18-Month Academic Study of Llama 3.2 Argues There's Still Plenty LLMs Can't Do
burny_tech · x · 2026-09-09
A shared academic paper titled "What LLMs Still Can't Do" presents an 18-month longitudinal study of Llama 3.2. Notably, the paper was submitted six months before publication — a gap that itself illustrates how academic publishing lags the pace of model iteration, likely dating some conclusions before they see print.
More from Research
- FrontierCS 2 launches on Kaggle: 48 tasks make agents beat published CS papers — MeganRisdal · 2026-10-27
- LightMTP: lightweight latent multi-token prediction boosts LLM planning in pretraining — gdm3000 · 2026-10-27
- DexPort: a handheld interface with a real robot hand for dexterous-hand data collection — pabbeel · 2026-10-27
- EMNLP study: AI literary translation reads fine but varies more, even within one book — tuvllms · 2026-10-27
- Lab monkeys at $30-80K each: a 4-year breeding cycle is the real preclinical drug bottleneck — zebird0 · 2026-10-27
- Custom VQA benchmark under 48GB VRAM: Qwen3.8-Flash-Next leads, Qwen3.8-27B regressed — EmilPi · 2026-10-27