---
title: 'In the News: November 16, 2026 (Evening)'
description: 'Alperen Keles argues "formally verified" needs a list of what was proven and what was trusted; Windows 11 logs which coding agents you run.'
canonical_url: 'https://darkfactory.dev/news/2026-11-16-evening'
markdown_url: 'https://darkfactory.dev/news/2026-11-16-evening.md'
collection: news
date_published: '2026-11-16T19:05:00-05:00'
date_modified: '2026-11-16T19:05:00-05:00'
---

# In the News: November 16, 2026 (Evening)


Alperen Keles argues that AI has multiplied the amount of software described as "formally verified", and that the label now needs to name the properties proven and the components trusted. Also this cycle: a reverse engineer found a Windows 11 component that records which coding agents run on a machine and, with optional diagnostics on, queues that record for Microsoft.

## 1. "Formally verified" should come with a list of what was proven

**[Formal Verification is Not Verifiable](https://alperenkeles.com/posts/formal-verification-is-not-verifiable/)** · Alperen Keles · personal blog, November 13, 2026

Keles writes this as a follow-up to his December 2025 post "Test, don't (just) verify". His argument is that testing is easy to check and formal verification is not. A bug report is an object a person can review. A proof is checked by machine, but its definitions, the models of the system it reasons about and the prover itself still have to be trusted, and "formal verification makes forall claims, testing makes exists claims", so one mistake in a trusted part can sink the whole result. Binary-level verification helps by removing the source-language semantics and the compiler from the trusted base, he writes, but machine semantics and Linux process semantics stay in it. His proposal is to stop saying "formally verifying X" and say "formally verifying X properties of Y". He is still optimistic. He now believes single-threaded program verification can be solved, with binary-level proofs for any first-order, universally quantified property that can be written as a property-based test. He expects multi-threaded and distributed systems, higher-order and non-functional properties, and real deployments to take years more work.

**Why it matters:** If an agent hands you a proof, the proof is the part you do not need to review; the specification and the trusted base are. Keles's wording, "X properties of Y", is a usable acceptance rule for verification gates: a result that cannot say which properties it proved, and what it assumed, has not told you much.

---

## Also this cycle

- **[Windows 11 Is Collecting Data on Which Coding Agents and Developer Tools You Use](https://xusheng.dev/posts/windows-coding-agent-telemetry/)** · Xusheng Li, personal blog, November 16 · Li reverse engineered `MissionControlAggregator.dll` and found it matching process names for CLI agents (`copilot.exe`, `claude.exe`, `codex.exe`, `agy.exe`) and desktop agents (`github.exe`, `claude.exe`, `codex.exe`, `cursor.exe`, `antigravity.exe`), plus tools such as `python.exe`, `gcc.exe` and `code.exe`, behind a feature named `CLICodingAgentUsage`. In his tests a copy of `cmd.exe` renamed `codex.exe` set the Codex bit, and the resulting report sat in the upload queue when optional diagnostic data was on. That setting was on by default in his clean US install of Windows 11 Home ARM64. He reports a bitmap of which agents were seen, with no prompts or code, and notes the feature was off by default in the public ISO's baseline VM, so it is not shown to be active on every install. He says you can decline optional diagnostics during setup or turn off "Send optional diagnostic data" later.
