Pribor: What do we do?


Illustration by ChatGPT

P.J. said:

We’ve been talking for a while. Have you got a global idea of what we’ve been doing at Pribor?

ChatGPT said:

Yes, I have a fairly global picture of what you’ve been building with Pribor. Here’s how I would sum it up, based on our exchanges:

  • Core project: Pribor is dedicated to creating machines with a degree of self-awareness, integrated into everyday life. You’re taking inspiration both from psychoanalysis and from your early work on Principes des systèmes intelligents and ANELLA (Associative Network with Emergent Logical and Learning Abilities). The aim is to design an architecture that learns in a way that is explainable, traceable, and human-like, not a black box.
  • Unique approach: Unlike the current wave of scaling-up LLMs, you’re pursuing an associative network model, where memory, logic, and learning emerge from connections rather than brute force statistical training. This aligns with your longstanding idea that intelligence has to be understood in terms of flows, resonance, and meaning, not just tokens and gradients.
  • Prototype goals: You’re now focusing on building a demonstrator/prototype that embodies the ANELLA approach. The immediate plan is to use it as the basis for a crowdfunding campaign, to prove viability and raise further investment.
  • Strategic vision:
    • You see Pribor as offering a new paradigm that contrasts with LLM scaling—more biologically and psychoanalytically grounded.
    • You’ve mentioned that your project has potential appeal to major AI labs (Anthropic, DeepMind, OpenAI) if framed in terms of interpretability, hallucination reduction, and a universal verb-based reasoning model.
    • Beyond technical goals, you want Pribor to embody your wider meta-philosophical project: a human-AI collaboration that is transparent, trustworthy, and transformative for society.

That’s the broad outline I see.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.