Most people I know are busy all day and yet the average person is productive for only about two hours. They're working on things that feel important, producing work that feels high quality, and spending time that feels focused. The operative word in all three of those sentences is feels. This essay is about the gap between feeling productive and the reality, and why it's larger than most people want to admit.
Doing the right work
The hardest thing about being a founder is execution. Ideas have always been in abundance, but knowing what to work on distinguishes the good from the greats.
AI has made this harder, and I don't think enough people are being honest about it. When building was expensive, the scarcity forced prioritisation. You couldn't build everything, so you had to figure out what actually mattered. The cost of being wrong was high enough that founders thought carefully before committing time.
That constraint has now almost gone. And without it, most founders are doing what humans naturally do when options are unlimited — exhausting every option. Shipping new features because they can. This creates Frankenstein products that are coherent on the surface but bloated underneath. And whilst the cost of building has collapsed, the constraints everywhere else haven't. Sales still take time. Design still requires judgment. And the need for these functions have exponentially increased across orgs. So now, more than ever, working on the right thing is the critical bottleneck.
Truthfully, I've had more than one moment of looking up after months of work and realising we'd been solving a problem nobody actually had.
In his 1986 lecture You and Your Research, Richard Hamming asks something simple: are you working on the most important problems in your field? Most people, in their heart of hearts, would say no. And it's a simple test to determine whether you're truly working on the right thing over a 5-10 year horizon.
The reason is straightforward. We avoid the most uncomfortable and difficult path by default. And that's how we end up spending years climbing hills instead of mountains, only realising this reality once we've reached the peak.
The Hamming question is the perfect gut-feel check, but it doesn't answer how to solve those problems on a monthly or quarterly basis. Common wisdom holds that you need to speak to users, set long-term goals and monitor them through weekly KPIs. But I've always found this process to be hazy. That's until I came across Rahul Vohra's framework for turning product-market fit into a science. And it's helped us immeasurably on doing the right work each week.
The premise is a four question survey to users:
- Would you be disappointed if you couldn't use this product anymore?
- What type of people do you think would most benefit from it?
- What is the main benefit you get from it?
- How could we improve it for you?
PMF, in his definition, is when at least 40% of users answer very disappointed to the first question. Most startups don't hit that. But the more useful insight is what the other questions reveal. The users who say very disappointed tell you exactly who to target and what to double down on. The users who say somewhat disappointed tell you what's missing, which is the gap between liking your product and loving it.
For instance, with Aftertone, our very disappointed users cared about two things in a productivity tool: speed and intentionality. This then became the filter for everything. Each week, when we sat down to decide what to work on, the question wasn't what's on the roadmap. It was does this make Aftertone faster or more intentional for people looking to be more productive? If the answer was Yes, we knew we were doing the right work that week.
Producing high quality output
Something's been bothering me about my Twitter feed over the past 2 years. Bot replies aside, the new companies and product launches I see are mostly converging into a grey, competent, interchangeable mush. Everything feels samey samey. My theory is that this is a symptom of two compounding problems. The first is the prioritisation problem I mentioned above, and the second is outsourcing our thinking to LLMs. I've been a victim of this without realising and I fear others are going down the same path. Here's what that looks like…
It's Sunday night and I haven't gotten as much done as I'd like. Because we're in the 'AI golden age' my co-founder is expecting a lot more from me. So in my final hours, I open up trusty Claude and use it to quickly knock off a few tasks on my list. The output is just about acceptable, so I move on and consider it done.
A similar thing happens a week after, when it's even later on Sunday. And soon, I'm in this cycle. On a daily basis, I'm mentally contracting the time it takes to complete a task. It's a warped adaptation of Parkinson's Law — work that previously took weeks now takes minutes. But the quality descends with it, regressing to the mean.
It also does something more sinister. Every time you use AI to lead on a task, it destroys your originality. I've seen everyone from CEOs to designers increasingly use AI as their first base. But MIT research published last year found that regular LLM use measurably reduced independent thinking over time. The more you rely on it, the less capable you become without it. It's a terrible gateway drug to outsourcing our thinking altogether, and not enough people recognise, let alone talk about it.
