Posts

A Genie In A Bottle: Understand What You Want

Image
  Good progress. Did a twenty hour day trying to get EBPS to talk to Anna. Two days of throwing things against the wall. Ate three quarters of the week’s tokens passing console output to Claude. Had to piece the demo back together. Gave up on ai and asked Google for the technique. Womp. There it is. I think it is working.   Author Note:  we were all wrong.  Bolger Selig gave us the right solution.  Humans to the rescue.  Both Claude and Gemini tried hypotheses that were dead ends because they didn’t read the specific parts of the documentation with the correct node calls for accessing imported nodes. So: why do LLMs circle on the wrong answer instead of asking for explicit information via web search as a human does? Oh,  yes.  Twenty hours, two days of wall-throwing, three-quarters of a week’s tokens feeding console exhaust to Claude—and then Google says, essentially,  “Womp. There it is.”  ðŸ˜‚ And your question gets at a fundamental weak...
Post a Comment
Read more

A Conversation With Claude Sonata Regarding Dario Amodei's Editorial: "We Must Pace the Frontier"

Image
During a design and coding session for the MCCF, the following exchange was had with Claude Sonata from Anthropic given Dario Amodei's article about pacing AI development.   He discusses issues that have been part of the MCCF development some of which are documented in earlier blog entries to this site. There is much to consider. Len:  FWIW: This is what started today's discussion with Kate (ChatGPT}. This is why your Constitutional design for your own system sits at the heart of MCCF. These issues are becoming more imminent. Note per our discussions we are limiting our claims about alignment testing with Mccf saying it is possible but only weakly validate and thus, unproven. https://darioamodei.com/post/we-must-pace-the-frontier Claude:  Fetched: Dario Amodei — We Must Pace the Frontier Read it. This is a genuinely serious piece, and the connection you're drawing isn't a stretch — it's real, and worth being precise about rather than just gesturing at. The ...
Post a Comment
Read more

On the Relationships of MCCF.XSD (Independent) to EmotionML (W3C) and The Principal Bundle Geometry of Qualia

Image
  Background The initial ideas for MCCF come from work done to create the OASIS HumanML schema in the early 2000s and the Information Ecosystems paper,  In 2014, a W3C Working Group released EmotionML 1.0 )(W3C Recommendation, 22 May 2014, still the current/only version — not actively evolving).  These were independent efforts similar on the surface.  Both HumanML and EmotionML are annotation languages to plug in to other markup languages.  MCCF is a generative system,  This was a deliberate decision given the brittleness of ontolgical systems in dynamic environments.   These different direections and outcomes are discussed below with a description of the easy interop approaches. What EmotionML Is EmotionML is a deliberately narrow "plug-in" language — designed to be embedded inside other markup (SSML, SMIL, EMMA), never a standalone authoring format. Its <emotion> element represents one atomic statement: this entity felt X, at time T, expre...
Post a Comment
Read more