The rapid adoption of features like remote control and scheduled tasks by Anthropic, Perplexity, and Notion is not about copying the open-source OpenClaw project. Instead, it marks the industry's recognition of a new set of fundamental "primitives" for agentic AI: persistent, remotely accessible, and autonomous operation. These are becoming the new standard for AI interaction.

Karpathy identifies a key missing piece for continual learning in AI: an equivalent to sleep. Humans seem to use sleep to distill the day's experiences (their "context window") into the compressed weights of the brain. LLMs lack this distillation phase, forcing them to restart from a fixed state in every new session.

AI agents like OpenClaw learn via "skills"—pre-written text instructions. While functional, this method is described as "janky" and a workaround. It exposes a core weakness of current AI: the lack of true continual learning. This limitation is so profound that new startups are rethinking AI architecture from scratch to solve it.

OpenClaw feels more alive than other AI tools because of two key concepts. The "soul" is a file defining its identity and personality. The "heartbeat" is a scheduled job that makes the agent check for tasks proactively (e.g., every 30 minutes), creating the feeling of a collaborative, ever-present assistant.

Long-running AI agent conversations degrade in quality as the context window fills. The best engineers combat this with "intentional compaction": they direct the agent to summarize its progress into a clean markdown file, then start a fresh session using that summary as the new, clean input. This is like rebooting the agent's short-term memory.

Unlike other AI models, OpenClaw can be tasked to figure out how to interact with a new service (like email) and write a reusable "skill" for it. This self-learning capability allows it to continuously expand its own functionality without manual coding.

The "always-on" nature of agents like Clawdbot enables a new work paradigm. Users can assign complex tasks before sleeping and wake up to completed work, effectively turning sleep hours into productive hours for their digital assistant.

The key to continual learning is not just a longer context window, but a new architecture with a spectrum of memory types. "Nested learning" proposes a model with different layers that update at different frequencies—from transient working memory to persistent core knowledge—mimicking how humans learn without catastrophic forgetting.

Unlike session-based chatbots, locally run AI agents with persistent, always-on memory can maintain goals indefinitely. This allows them to become proactive partners, autonomously conducting market research and generating business ideas without constant human prompting.