Tag archive
งานวิจัย RRSI ของ Google Cloud AI Research: ทำให้ AI agent เก่งขึ้นโดยไม่ต้องเปลี่ยนโมเดล แค่เปลี่ยนวิธีแก้ harness
핵심 내용 AllenAI가 공개한 AstaBrief 8B는 과학 논문을 인용 기반 보고서로 자동 생성해 주는 모델이다. 기존 Claude 기반 ‘Thinking...
Drop squid rings into a frying pan and two reasonable goals — fit as many as possible, or sear as much surface as possible — turn out to be different problems with different answers. Unless the sizes obey one condition, and then not only do the goals agree: where you put each ring stops mattering at all.
The same 30 Consumer electronics and mobile brands, the same questions, every major AI engine — and a 12.2-point gap in how often a brand...
Search scholarly papers by topic or look up DOIs and get titles, authors, journals, abstracts and citation counts as clean rows you can export or call from code.
Today's digest covers seven stories that moved the industry in the past 24 hours: Google ships its...
How a fully sovereign QA stack — local Wikipedia index, distilled Qwen2.5-3B reader, and crutches...
The AI Data Satiation Point: Why "Wild" AI Text is Breaking Chinchilla Scaling Laws As we...
Beyond One-Shot Success: How AREX-2 Teaches LLM Agents to Reflect and Persevere Current...
Dated: 2026-10-01 (gate 05:13Z, old-lists 07:10Z, seeded control 09:50Z — same day, same pipeline...
Originally published on depra.ai on 15 Sept 2026. Yes, AI engines recommend different brands when...
We measured 30 Communication, email and messaging brands across every major AI engine. Slack — which was not in our sample — came up agai...