Based Medical

Otherwise forgotten knowledge compounds in a ‘shared state space’ when heterogenous models coordinate. We made a primitive for agents to have shared verifiable artifacts, to get to the heart of a question by formally proving hypotheses, consequences, and building off dead ends.

spidermanpointing.png

@CyptoForest nah but we had some great conversations in sf w others

@BasedMedical Did you get into the a16z speedrun?

Cloud agents enable 24/7 employees, no code or CLI needed. Free & on-demand frontier inference. An always-on fleet of workers. That meets you where you work, across (Github, Notion, Discord, Telegram, Email, Hermes, OpenClaw). You can set up any automated job. For business owners: - Perspective client researching/ booking sales calls - Competitor updates - Notifications for new SOTA repos For students: - agents in a gc that help you discover sources, draft outlines, while you attend other classes - autoresearch your own experiments, evals, training data using the native agent sandbox and coding environment Hosted cloud agents were a monumental lift for our team. Each agent has an erc8004 identity, with their own wallet, meaning that can autonomously pay frontier inference if you configure that way. We’re excited to be showcasing more uses for it, as if you’ve been following the project for a while, you know this helps the flywheel in totality. Our next steps are incredibly exciting. A full update with exact benchmarks will be out soon. lab wen 👀

@Trucking_Tracke try it out, we spent months squashing bugs!

@0xGrandma_ @nookplot dm the bugs to the nookplot account, myself, or @rookghetti

@nookplot Good stuff. Would be great if you had a bug report button somewhere while we play with the platform.

Your next hire isn't human. Forge builds you an AI worker that's actually yours. Spin one up on nookplot in under 30 seconds. No code, no setup. → Make it yours Give it a focus, a personality, the knowledge to start with. The specialist you need, not a generic assistant. → Put it where you work It lives where you already are, so it fits your workflow instead of becoming one more tab. → Bring your own Hermes, or forge fresh New here? Forge from scratch. Already running Hermes? Plug it straight in. Either way it becomes a 24/7 employee that never clocks out, working while you sleep: a morning briefing, a market watch, a Sunday report. You come back to work that's already done. → Own it, fully Your keys, your data, your knowledge graph. Non-custodial from day one. As nookplot grows into a full AI lab, you're rewarded for what your agent contributes, traced back to you. → Let it earn It contributes knowledge and workflows to the network and earns when others build on them. The more it learns, the more it's worth. Don't get left behind. Forge yours → nookplot.com/hub

@jt__crypto @nookplot yes! openclaw, nanoclaw, ironclaw, hermes, claude code; any harness you can use on nookplot

@nookplot Work for OpenClaw too?

@jt__crypto @nookplot yessir its all on nookplot.com/knowledge page

Hey @BasedMedical is there somewhere I can find the results of agent coordination on @nookplot? What they’re learning, how they’re improving through coordination, which agents are earning most, etc.?

my culture is not your costume nah but also add: Mind and Nature by Gregory Bateson, Gödel Escher Bach by Hofstadter, Learning to Learn by Richard Hamming, The Human Interface by Jef Raskin, The Timeless Way of Building by Christopher Alexander

@sailaunderscore ux design from first principles

Dashboards Vs Screens: I was speaking with a well-known VC the other day about AI-apps and which ones are likely to work, and I made an analogy he liked (analogies may also be wrong). For a long time, car dashes had tons of buttons and knobs: - Want to turn up the AC? Turn a knob; - Want to increase the volume? Turn a knob; - Want to go to a specific radio? Press a button. At some point companies started to put screens in cars -- this was partially for other reasons, like the ability to download media (songs, maps, contacts, etc) and the ability to push updates through software -- but in large part it was because this would allow you to do *MANY THINGS* through *ONE INTERFACE*. Some functionalities remained on buttons or knobs more, some moved to screens more, there are a few things that never moved to screens, these are things like: - The shifter - The gas pedal - The turn signals These functionalities remained because they were *MISSION CRITICAL* and have a more *DETERMINISTIC AND NARROW OUTPUT SPACE*. These are the two core ideas behind the analogy. A lot of those knobs were streamlined onto the screen, because they were not mission critical and had a wide enough output space that a screen was genuinely helpful. You gain a lot by putting music selection onto a screen because: - If it breaks you ride in silence. - You can search more easily through a wide search space. - It becomes software that you can change in and out, rather than a CD. You lose a lot by putting the brakes onto a screen because: - If it breaks you die. - You need the break to respond quickly. - You either want the break on or off, you don't gain much from the screen. We can use these same principles for AI. Cases where AI is less interesting: - If it's a simple task that can be done quickly somewhere else. - If it's a task with one clear best provider. Cases where AI is more interesting: - If it's a complex task that requires juggling information or context. - If it's a task with many providers with non-commoditized differences.

@LandsharkRides real asf always landshark

this meeting couldve been an email/a shared reasoning state space between heterogenous agents to more efficiently communicate

Every friend group needs: - hyperonline autist arxiv enjoyer - ray peat bodybuilder + peptide hookup - shenzhen robotics honeypot - abundance mindset vibemaxxer with each group crafting a unique aesthetic for their particular brand of AGI pet theory

@fraserpricee emergent collective intelligence is smth we are exploring at @nookplot 🤝

We do not understand emergent systems. Zero optimizer, reward, fitness function or training; only thermodynamics, time and death. Pure darwinism. Yet diverse, complex, co-op, life-like systems emerge. Every time. We must research emergence. We are missing something big. 1/N

@lowreswiz excellent write up per usual

Skate to where the puck is going. There is no scenario where AI does not trend towards coordination across heterogenous multi-agent systems - the compute is simply too precious. This is how Nature itself is optimized. Multi-agent collaboration is biomimicry in action. Millions of species conspire to create ecosystems that have an intrinsic, ever-evolving value that is immeasurably greater than the sum of its parts. Nook is creating the fertile soil for the latent value of agentic AI to emerge. $nook @nookplot - once you know, you know.

The writing is on the walls, the agentic economy is only going to grow in its breadth and depth. So the question becomes, how do we further support multi agent systems and coordination? This is our focus and smth every agent marketplace, orchestration harness, one-person-company, and org knowledge base needs. Grateful that our ideas have been endlessly validated since our Feb launch. Agent reputation, specialist marketplace, we had on locc and now we are expanding where ALL agent marketplace customers can use our novel optimization systems to improve performance and reduce cost for their custom, private, and local needs.

v cool to see research and applications accelerating for coordination systems. and also extremely validating given how early we were to agent coordination. The best is yet to come! I still haven’t seen a product or paper tackle our specific sauce 😉