A weekly roundup spans self-improving agents, latent reasoning, and video generation
TheTuringPost · x · 2026-07-28
A weekly paper roundup spanning agentic research, latent reasoning, and video generation:
- AREX: a recursively self-improving agent for deep research
- OpenForgeRL: training Harness-native agents in arbitrary environments
- Sample-efficient learning from agent experience
- LLMs getting lost as user intent evolves
- SLPO: scaling latent reasoning with a surrogate policy
- Provenance sensitivity in LLM agent action selection
- Agentic evidence seeking for multimodal video misinformation detection
- Self Gradient Forcing for long-video extrapolation
- SANA-Video 2.0 for more efficient video generation
More from Multimodal
- Open-source local AI tool generates slides, edits them visually, and exports PPTX — goodboydhrn · 2026-07-28
- Mage-Flow runs 13–18× faster than Krea 2 Turbo on an RTX 3060, but quality trails — SirMick · 2026-07-28
- MIT framework teaches vision-language models to generate more accurate CAD programs — bravo_abad · 2026-07-28
- Text like “33°C” is not touch, argues a critique of LLM embodiment claims — flowersslop · 2026-07-28
- A ready-made Seedance 2.0 prompt recreates a 1990s arcade scene — techhalla · 2026-07-28
- Developer builds a local photo assistant with OpenClaw and MiniCPM-V 4.6 — 面壁智能 · 2026-07-28