Published on

Rethinking Focus in the Age of Agentic AI

Authors

The average person spends at least 2 hours every day being distracted. That is around 1 month of time out of every year. 12% of your time. With the current average life expectancy of 75 years, that is 9 years, at least, that people will spend during their life being distracted. How much would you give to get that time back? Well, that's the cost of distraction.

If distraction was a sport, then I would consider myself an ex-pro athlete. Fortunately, I'm way past my prime and trying my best to retire completely as soon as possible. As a result, I have been quite obsessive about how I spend my time in the past few years. Distraction has been one of the hardest problems to solve in that regard. It's like a tumour that keeps regrowing that I have to constantly re-diagnose and treat. To be clear, I'm not one of those productivity freaks who tries to make every second of the day work-related. I also most certainly do not have a "morning routine". My main goal is to be intentional and feel in control of my focus and time, regardless of whether I'm being productive or not.

I have tried many tools, apps, and techniques to avoid distractions over the years. Through my personal trial and error, I have become a self-proclaimed expert on what works and what doesn't. Up until recently, I wasn't convinced of the existing solutions that I tried as a reliable solution to distraction. If distraction was a disease, then the solutions we have so far have been only effective at temporarily treating symptoms. In fact, some of them also lead to negative side effects. With recent frontiers in tech, mainly AI, I believe we finally have the tools necessary to develop a cure. I find that incredibly exciting.

For this post, I will try to convince you that we do need tools to deal with distraction, why the tools that exist now to deal with distraction are simply not sufficient, and introduce some of my first thoughts on a potential solution.

We need help, the algorithms are too good

Most of us are fighting a losing battle against recommendation algorithms. They have been optimized to the moon and back over the past two decades. There are multibillion-dollar companies built on top of recommendation algorithms that throw an endless amount of resources and engineers to work on keeping us distracted. And they are multibillion-dollar companies for a reason; their products work and people are most certainly distracted. It works too well.

Personally, I have been struggling with what I call "browsing tics". These are subconscious muscle memory impulsive behaviours that I just automatically do while browsing. My current browsing tic is opening YouTube. Sometimes in the middle of work, my brain just decides to open a new tab, I automatically hit "y" on my keyboard, my browser autofills it to "youtube.com", and I hit enter. That whole process takes like a second. That's all it took for me to get into a distraction loop.

Putting this into perspective, I hope, paints a good picture of why it's such an uphill battle for most people to fight against distraction on the internet. The notion that we can consistently win this battle through sheer willpower is somewhat unrealistic. Even if you can, it's such a silly battle to waste your brain power on when there could be software to do it for you. The battle for our attention needs to be automated because the other side already is. The same way that software has become quite effective at training us to get distracted, it can also become equally effective at training us to stay focused.

The Problem with Current Solutions

Hopefully by now, you're convinced that distraction is a problem. So why haven't we solved this yet? There are a ton of apps targeting distraction out there, surely something has solved this? Current solutions for managing digital distractions typically rely on static URL block lists. While this approach seems logical at first glance, it fundamentally misunderstands how we actually use the internet and why we get distracted.

Distraction is a Behaviour, Not a Website List

For a long time, I viewed my distractions as a list of problematic websites to avoid and a focus app that blocked those sites was a perfect solution. In my mind, the problem and solution seemed straightforward: I spend too much time on Reddit and Instagram, so I'll just block those sites and I'll be fine. Time showed that my approach was fundamentally flawed.

What I discovered is that distraction at its core isn't a list of websites, but it is a behavioural pattern that adapts and evolves. In my case, my core distraction was my addiction to feeds. Every time I refreshed a feed and saw new content, my brain got a quick dopamine hit. When I blocked certain sites, my brain didn't stop craving that hit, it simply looked for new sources. Regardless of how many sites I blocked, the internet always found a way to shove a new feed in my face. For several years, I completely blocked social media to stop getting distracted by their addictive feeds. My "feed-checking" behaviour simply migrated elsewhere. I started doom scrolling Hacker News, then my email, and at some point even my GitHub feed. My poor brain was scraping the bottom of the "feed" barrel trying to get that dopamine hit.

