Why a productivity obsessive is suddenly the voice of reason on AI
The AI conversation in workplaces usually arrives with the volume turned up too high. One person says the bots are about to rewrite office life. Another says half the point is to sound like you’re rewriting office life. In the middle of that noise sits Pierce, who has made a career out of poking at software until the seams show.
He’s not coming at this as a casual user with a half-full downloads folder. Pierce has spent years as a tech editor, a newsletter writer, and a product critic, which means he has made a habit of living with tools long enough to see what survives the honeymoon. His most memorable claim is also the simplest: he has installed, used, and then abandoned roughly two hundred to-do apps. That’s not normal behavior unless you count software archaeology as a hobby.
Which is why his AI advice lands differently. He isn’t asking whether artificial intelligence can sound impressive in a product demo. He’s asking a far less glamorous question: what can people actually do with AI at work right now? Not next quarter. Not after the next round of policy memos. Right now, in the middle of inboxes, meetings, chat threads, and the usual office sludge.
The most useful AI advice usually comes from someone who has already watched a lot of tools fail at ordinary human habits.
That framing matters because the gap between promise and practice is still wide. AI vendors keep talking as if white-collar work is about to be transformed overnight, but most office routines are stubborn little creatures. People still copy text into documents. They still lose track of action items. They still spend too much time sorting through email that should never have been sent in the first place. The tools are newer. The habits are old.
Pierce’s perspective cuts through some of the usual tech news theater because he has seen this movie before in smaller form. Every new app claims it will fix attention, simplify planning, or make busy people feel less behind. In digital culture, that pitch rarely dies; it just gets a shinier interface. AI is doing the same thing now, only with bigger promises and more expensive slide decks.
That doesn’t make the software useless. It just means the useful parts are more modest than the hype suggests. The real story, for now, is not whether AI can replace work. It’s whether it can shave off the parts of work that are repetitive, annoying, or embarrassingly mechanical without pretending it knows how you think.
And that is where Pierce’s background matters most. Someone who has spent years trying to make productivity tools behave is well positioned to notice when AI is actually helping and when it is just wearing a cleaner suit. In the next section, his long run through task apps explains why his standards are so low in the best possible way.
The endless hunt for the perfect task app
Pierce’s advice sounds calmer than the average AI hot take because he’s spent years poking at the machinery of getting work done, one app at a time. His first App Store download, back when the store still felt a bit like a magic trick, was Evernote. That detail matters. Plenty of people dabble in productivity tools for a weekend and move on. Pierce has been in the weeds long enough to know the difference between a shiny interface and a system that actually survives contact with a busy Tuesday.
Since then, he’s cycled through the usual suspects and then some. Todoist has been in the mix. Apple Reminders too. So have NotePlan, Obsidian, and Craft. The list reads less like a permanent setup and more like a long series of compromises, which is probably the honest version of how most people work anyway. There’s always some new app promising a cleaner brain, a calmer inbox, or a more civilized morning. Then a deadline shows up and the whole elegant scheme starts limping.
The best task app is the one you stop thinking about long enough to do the actual work.
That’s the catch Pierce keeps circling back to. He doesn’t seem interested in owning the perfect system so much as building one that fails gracefully. He wants capture to be easy, ownership to be real, and retrieval to be less annoying than memory. A lot of his thinking comes down to a simple preference: if he writes something down, he wants it stored in plain text files he controls, not trapped inside a proprietary format that could vanish, change, or start charging rent.
That’s where tools like Obsidian and NotePlan have had appeal. Plain text is boring in the best way. It’s readable, portable, and not tied to one company’s app store mood swings. If you’ve ever watched a favorite tool become slower, uglier, or paywalled overnight, the appeal is obvious. This is especially true for people who treat notes as working documents rather than digital souvenirs. When the file lives in a format you can open elsewhere, you’re not negotiating with the software every time you need your own notes back.
Craft has pulled him back in recently, at least for now, because it added task support and made room for Apple Reminders. That kind of integration sounds minor until you live with it. The difference between “I can keep this in one place” and “I need to remember which app the thought landed in” is the difference between a usable day and a small administrative headache that repeats itself fifty times. Pierce seems to like tools that cut down on that little friction without pretending to solve human attention.
He’s also settled on a daily-note workflow, which is less glamorous than it sounds and more useful than most software demos admit. Calendar items, tasks, and notes live together in one daily page instead of scattering across five apps and a prayer. A meeting note can sit next to a task. A stray idea can sit beneath an appointment. The day becomes one file, or close enough, which means less hunting and less second-guessing. There’s a certain relief in that. You open one place and see what the day contains.
