A Step Closer to Digital Human - Agentic Browser

Recently, we saw the announcements of ChatGPT Atlas and Perplexity Comet.

Both of these seem like somewhat useless products and don't appear to add much on top of a standard Chrome browser.

However, I believe that these "agentic browsers" are going to have huge impact.

Why? The biggest problem with current AI agents is that they cannot see what we see. They cannot consume what we consume, so they cannot act the way we act.

We live much of our lives digitally and online, but our agents do not.

If I were able to give my agent access to my Instagram, YouTube, WhatsApp, email, calendar, etc., it would start to think more and more like me.

Since the browser already has access to all of this, it is the key component required.

There are one or two more features that would be required to make a big impact. 

First, we would need the ability to schedule tasks. This way, my agent could be like me, checking WhatsApp, email, and YouTube periodically.

Second, like Cursor and Claude Code, it should be able to write to disk so that it can create and manage memory.