A the websites changed, the underlying behaviour remained exactly the same. The crucial realization is that we don't get distracted because certain websites are inherently distracting, we get distracted because we've developed behavioural patterns of seeking quick dopamine hits through constant information consumption. Traditional blockers fail because they target the symptoms (specific websites) rather than the underlying behavior pattern. It's like playing an endless game of whack-a-mole, where your distraction impulse simply finds a new, often "productive-looking" outlet.

The All-or-Nothing Approach

The current approach to website blockers in policing your browsing activity is to simply block a given website completely, or not at all. This is a bit problematic for me with the duality of certain websites. For example, consider YouTube, a typical website to label as a distraction and add it to a block list due to the abundant amount of brain rot that exists. However, YouTube is simultaneously one of the world's largest educational platforms just as much as it is one of its biggest time sinks. On the same platform where you might lose hours to mindless content, you can also get a whole degree's worth of education on almost any topic. With the block list approach, you either get exposed to the brain rot or lose on the potential educational content. It seems like in either case, it feels like you're being punished.

Personally, I decided to not block YouTube due to the educational benefit I get from it. One trick I do though is that I watch any new YouTube content on private browsing mode. Why? Because the moment I relent and watch one of YouTube's brain rot recommendations, the algorithm will turn my feed into garbage. That way, I stay in total control of the recommendations I get in my feed. If I come across a new video that I really enjoy, only then would I actually open it on my actual account for the recommendation algorithm to register it. This has mostly worked and every now and then I get the occasional garbage time-wasting video like Mr. Beast in my feed which I promptly remove.

Context Matters More Than Content

Traditional blockers operate on the flawed assumption that certain websites are inherently distracting, when in reality, it's all about context. Just because you are doing something "productive" does not mean that you are not distracted. This is something that not only focus apps totally miss, but also people struggle to reason with this. The simple question I ask myself is: What am I supposed to be doing right now? If the answer is no, then I'm simply distracted.

As a personal example, I do a lot of software engineering and one of the less exciting parts of my work, in my opinion, is documenting the work I do for other team members to have as a resource. When I'm writing documentation, even productive activities like engineering new features become total distractions.

What AI Unlocks: Intent-Based Browsing

Solving the problem of distractions requires a subtle but fundamental shift in how we interact with the internet. The true problem isn't about identifying what distracts us and blocking it, it's about honouring what we originally came online to do. It is a shift from asking "what sites are distracting?" to "what am I trying to accomplish?". I call this "intent-based browsing".

Consider opening your browser like walking into a room. You should always know what you came into the room to do. If you do not, then you spend time figuring it out or simply leave. By starting every browsing session with setting your intent, blocking distractions becomes a task of determining whether every webpage you visit aligns with your stated intent or not. This seemingly simple human judgment of 'does this page serve my intended purpose?' was previously quite difficult to implement software. LLMs have changed that, bringing good enough reasoning capability to this context sensitive problem. When you set your intent as 'I'm researching medieval architecture', an AI can now correctly assess whether that Reddit thread about cathedral construction genuinely serves your goal or is just another rabbit hole. It's a trivial task for software now, and that is incredibly exciting.

The Power of Setting an Intent

I want to double down on why setting an intent is so important even though it such a simple task. It is astounding how often we browse without any clear purpose and how dangerous that is. It's a familiar scene: you're bored, or maybe you're feeling a bit tired, or you have some extra time in the day, and almost reflexively, you open your laptop. There's no goal, no specific thing you're looking for, it is just an unconscious reach for the easiest form of entertainment available. Just mindless browsing.

Without a clear intention, we're at the mercy of whatever captures our attention next. Social media feeds, news sites, and recommendation algorithms are explicitly designed to exploit this aimlessness, pulling us deeper into content we never actively chose to engage with.

