NPQG LangChain

2 minutes

·

·

After watching this video I imagines creating exploration machines that integrate AI LLMs, Langchain, NPQG simulators, CERN, and more, like Lee Cronin’s chemputers. Let’s brainstorm!

Imagine a combined algorithmic and experimental approach that integrates the following:

  1. a reaction to be explored or optimized
    • e.g., molecular, atomic, nuclear, standard model assembly, point potential
  2. input reactants
    • spacetime
    • photons (diffuse sources or lasers of specified frequency and phase)
    • neutrinos (are there neutrino detecting reactions we could discover / exploit?)
    • standard model, atomic, or molecular objects
    • In short, the ingredients to any reaction
  3. reaction conditions
    • Scenario chosen from the set {any reaction, anywhere in the universe} e.g.,
      • planetary (e.g., earth, mars): ranging from leading edge experimental to advanced manufacturing
      • deep space
      • stars, neutron stars, black holes, SMBH, etc.
  4. optimization and constraint factors
    • desired products,
    • yield goals,
    • safety constraints, i.e., don’t produce hazardous byproducts or harmful radiation
    • cost goals,
    • production rate goals
  5. point potential simulation libraries
    • purpose built
    • best in niche
  6. Ai LLMs
  7. LangChain
  8. perhaps Lee Cronin’s chemputer
  9. an API to CERN LHC directing experiments in real time.
  10. etc.

How does our exploration machine work, given detailed prompt information?

  • It uses existing knowledge to formulate some areas of interest to simulate.
  • It seeks new information by enumerating a distribution of reactions over the constraints
  • It simulates all those reactions
  • The results guide future explorations according to optimization and constraint factors.
  • Continue until recommendations are made.
  • Prove those reactions in real world experiments.

We formulate prompt based questions for which answers would be valuable and then we submit them to our exploration machine. Here are some of many questions that I think could be important.

  • Identify scalable processes to efficiently remove carbon dioxide from our atmosphere.
  • Identify photon-neutrino reactions that improve our ability to work with neutrinos.
  • Explore ways to improve a chip production process.
  • Design compute chips that are optimized for path based simulation in R4.
  • Find processes that can turn spacetime into water or oxygen or other atomic materials.
  • Optimize molecular nanobots that can work inside the human body to repair or augment.
  • Explore linkages that could improve advanced gene editing, e.g., CRISP-R , etc.
  • Explore ways to implement computation and memory at scales not yet imagined.
  • and so on

This idea is one that could be implemented within six months given the right team. Point potential simulation is not difficult and will soon be a reality. Everything else exists.

Perhaps the best part of our exploration machine is that we can use it to improve every part of the machine itself, for example finding improved ways to simulate various point potential simulation niches.

J Mark Morris : Lynn : Massachusetts

After watching this video I imagines creating exploration machines that integrate AI LLMs, Langchain, NPQG simulators, CERN, and more, like Lee Cronin’s chemputers. Let’s brainstorm! Imagine a combined algorithmic and experimental approach that integrates the following: How does our exploration machine work, given detailed prompt information? We formulate prompt based questions for which answers would…