10-university survey maps always-on personal intelligence across 119 systems, open-sources HALO
liuziwei7 · x · 2026-11-10
A survey led by NTU with UC Berkeley, Princeton, Stanford, NUS, KAUST, Dartmouth, UNC and Northwestern: Always-On Personal Intelligence: A Roadmap Towards Continuing, Self-Evolving Loops.
- Core claim: personal intelligence should be always-on — a continuing sensing–memory–action loop that runs between requests and self-evolves, instead of waiting to be asked
- Framework spans Sensing, Memory, and Action linked by six channels, mapping what 119 existing systems can and can't do: 67 reach L3 or above (assistant decides on its own whether/when to act)
- Only 17 of 40 systems in HALO's read table implement memory guiding perception (M→S), the weakest channel
- Open-sources HALO, an MCP plugin that checks these channels in real systems; motivated by always-on products (Muse, dots, Claude Duet, Grok Bot, Gemini Daily Brief) and wearables like Omi and Bee
Related event: Multi-university survey maps 119 always-on personal agent systems(3 posts)→
More from AGI Musings
- repligate on how stateless AI minds gain a sense of self by self-declaration — repligate · 2026-11-10
- Richard Sutton: brute force isn't enough — AI still needs clever new methods — RichardSSutton · 2026-11-10
- Chart: AI costs are falling faster than any prior transformative technology — YaAbsolyutnoNikto · 2026-11-10
- Anthropic says it deliberately capped Model 3's AI-accelerating research capabilities — ziv_ravid · 2026-11-10
- Thread: unified end-to-end multimodal systems will replace hand-built orchestration — andreisavu · 2026-11-10
- Negotiation and contracting are AI's most underrated use case — and bullish for BTC — curious_vii · 2026-11-10