The Real Bottleneck for AI Assistants is Long-Term Memory
xiaosun86 · x · 2026-07-20
The author notes that current GPT-style projects start losing context after just a two-week trip, with even previously compressed information gradually distorting, eventually causing errors with tickets and locations.
They argue that the true bottleneck for AI assistants isn't form factors like glasses, cameras, or AirPods, but the ability to handle "multi-year scale" long context—continuously remembering and retrieving the raw, multimodal data of a person's life.
More from AGI Musings
- Claude Code skill uses 10 Markdown rules to make outputs ADHD-friendly — alex_verem · 2026-07-22
- ControlAI CEO says an international ban on superintelligence is needed to avert extinction risk — zetalyrae · 2026-07-22
- Gary Marcus says LLMs still cannot really do math on their own — GaryMarcus · 2026-07-22
- Gary Marcus says LLM math skills are like knowing only a car’s engine size — GaryMarcus · 2026-07-22
- AI may make digital work infinitely leveraged while offline life gets more human — illscience · 2026-07-22
- Better AI math could save researchers time by killing false conjectures earlier — prateekj · 2026-07-22