drexler

@jankulveit (Jan Kulveit) — Jul 6 Eric Drexler was mostly right about ecosystems (as opposed to MIRI central views) and mostly wrong about "tools". The problem is 'agents' are a highly convergent solution. Evolution also does not somehow intrinsically want agents: genes want a tool, a design stance system, to replicate themselves. Yet the convergent solution are agents. Humans want to coordinate, a design stance non-agenty systems like contracts... and somehow the 'tools' often end up having the shape of an agent-like organization. And so on. Sure, you can engineer whatever, but the engineered solutions live in a competitive landscape (compare: you can also engineer cubical submarines). When ML research stumbled upon the most non-agenty edge of active inference systems - pure predictor LLMs - the next quest which almost every serious competitor went on is 'how we can make them more agent-like', and what everyone is competing on now is the horizon of autonomy. > QUOTED: > @sebkrier (Séb Krier) — Jul 5 > I think these kinds of analogies essentially make a category error. It's a mistake to treat an AI as some sort of persistent situated entity with goals as one would a different species. A lion is a product of Darwinian selection, an AI is not; ... [truncated]

Dr Singularity ✓ @Dr_Singularity I just asked Fable 5 to help design a functional molecular assembler (replicator) from the ground up. A complete engineering blueprint: architecture, subsystems, components, manufacturing methods, assembly process, control software, and detailed diagrams. [Embedded document images, four panels visible, containing diagrams and text such as:] "This document specifies a positional-mechanosynthesis nanofactory in the tradition of Drexler, Merkle, and Freitas: a machine that builds objects by placing atoms and molecular fragments under mechanical control, the resulting parts joined by convergent assembly. It is not a matter-from-energy replicator (that violates physics). The core reaction was first demonstrated in the lab in 2025-2026; the full machine has never been built. Claims are tagged [DEMONSTRATED], [NEAR-TERM], or [THEORETICAL]." Contents: 0. Purpose and scope; 1. Design philosophy and system architecture; 2. Physical layout and subsystems; 3. Core operation: positional mechanosynthesis; 4. Component catalogue (bill of materials); 5. Manufacturing methods (how to build the machine) "1. Design philosophy and system architecture — The nanofactory is organised as a directed pipeline wrapped in two crossing-service layers. Feedstock flows left to right through the functional stages; a control and computation layer drives the active stages from below, and a metrology and error-correction layer watches them from above. Power and a shared clock are distributed underneath everything." Diagram: "Assembler module — UHV enclosure, 4 K, vibration-isolated" with boxes: Intake and purifier, Assembler bay (arm array + tool turret), Convergent assembly (part stacking), Control unit, Feedstock buffer, Egress port. 1:39 PM · Jul 1, 2026 · 20.6K Views [reply] 30 [retweet] 42 [heart] 364 [bookmark] 115 [share icon] Relevant ⌄ View quotes > Dr Singularity ✓ @Dr_Singularity · Jul 1 [Further embedded diagram panels, partially cut off at bottom of screenshot, showing "2.2 Tooltips and the tool library", a "Build cycle (per unit placed)" table with columns Component/Role/Material/Approx. count, "2.5 Metrology and error correction", "3. Core operation: positional mechanosynthesis", a "BOOTSTRAP" callout box, and "6. Assembly process and scaling"]