AI models hate Python and love PHP: we asked 12 of them 1,200 times
Blog post from JetBrains
An experiment asked 12 AI models from 10 labs to produce 1,200 short programming-language hot takes, using two prompts and fresh conversations for each response. When asked to choose a language “at random,” models named only 14 languages, with Go and Python accounting for more than half of the selections; overall, 75% of the takes criticized their target. Python, Rust, and TypeScript drew the most criticism, while PHP and Java were consistently defended, and opinions about Go’s explicit error handling were evenly split. Models often focused on languages they had selected themselves, repeated familiar arguments or phrasing, and sometimes contradicted their own positions; none made C# the subject of a take. The results suggest that prompting models for controversial views produced more shared conventions than distinct perspectives, though the authors caution that “random” was only an instruction, not a verifiable sampling method, and that 50 responses per model per prompt may not capture smaller differences.
| Trend | Post Mentions | Total Month Mentions | Posts | Companies | MoM |
|---|---|---|---|---|---|
| Opus 5.5 | 25 | 245 | 65 | 39 | -61% |
| GPT-6.1 Sol | 18 | 78 | 21 | 14 | -69% |
| Gemini 4 Argon | 15 | 91 | 23 | 18 | -73% |
| GPT-6 Astra | 14 | 132 | 35 | 24 | -60% |
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.