For anyone juggling work, household errands, or even the kind of low-grade admin that follows you from email to message thread to calendar, that setup can feel saner than the app-picking ritual itself. It also fits the broader case he’s making about plain-text productivity tools: the goal isn’t to build a shrine to organization. It’s to keep enough structure around your notes and tasks that you can move without thinking about the container all day.
And that may be why his system sounds a little fragile on paper but works in real life. It leaves room for mess. It assumes he’ll forget things unless they’re captured somewhere obvious. It doesn’t pretend every task deserves a fancy home. In a world full of AI productivity tools promising to think for you, that restraint feels almost rebellious. A notebook, a task list, and a daily page still do most of the heavy lifting, which is exactly the sort of unflashy arrangement that usually survives longest.
For a broader take on why software should stay in its lane, his argument about what ChatGPT can’t plan for you lands in the same neighborhood. The plan, if there is one, is still mostly yours.
What AI really helps with at work—and what it still botches
The useful part of AI at work is oddly unglamorous. It’s the stuff nobody wants to do after lunch, or before lunch, or ever: converting a pile of files into the format your finance team actually uses, clearing a clogged inbox, or finding a plumber who can come by before the ceiling decides to become a water feature. That’s where David Pierce thinks the tool earns its keep. Not in some grand reinvention of the office, but in the drudge work that eats time without asking for much judgment.
AI is best when the job is annoying, repetitive, and easy to verify.
That’s a much smaller promise than the one sold in most product demos, but it’s a real one. Pierce has tested AI on the chores that usually make people sigh and reach for a second coffee. Batch conversions? Sure. Inbox cleanup? Often useful. A quick search for service providers, like a plumber or a local repair person? Also fair game. The model can sort, summarize, and narrow the field faster than a person who’s been staring at the same tab pile for twenty minutes.
The trouble starts when people ask it to do the harder thing: tell them what matters.
Pierce said he tried running AI against his main inbox, and the results were mostly wrong. That’s not because the software is lazy. It’s because email senders are not. They’ve spent years learning how to make a message sound urgent even when it’s not. A billing notice can be phrased like a final warning. A calendar invite can look like a crisis. A cheerful “quick follow-up” can hide a request that will chew up the rest of your afternoon. AI can spot obvious clutter. It has a much harder time sorting genuine priority from the theatrical panic that already lives in corporate email.
That problem gets worse because people have figured out the new trick. If AI tools are trained to look for urgency, then every sender has an incentive to write like the office is on fire. The result is a small arms race of fake urgency, with subject lines and message bodies dressed up to look like emergencies. Humans did not need another way to turn a simple ask into a siren, but here we are.
Even when AI is used to scan across email, chat, and task lists for missed items, the results seem mixed. It can pull together scattered notes, surface something you forgot to answer, or catch a meeting follow-up that fell behind a pile of Slack messages. Then again, it can just as easily surface the wrong thing with great confidence. That makes it useful as a helper and unreliable as a manager. One can save time. The other can waste it in a fresh and slightly annoying way.
If you want the broader argument behind that view, it’s the one in this piece on whether work actually needs an AI revolution. The short version is that a lot of office labor still comes down to sorting, noticing, and deciding, and AI only handles the first part cleanly.
The best results, so far, come from low-stakes cleanup. AI can sweep the floor. It can label the boxes. It can dig out the phone number for the plumber. It can probably help you find the document you misnamed at 11:47 p.m. Last Tuesday. What it can’t do, at least not reliably, is think for you about what deserves attention first. That’s the messy part, and it’s still human work.
Pierce’s skepticism lands because it’s practical, not gloomy. He isn’t arguing that AI is useless. He’s arguing that most of the hype skips right past the part where office life is already full of judgment calls, half-finished threads, and inboxes that behave like tiny propaganda machines. For now, AI looks best when it takes the chores nobody brags about. The second it starts auditioning for your boss, things get weird.
And if the question is how to keep from losing track of the rest of it, the answer may have less to do with smarter AI than with a better place to put your own thoughts. That’s where the next part of this conversation gets interesting.
Single source of truth beats a perfect system
The useful part of Pierce’s advice is that it doesn’t ask people to become monks of productivity. It asks them to stop scattering the same thought across five places.
