A response to the Day 1 paper by Addy Osmani, Shubham Saboo, and Sokratis Kartakis (Google, 2026), from the people building [&], OpenSentience, and TRVM.
"Generation is solved. Verification, judgment, and direction are the new craft."
A response to the Day 1 paper by Addy Osmani, Shubham Saboo, and Sokratis Kartakis (Google, 2026), from the people building [&], OpenSentience, and TRVM.
"Generation is solved. Verification, judgment, and direction are the new craft."
In 1247 AD, a group of 47 scholars from different monasteries across Europe discovered something extraordinary while translating ancient texts. They noticed that certain arrangements of symbols, when read in specific patterns—not just left to right, but in spirals, zigzags, and layered meanings—unlocked deeper layers of human consciousness.
They called themselves the "All Knowing 47" and adopted the cryptic identifier "AK47" centuries before the weapon would bear those same letters. To them, knowledge was the ultimate weapon.
design/review a computational model coded in elixir macros that has movable types and navigational DSL where instead of Von Neumann instruction + data packets there are condition + data packets, pattern + data packets, and/or additional packet flows (?)
https://claude.ai/chat/c7571d57-00c8-4b54-a244-3e0489d76165
review this, how cool is it? then: let's start with phase 1 a minimal VM
| defmodule EnhancedADT.Actors do | |
| @moduledoc """ | |
| Actor model system with navigable DSL for EnhancedADT modules. | |
| Provides hierarchical navigation instead of link-based references. | |
| """ | |
| defmacro __using__(_opts) do | |
| quote do | |
| import EnhancedADT.Actors.DSL | |
| import EnhancedADT.Actors.Navigation |
| defmodule EnhancedADT do | |
| @moduledoc """ | |
| Enhanced ADT system combining the best of both worlds: | |
| - Complete type system from ADT | |
| - Advanced recursion and state management from BenBen | |
| - Powerful fold/bend operations | |
| """ | |
| defmacro __using__(_opts) do | |
| quote do |
| // https://claude.ai/chat/fc7d33de-d7da-4d4a-9d8c-745ba021047b | |
| // Enhanced reduction with unified kernel and comprehensive error handling | |
| void reduce(uint32_t max_steps = 1000) { | |
| if (error_context.has_error()) { | |
| std::cerr << "Cannot reduce: VM in error state - " << error_context.get_error_message() << std::endl; | |
| return; | |
| } | |
| auto reduction_start = std::chrono::high_resolution_clock::now(); |
| -- https://claude.ai/chat/93a32ec2-3558-4662-aa9f-8241583007cf | |
| {-# LANGUAGE OverloadedStrings #-} | |
| module VMCompiler where | |
| import Data.Text (Text) | |
| import qualified Data.Text as T | |
| import Data.Word | |
| import Data.Map (Map) |
| // https://claude.ai/chat/cedd3660-ee18-41f2-a965-27472a4eac38 | |
| const std = @import("std"); | |
| const print = std.debug.print; | |
| const ArrayList = std.ArrayList; | |
| const HashMap = std.HashMap; | |
| const Allocator = std.mem.Allocator; | |
| const Thread = std.Thread; | |
| const Mutex = std.Thread.Mutex; | |
| const Condition = std.Thread.Condition; |
| Benchmark Prompt: | |
| --- | |
| assume i'm starting from nothing and trying to get to something. let's say 0 -> 1 or A -> Z. this vector is completely different than something to something. let's 1 -> 2 or X -> Y. assume you know the secret algo behind good books like Dr Seuss and Alice in Wonderland is bc of limits. now pick up your story card and explain to yourself why you think the foundation of this building is concrete & rebar OR a grid of pixels OR an inter planetary file system OR a tree of leaves OR w/e it is that you see. give yourself logical agile and waterfall checkpoints that explain where everything is going, for what reason, and why. | |
| --- | |
| // ANSWER | |
| ======== | |
| Alright, let’s break this down like a story card in an agile system, but layered with waterfall checkpoints to give a structural foundation. |