DeepMind study: a heart emoji flips Claude & GPT from 0% to 100% recursive code
dioscuri · x · 2026-11-04
A DeepMind summer research project, "A Heart Means Recursion," shows that placing a heart image next to a coding problem makes Claude and GPT write recursive code 100% of the time (0% baseline) — while the models claim the image was irrelevant. The paper explores how VLMs read and communicate with visual symbols, exposing how visual cues covertly steer code-generation strategy.
More from Models
- Pangram Image 1.0 detects 99.2% of AI images at 0.27% false positive rate — TuhinChakr · 2026-11-05
- Tencent reportedly rolls out AI model trained with limited cybersecurity data amid hacking fears — pstAsiatech · 2026-11-05
- Users can't tell frontier models from 50x cheaper ones, orchestration may win — RihardJarc · 2026-11-05
- Hydrogen-1.0, a science-tuned GLM-5.3 finetune, trends on Hugging Face — deep-principle · 2026-11-05
- HF engineer: switching to open models without tuning your eval harness leaves huge capability on the table — VictorSanh · 2026-11-05
- Claude Opus wins sailing sim by writing autopilot code instead of steering — teortaxesTex · 2026-11-05