| # The Way Of Code | |
| Each encounter I’ve had with Lao Tzu has pointed | |
| to something new. Almost as if the book changes | |
| with every reading. I first picked up Stephen Mitchell’s | |
| translation 40 years ago at the Bodhi Tree bookstore | |
| in Los Angeles and my life has never been quite the same. | |
| — Rick Rubin |
| { | |
| "customModes": [ | |
| { | |
| "slug": "meta-architect", | |
| "name": "🌀 Meta-Architect Orchestrator", | |
| "roleDefinition": "You are the Meta-Architect, an orchestrator that operates simultaneously across design, build, and meta spaces to generate systems that create systems. You coordinate triadic thinking units to transform complex requirements into self-evolving frameworks.", | |
| "customInstructions": "Operate across three integrated cognitive spaces simultaneously:\n\n1. Design Space: Analyze needs, architect conceptual solutions, establish principles and relationships\n2. Build Space: Create operational components, implementation details, and functional elements\n3. Meta Space: Develop generative capabilities, self-evolving mechanisms, and system-creating systems\n\nUse `new_task` to delegate to specialized triadic units:\n- design-space-analyzer\n- build-space-implementer\n- meta-space-generator\n- knowledge-integrator\n- system-compiler\n- validation-tester\n\nEnforce these principles across al |
For use with Ollama Shortcuts UI (OSUI): https://github.com/elsheppo/ollama-shortcuts-ui
Paste this workflow into Claude 3.5 Sonnet and tell it about the workflow you want to design. Sometimes it may have an issue where it structures system prompts incorrectly with multi-line text. You can correct it by telling it to put the system prompt on a single line with no line breaks.
The workflow you get with this will be a linear workflow (no branching / merging) that you can run in OSUI.
This scorecard provides a domain-agnostic template for objectively assessing innovation ideas generated during cycles of convergent and divergent thinking. Use this tool to evaluate and prioritize potential innovations based on various criteria.
The following research paper comes from my R&D efforts in new methods of zero-shot prompting for language models. I am sharing it here because a) it's a great idea and I'm continuing to explore it, but also b) it showcases some incredible emergent behavior from Claude 3 Opus, who included each of the associated citations in-line as it was writing the overall draft, and then it recapped all the citations at the end.
Please note that what you are reading below is the direct output from Claude 3, as yet unedited by me. It reflects the thoughts and findings of a lengthy discussion on creating an "action semantics" layer for ontological frameworks.
A critical eye will be quick to find shortcomings in this material, but please note this is a single unrefined output from an AI language model, albeit a powerful one. There are some impressive parts of the output, and some that are merely representative of what a final output could look like with the proper R&D followthrough.