AI code review

Reality Check

“Loser Thinking”: Inside Proton’s AI Push, What Is Confirmed, and What It Means for Your Data

A plain white envelope held shut by a small steel padlock on a bright desk

Proton’s CEO reportedly called a request to pentest new code “loser thinking” as the privacy company pushes staff hard towards AI. What Proton itself confirmed, what rests on 404 Media’s reporting, why Lumo’s privacy is a promise rather than math, and what companies and users should check now.

Thorsten Meyer
October 30, 2026

View Post

Reality Check

AI Agents Now Start 4 in 10 Pull Requests on GitHub. The Bottleneck Has Moved to Review

A long line of white paper folders on a conveyor belt leading to a small wooden desk with a brass lamp and a magnifying glass

GitHub’s Octoverse 2026 says more than 40% of pull requests in September started with an AI agent’s first draft, up from 17% in March. What GitHub actually counted, why a 90% merge rate is not a quality score, what Copilot code review Max, approvals and Fusion cost, who answers for agent code that ships, and what it means for companies and developers.

Thorsten Meyer
October 29, 2026

View Post

Reality Check

FrontierCode: Would a Maintainer Merge Your AI’s Pull Request?

Glowing glass gateway with streams of code passing through and a luminous branch growing above it

FrontierCode grades AI patches like a maintainer would. Claude Opus 5.5 leads at 52.2% but fails 42% of attempts; GPT-6.1 Sol scores 49.1% for $0.81 per task against Opus’s $10.15.