I've been digging into some of the recent research from ICLR 2025, and I saw 20+ papers !!

Tons of new work on web automation, like WebCanvas, AgentOccam, VeSX, and NNetscape Navigator, is pointing to a future where smart agents handle more of our online tasks for us. But with all this progress, a bunch of questions come to mind about how this is actually going to change things.

Figured I'd throw these out for discussion:

1. Is Full Automation Close? – Most of us start any task online with a search and follow a trail from there. Today, there’s no single tool that can do a whole complex task from start to finish. How close are we to an agent that can take over a complete workflow ?

2. What About Ads? – Big companies depend on ads to make money, but if agents can just skip them, what's going to happen? Will they need new ways to keep our attention, or will ads shift to target the agents themselves?

3. Agents Running 24/7 for Us? – Imagine agents constantly monitoring things for us—like tracking prices, reading industry news, or finding the best deals—all day, every day. Sounds cool, but are we ready for this level of constant automation?

4. Agent-Friendly SEO? – Right now, SEO is all about human clicks. But what happens when we start optimizing for agents? Are we going to see a new wave of "Agentic Content Optimization"?

5. Web Macros as a Standard? – What if every website published “web macros” (like example.com/webmacro1) that agents could find and use to complete tasks? Would this make automation easier, or would it lead to a whole new type of SEO arms race?

6. APIs or Macros, or Both? – APIs are the go-to right now for any programmatic interaction, but maybe a mix of APIs and macro mining is where things are headed. Which one will become the standard for web agents?