Urgent.News

What's breaking now, across thousands of outlets.

AI

AI coder gets Doom running on a custom CPU designed by GPT-5.6 Sol — game viewport is overlaid on a pulsing schematic of the CPU in Turing Complete's sandbox environment

An AI computing enthusiast has demonstrated Doom running on a custom CPU designed by GPT-5.6 Sol.

AI coder gets Doom running on a custom CPU designed by GPT-5.6 Sol — game viewport is overlaid on a pulsing schematic of the CPU in Turing Complete's sandbox environment

An AI computing enthusiast has showcased the impressive capability of GPT-5.6 Sol by demonstrating Doom running on a custom CPU it designed. Called Codex-R32, the CPU runs within the Turing Complete game environment, where a pulsating schematic of the CPU provides a live visualization of processor gates, registers, memory, ALUs, and other components. The demonstration used the C-based PureDOOM port, compiled into native RV32IM machine code that runs directly on the custom Codex-R32 CPU.

The AI model's achievement has been hailed as truly insane, as it pushes the boundaries of what is possible with agentic coding. While some may argue that there is always room for more, the AI confidently responded to a challenge to tackle Crysis next. In response, the AI proposed building a GPU, adding gigabytes of RAM, and making the schematic visible from orbit.

It's important to note that although Doom may seem insignificant compared to more demanding games like Crysis, the achievement still showcases the model's potential and performance within the Turing Complete sandbox. The live stats offered during the demonstration, including the cycle counter, sim speed, memory/register values, and more, provide valuable insights into the capabilities of GPT-5.6 Sol's custom CPU.

Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at tomshardware.com →

More in AI

Why LLM reasoning isn't enough for medical scheduling math

I’ve seen plenty of people try to make Claude or GPT-4 act like a specialized scheduler. They prompt it heavily: "You are a precise medical assistant.

  • LLMs excel at simple scheduling but struggle with complex constraints
  • Probabilistic nature of LLMs leads to hallucinations in deterministic scheduling
  • Injection Day Alignment MCP server offers structured tools for medical scheduling

Passing Once Isn't Reliable — This Week's Agent Engineering Puts the Harness Before the Model

This digest covers AI agent developments from 2026-08-18 to 2026-08-25: orchestration patterns, tool/function calling, memory, planning loops, multi-agent coordination, and agent evaluation.

  • AgentWeave reduces tool exposure by 70% and latency by 51%
  • Pass@1 in Thinkingbox drops from 65.36% to 25.25% despite initial success
  • AutoSaddler optimizes agent harness, gains 9-10 points across benchmarks

More from Tuesday 25 August →