Life as a Service, too many AI agents.

I recently heard about a tool called OpenClaw that’s been taking over the agentic automation space by storm, along with its waves of security vulnerabilities. It’s a free and open source autonomous AI agent capable of doing anything a digital personal assistant can, all handled through a messaging app like Discord or WhatsApp. While this tool sounds fantastic in theory, it sounds like a headache to maintain and absolutely overkill for anything I need to do.

For example, many people ask their OpenClaw agents to summarize their new emails in the morning, text the list of important ones to them and draft up replies. However, I’ve spent over 8 years maintaining inbox zero for all my email clients (Gmail, Outlook) and I honestly don’t spend much time reading and replying to them. So having an automated agent write text summaries to me and draft up replies seems like extra bloat to me.

This is one of the many examples of why I wouldn’t need an agent to do something that I can just do myself and be more confident in the result. I do use automation tools like n8n or Windmill, I just find these full agentic systems like OpenClaw to be too non-deterministic, thus unreliable.

So, all of this “OpenClaw talk” leads me to my point, how far are we planning to abstract things? We have IaaS/PaaS (Infrastructure as a Service/Platform as a Service) where we get to manage the Cloud Infrastructure of our business or personal projects. Then we have SaaS (Software as a Service), with tools like Gmail or Canva, where the business handles all things technical like the servers, storage, networking, and the user only worries about using the application. In the last couple of years, we have had MaaS (Model as a Service), where many companies are integrating chatbots into their products, allowing you to skip the clicks and just ask the chatbot to handle the task.

With the emergence of tools like OpenClaw, I am starting to see this new abstraction layer show up, OaaS (Outcome as a Service), where you don’t need to manage anything, you just ask the agent to do something and it will produce the outcome you want. It’s great we are in an era that allows for this, but I worry it’ll abstract too far to the point of losing our own agency.

I love automation more than most, but I worry that we are relying too much on agents to the point where we start offering “LaaS” (Life as a Service). This is the extreme case, but I think it gets my point across.

Abstracting tedious and repetitive tasks are fine, but it’s important we don’t abstract the parts that make life fun. If you are wealthy and busy, it makes sense to have someone book your flights for you since you might be taking flights every week. But if you enjoy planning a vacation, scheduling the day, or ordering the groceries for a healthy lifestyle, don’t let an agent do it just because it can.

Effort creates meaning. If we remove the human from every loop to maximize efficiency, we end up with a perfectly optimized life that feels completely empty.

References

Lindzon, J. (2022) Why a work challenge is a good thing, Rotman Insights Hub. Available at: https://www-2.rotman.utoronto.ca/insightshub/leadership-career-development/meaningfulness-effort.