OpenAI and Perplexity rush out 'decision models' to compete with Jev, dev shares dead-code experiment
bendee983 · x · 2026-10-02
Blogger bendee983 claims every AI lab is scrambling to release a competitor to Jev: OpenAI has shipped a Decisions API and Perplexity has entered with its own decision model — reportedly urgent responses without Jev's months of preparation and testing. He sees limitless use cases for decision models.
He also ran a related experiment: AI-generated code often contains dead code (never-called functions that can be externally invoked and become security liabilities). A reasoning model alone missed chunks, so he had an AI build function-call graphs and traced execution graphs across modules to find unused code — more accurate, but the thread cuts off mid-sentence on its hallucination issues.
More from coding & agent
- Matt Pocock: GitHub Actions is the easiest way to start with software factories — mattpocockuk · 2026-10-02
- Daily brief: Grok 4.7 rolls out as base model everywhere, Claude Code ships Mods — testingcatalog · 2026-10-02
- Hjarni notes app ships MCP server so ChatGPT can read your notes — tehmadnezz · 2026-10-02
- Screen Studio trained on 5,000 fake macOS apps to build a UI-specific upscaler — CurieuxExplorer · 2026-10-02
- Codemode is just a subagent if you squint, argues developer Lucas Meijer — lucasmeijer · 2026-10-02
- Reddit user finds one prompt line makes Claude communicate visually — ZgemboAdislic_ · 2026-10-02