The Case for Intentional Wandering

One of my dilemna's with being distracted on the internet is that a some of my best finds on the internet have come from mindless browsing. And it's totally true, the internet's ability to facilitate serendipitous discovery is one of its most magical qualities. I've stumbled upon life-changing articles, fascinating communities, great ideas, and fire memes through what seemed like aimless browsing.

However, it is crucial to note that even exploration and discovery can be intentional. There's a stark difference between saying "I'm going to spend the next hour exploring interesting corners of the internet" and mindlessly opening Reddit for the tenth time today because you don't know what else to do. When you set aside dedicated time for exploration, you're making a conscious choice about how to spend your time and you're able to be fully present in the discovery process without guilt or the nagging feeling you should be doing something else. The goal isn't to eliminate wandering, but to make it purposeful.

Focal, an early experiment

I strongly believe that intent-based browsing should exist, and I hypothesize that it targets the core problem with distractions on the web. To test out my hypothesis, I built an early experiment called Focal, one of the first intent-based browsing tools to test my hypothesis. Afterall, I did claim it was a "trivial" task for software now. You can check it out here: https://www.focalbrowse.com/

Currently, Focal is a browser extension and it currently works like this:

  1. State Your Intent: On your browser, every time you start a browsing session, a prompt pops up that lets you state your intent for browsing and for how long. You will not be able to proceed without stating that intent.
  2. AI Analysis: As you browse, every webpage you visit is periodically analyzed by an AI to check whether the webpage's content aligns with your stated intent.
  3. Dynamic Blocking: Focal will close tabs that don't align with your stated intent, citing a reason for the closure.

Focal has been a promising experiment for me personally. I've been able to get much more out of my time on the web, and most importantly, I feel in control of my focus online where software is doing all the heavy lifting for me. In fact, by having to state my intent at least a hundred times by now, I feel like my brain has been slowly adjusting to stopping mindlessly open my browser unless I have a clear intent of what I need to do. I've also gotten much better at describing my intention by being punished for providing a poor description and getting blocked on things I need. Is that a feature or a bug? Too early to tell. There is definitely still a lot of work to be done, but the fundamentals are there and, in my opinion, they are sound. That is very exciting.

Next Steps and Vision

Focal is currently merely an experiment at its current state. I am highly considering to open-source it after cleaning up the code and would love for people to contribute. In all honesty, I believe that the eventual point of having good ideas is seeing them executed at some point, regardless who executes them. Whether it’s Focal or not, I hope some product like this reaches the masses if proven useful.

Looking beyond Focal, I believe the ideal intent-based browsing product would include several key components to create a complete solution:

  • Dynamic Page Filtering: Rather than blocking entire pages, an advanced intent-based browsing agent would selectively filter content within pages that doesn't align with the user's intent. Imagine visiting YouTube with the intention of learning math where the system would hide all unrelated videos, dynamically reshaping the site to match your specific purpose. Every website would essentially become a personalized space tailored to the current intention.
  • OS integration: Distractions do not live just in the browser. A truly complete solution would integrate at the operating system level, applying intent-based filtering across all applications, not just web browsers. This also applies across devices as well, mainly mobile and desktop.
  • Agentic Assistance:
    • An AI agent that helps decompose complex projects into manageable browsing intents, creating a structured approach to online research and work.
    • Systems that use browsing history and past intents to suggest productive next steps, helping users maintain momentum toward completing projects.
    • Intent-based context that powers AI assistants to proactively find relevant resources. For example, if your intent is "find a book to learn about sales techniques," the agent could immediately suggest highly-rated options without requiring you to sift through search results.

Final Thoughts

The battle for our digital attention is due for an intervention. Intent-based browsing represents a fundamental shift in how we relate to technology, putting us back in control of our time on the web. If this is something that excites you, you have some thoughts, or you want to contribute, please reach out, I'd love to chat with you! You can find me can find me here or email me here.

Thanks to Liam Hough for helping editing this article.