is the real “digital transformation”. The internet and what got built on top of that was transformative, but I think it would look small in hindsight. However, it was an absolute prerequisite to get everything connected (as infra) to then layer on intelligence. As John Chambers used to espouse: “the network is the platform”. He was right, but also wrong where the value accrues. Companies will have no choice or they’ll just find themselves to be slow, laggards and under massive margin pressures. Things will just get “repaved” with AI as the substrate. I think we’ll see the first big waves of early adopters move at scale in 2026.
There's a fruit fly walking around right now that was never born. @eonsys just released a video where they took a real fly's connectome — the wiring diagram of its brain — and simulated it. Dropped it into a virtual body. It started walking. Grooming. Feeding. Doing what flies do. Nobody taught it to walk. No training data, no gradient descent toward fly-like behavior. This is the opposite of how AI works. They rebuilt the mind from the inside, neuron by neuron, and behavior just... emerged. It's the first time a biological organism has been recreated not by modeling what it does, but by modeling what it is. A human brain is 6 OOM more neurons. That's a scaling problem, something we've gotten very good at solving. So what happens when we have a working copy of the human mind?
