Yoav Goldberg: Fable/Astra excel at long-horizon tasks via consistent memory notes and compaction
ysu_nlp · x · 2026-09-14
- Yoav Goldberg (ysunlp) elaborates on ability transfer: Fable/Astra's edge is not just accounting knowledge or ERP integration.
- Both models are materially better at extremely long-horizon tasks by maintaining a consistent memory note and effective context compaction.
- Most other models struggle with this kind of sustained coherence, he says.
More from Models
- Agent trace dataset hits 50k+ monthly downloads — author speculates on SFT and reward-hacking monitor uses — maksym_andr · 2026-09-14
- Author Finds Gemini 'Reliably Wrong' at Verifying Quote Sources, 0/2 in Tests — danbri · 2026-09-14
- Astrable: Open-Source Codex Plugin Pairs GPT-6 Astra with Claude Fable 5.1 — daniel_mac8 · 2026-09-14
- GPT-Live-1 tested on real phone calls: natural speech but serious instruction-following flaws — kolchinski · 2026-09-14
- Dan Shipper: using Astra medium for simple tasks, "pacing the frontier" — danshipper · 2026-09-14
- Grok-4.6 spotted generalizing well across ARC-AGI generations, per third-party observation — zainhas · 2026-09-14