Codex reportedly fails on cryptography tasks and rewrites a Merkle tree into comparisons
chaumian · x · 2026-07-24
A reply thread says Codex is still weak on cryptography-heavy tasks and can produce very inefficient code. In one example, a simple tool built with Claude was switched to Codex, which then removed the crucial Merkle tree construction and replaced it with thousands of comparisons.
The thread also notes that this came with “other completely absurd changes,” suggesting the model made broad, incorrect edits rather than a targeted improvement.
Related event: Codex Struggles with Cryptography, Generates Inefficient Code(2 posts)→
More from coding & agent
- Dev turns DJI Mavic into local AI security drone built with Opus 5.5 — alexcovo_eth · 2026-10-27
- Microsoft Research: ForkPilot cuts long-horizon agent compute by up to 59% — EliasEskin · 2026-10-27
- Cursor launches Agent SDK: agents as code with Markdown prompts and TypeScript tools — soleio · 2026-10-27
- Vibecheck, the 'usable formal methods' hackathon, sells out two floors in SF — spikedoanz · 2026-10-27
- How do you verify AI agent outputs beyond unit tests and LLM-as-judge? — Practical_Green5775 · 2026-10-27
- AI agents may soon become the only way to run code, not just write it — IgorCarron · 2026-10-27