Blog
All posts
Why the Web Needs Automating
Primal
September 12, 2024
8 min read
Primal Labs

Technology was supposed to revolutionize our lives. There were promises of 20-hour work weeks, robotic servants to do our bidding, and leisurely weekday afternoons in the sun. That was a fantastic dream. So what happened along the way?

Today, we face the grim reality that most of the technology we build simply enables people to do more work.

Your PC is perhaps the best example of this. Sure, it’s a powerful tool. But it’s one that can do almost nothing without a human driving it. You respond to your emails. You browse the Web. You write that report. And you fix it when it breaks.

Could a computer do some of that work for you?

The promise of automation

Automation is the missing link that’s needed to make technology work for us. When a technology automates something, it completes a task that’s valuable to us, but in a manner that doesn’t require our personal oversight.

There’s evidence of automation in our lives today. We have simple assistants that can finish tasks like “wash and dry these dishes” (your dishwasher) or “vacuum this carpet” (a Roomba). Even your PC, problem-generating nightmare that it is, can handle a command to “find and correct the obvious typos in this document” (a spell checker).

But “write my science paper for me” is not a task your computer is prepared to handle. Neither is “buy an attractive couch for my living room” or “think of a good topic for my next blog post.”

Combining automation with the Web

To accomplish these kinds of tasks, most people use the Web.

Unfortunately, the Web today is a lot of work. It’s full of repetitive chores like entering keyword variations, experimenting with different services, paging through search results, and patching together information from many sources.

To accomplish a simple task, you might find yourself with 10 browser tabs open, each with a mile-long history of different services and search terms you’ve explored. Quite likely, you’ve also patched together an email to your colleagues, with notes you’ve hastily copied and pasted.

Using semantic technology to enable new kinds of automation

Before a computer assistant could write a document or go shopping for you, you would need a way to communicate your requirements to it. Your assistant needs a sense of the boundaries it has to work within.

Semantic technology provides a way to draw those boundaries. It provides a formal way for humans to tell computers what it is they need. Without semantics, a computer assistant tasked with writing a paper about “Tesla” couldn’t perceive the topic as anything more than a string of letters. It could do nothing more than patch together every paragraph it finds containing the word “Tesla”.

With semantic technology, “Tesla” becomes a concept rich with meaning. Suddenly, your assistant can understand that you were referring to Tesla the scientist and not Tesla Motors. Even more, if this knowledge model represents your specific ideas and perspectives, as opposed to some universal worldview, your assistant could perform highly personalized work.

With this knowledge model, your computer assistant can perform more complex automation. In fact, it could create a whole website for you about your personal perspective on Tesla.

After all, shopping assistants aren’t much helped by knowing couches are pieces of furniture. To fulfill the promise of true automation, they need to be driven by your particular thoughts and intentions.  The shopping assistant needs to know where you plan to put the couch, why you want it, and what your budget is.

The Web as your virtual assistant

The Web experience of the future will feel more like interacting with a virtual assistant and less like doing a bunch of chores. By providing software agents that understand your interests and intentions, the Web will begin to relieve your burdens rather than add to them.

Maybe there’s still a fighting chance for those leisurely weekday afternoons in the sun.

Editor's Note: This blog was originally published on September 12, 2010. It was updated on September 5, 2024.