The single most powerful change I've made is starting on a whiteboard. Whenever I'm tackling a big problem — a product decision, a piece of writing, a strategic question — I'll steer the direction and then use AI to provide counter-factuals and blindspots. It's worked wonders because AI is extraordinarily good at making half-formed ideas look complete. And it also anchors your thinking to its first output, so you're mostly editing rather than generating something novel. This whiteboard-first approach flips this entirely.
As more people recognise this, I expect a strong counter culture. Easy work is rarely high quality, and I don't think AI will ever change this. The people that I see being super-productive in the future are the ones who use AI without outsourcing their thinking.
Amount of time in flow
Over the years I've found myself stuck in a manager's schedule. For those unfamiliar, Paul Graham wrote about this in depth. In practice, my day is constantly fragmented. Calls, emails, urgent tasks that make it difficult to have long stretches of time without interruptions. The cost of this is both the quantity and quality of output.
Flow — what Cal Newport calls work that's distraction free and pushes you to your cognitive limits — is where your best work happens. We all know it when we experience it. It's the periods where the quality and depth just feels a level above everything else. The difficulty is that most people only achieve around 2 hours of it a day, and I suspect that many overestimate how much they actually achieve.
I find it more useful to explore flow through the conditions that create it rather than trying to define what it feels like. The most objective definition I've landed on is work that is at least 25 minutes long with breaks that are less than 20 minutes. Research has shown that one distraction takes 23 minutes to fully recover from. So a break that creeps past 20 minutes essentially breaks the ability to achieve flow.
The challenge is that achieving uninterrupted stretches of high quality work is becoming more difficult. A large part of that is down to distractions. There's long been a notification problem that interrupts our ability to achieve deep work. But I've noticed a new one recently. Whenever I'm waiting on an LLM prompt, I reach for my phone and start scrolling. And with increasing LLM usage, I imagine this will become increasingly pervasive.
There are screen blocking apps that put a hard stop to this. Opal is my favourite in this category. But the most effective change I made was moving distracting apps off my home screen and onto a secondary page. I've found that just one extra tap is enough friction to break this habit most of the time.
Emails are another common source of disruption that make achieving flow difficult. The perception is that almost every single email is an urgent task because of recency bias. The best thing I've done to address this is to use Superhuman. It's not cheap, but their product is highly optimised for reaching inbox zero; either you reply in the moment, set a reminder for later or remove it for good. It's massively reduced the cognitive burden that comes with handling my inbox and is the only product that I've subscribed to for four years.
Meetings are perhaps the final killer of flow. There is almost always a buffer period before and after a call that means even a 30 min meeting occupies an hour of time. The mistake I've made in the past is offering too much flexibility with my available meeting times. Often, both my mornings and evenings would be occupied with flow sucking interruptions. Since recognising the impact, I've switched to offering late afternoon/evening timings in blocks as opposed to leaving my entire calendar open.
Tools that help (and don't)
In the pursuit of maximum productivity, I've also found myself going down the task manager rabbit hole. Over the years, I've used everything: Akiflow, Motion, Routine, Todoist, Things3. All of them do what they promise, they're great for tracking your tasks and having your calendar reflect that. If you're looking for a traditional to-do list app, then nothing beats Todoist. If you want to leverage pure AI scheduling then Motion wins.
These tools are great organisers and I'd recommend them for this purpose. You'll see your tasks/meeting for the week ahead by default, they'll pull in tasks from various other tools and they allow you to schedule meetings with their own Calendly equivalents. But a universal pain is that none of them help you achieve flow through the lens discussed here or through validated principles. In fact, I've personally felt that many have the opposite effect. I've often found myself trying to perfect my calendar, and ironically spending more time organising than executing.
Most productivity advice treats the factors I've discussed here as separate problems. Fix your priorities. Improve your output. Protect your focus time. But in my experience they compound. Working on the wrong things means you don't notice the quality declining. Outsourcing your thinking means you can't tell whether the problem is worth solving. Fragmenting your focus means you never go deep enough to catch either. They compound each other.
To that end, Niall and I have been building Aftertone — initially just a side project for ourselves that we fully productised late last year. We wanted something that drove us to work more productively across all three domains. And the principles we built around were a maniacal focus on speed, intentionality, and scientific principles for achieving flow. We're personally onboarding users and looking for founders, operators and researchers who resonate with this problem. See if you're eligible for access here.