That sounds simple until you watch a workday unfold. A request lands in email, the follow-up chat happens in Slack, the draft lives in Google Docs, the rough note gets dropped into Obsidian, and the actual task ends up in Apple Reminders or Todoist. By 4 p.m. Nobody can remember where the thing began, where the latest version sits, or which copy is the real one. The job hasn’t gotten harder. The handoffs have.
A single source of truth solves two annoyances at once: where to put something now, and where to find it later. That’s the whole trick. If every action gets a home, even a messy one, you don’t waste time deciding whether a client call belongs in a notes app, a task manager, or the inbox you already hate. If that same home is the place you return to when memory gets fuzzy, you also avoid the familiar scavenger hunt through half-finished drafts and old messages.
A slightly untidy system you actually use beats a pristine system you keep abandoning.
That’s why Pierce seems comfortable with a little visible mess. He isn’t chasing the fantasy of a spotless dashboard where every thought is filed and color-coded before lunch. He wants enough structure to keep moving, but not so much structure that the structure becomes the job. If a task keeps showing up in the same place every day, it does more than sit there politely. It nags. It refreshes memory. Sometimes it even does the odd creative work of making a connection you missed the first time around.
That idea runs against the modern urge to hide everything in a separate app and assume that means it has been handled. A task buried deep in a system can feel finished even when it hasn’t been done. A task sitting in plain sight has a different personality. It reminds you. It gives your brain something to worry over in the background, which is often where better ideas show up. That’s one reason he keeps returning to tools like Obsidian, Todoist, and Apple Reminders without pretending any one of them is magical. They’re containers. The value comes from how little friction they add when a thought needs a home.
That same practicality shapes his view of AI. He’s not arguing that people need to become better than the machine at every move, which is where a lot of the anxiety comes from. The race-to-stay-ahead story can turn into a kind of workplace superstition, as if everyone needs to outthink software before breakfast or get left behind. Pierce doesn’t buy that. The smarter move, he seems to say, is more boring than that: offload the dumb tasks, keep doing the work that actually teaches you something, and don’t confuse software with fate.
That’s close to the point Cal Newport has made in his writing on avoiding digital productivity traps and whether AI makes us lazy. The trap is rarely a lack of tools. It’s the overhead created by too many tools, too many tabs, and too much context switching. AI can take some of the dull sludge off the desk. Fine. Let it. But the part that builds judgment, pattern recognition, and actual expertise still has to be done by a human being who has lived inside the work long enough to know what matters.
So the practical advice lands in a pretty unglamorous place. Put the thing in one place. Keep it visible long enough for your brain to work on it. Use AI to clear the chores that don’t need your taste or attention. Then keep your hands on the work that does. That doesn’t sound like a revolution, which is probably why it’s more useful than most of the loud stuff around it.
The takeaway: treat AI like a tool, not a life plan
Pierce’s most useful advice may be the least glamorous one: calm down. The pressure to hop on the AI train or get left behind has grown loud enough to make ordinary work feel outdated before it’s even begun. That’s a bad way to think about software, and an even worse way to think about your day. If an AI tool saves time on one chore, use it. If it gets in the way, drop it. No ceremony required.
The smartest move is usually the boring one: let software do the grunt work, then keep your judgment where it still matters.
That framing matters because so much of the AI at work conversation gets tangled up in panic. People talk as if every spreadsheet, memo, inbox, and calendar invite needs a new machine-learning strategy. In practice, the question is narrower and much less dramatic. Which tasks are repetitive enough to automate without much risk? Which ones depend on context, memory, taste, or a little social reading between the lines? Those are still human jobs, even when the software around them gets shinier.
Pierce’s point is not that AI is useless. He’s saying the hype often asks for a full personality transplant when what most people need is a faster way to sort files, clean inbox sludge, or find a plumber without turning the whole afternoon into a scavenger hunt. That’s a pretty modest bar, but it’s the right one. A tool that handles a dull task well is useful. A tool that promises to run your working life may be selling you an expensive headache with a chat box.
The word that keeps cropping up here is software. Not destiny. Not a philosophy. Not some grand rewrite of how you think and work. Just software. That mental model helps strip away the performance anxiety that follows every new AI release. You don’t have to turn every app into a manifesto. You don’t have to rebuild your habits around each product demo that makes a crowd gasp for six seconds.
And that leaves the real decision where it belongs: with the person doing the work. Keep the tasks that teach you something. Automate the ones that merely chew through time. Leave enough mess visible that your own brain can still see what’s happening. Then, when the next wave of AI advice comes along claiming everything must change at once, you can do the radical thing and ignore most of it.
The future may be moving fast. Your task list doesn’t need to.
