Updated 2026-08-08

What Lifelogging Is, and Why I Record All Day

A Pastok user on why he records his whole day: what lifelogging is, the reasons people start, what actually comes back, and what it all costs.

A 16-hour day of audio takes 16 hours to listen to. The difference between keeping that record and ever reviewing it is why most people who try this give it up, and it is the first thing I would want to know if I were reading rather than writing it. I still do it. Not because I go back and listen, which I hardly ever do, but because the part worth having arrives without me going to look for it.

Why do people record their whole day?

Lifelogging means keeping a continuous record of your life. People do it with cameras, location traces and paper. I do it with speech, and I started for three reasons that turn out to be the common ones: the days had stopped being distinguishable, I kept losing arguments about what was said, and I wanted to know whether I was changing.

I am a customer here, not an employee, and Pastok wrote this up from what I told them. Speech is where I say the things I would never sit down and write, which is why my record is audio rather than photographs.

Memory is the big one of the three, and it is where I started. The days blur together. "Why can't I remember most of my life" is a question people genuinely type into search engines, and my honest answer was that nothing was keeping the days apart. A diary would have fixed it. I did not keep one.

Second is wanting a record of what was actually said. Somebody tells you they never agreed to that. A contractor says "I never quoted that." You end up arguing about the past out of memory, which is a fight you lose slowly. This is the reason people back with real incidents rather than a general feeling.

The third is harder to admit. I wanted to know whether I was changing at all, in either direction, instead of assuming.

What I tried first, and why none of it held

Keeping a journal failed the way it fails for most people. Friction, a missed day, guilt about the missed day, then nothing. The effort is paid every evening and the payoff is deferred indefinitely, which is a bad trade to ask anybody to keep making.

Trackers failed for a more interesting reason. A tracker can only correlate what you told it to watch. You pick your variables at setup, tag your days, and it reports how the things you nominated relate to each other. That is useful when you already have a hunch.

I would never have created a field for the neighbours' noise. That turned out to be the whole point.

Then I tried a recorder I pressed when something important seemed to be starting. It works for meetings. It fails at everything else, because you do not know in advance which two minutes will matter, and the moments worth keeping are the ones nobody announces.

What actually comes back?

Not an archive. A summary that builds during the session and arrives minutes behind the conversation rather than sentence by sentence, plus the things I said I would do with their timing attached. Then, occasionally, something I did not ask for.

The one that changed my mind was about the neighbours. I was not tracking noise. I did not suspect the noise. I complained about it sometimes, the way anyone does, on afternoons that had gone badly for reasons I would have named differently. What came back put the two things next to each other and pointed out that they kept turning up together. The same aside on a Monday and again on a Thursday is invisible from the inside. It is only a pattern from outside.

The second one crossed days. On a hot day I described how I was feeling in ordinary words, not as symptoms. Two other things were already in the record: I had slept badly the night before, and I had trained hard that morning. What came back set those three beside each other and raised the possibility that the heat had something to do with it.

That is also the limit. It connected something I said with activity already in the log and suggested a link. It did not decide what was wrong with me, and it cannot. No single recording could have produced it either, which is the real argument for running all day instead of pressing record when you think something is starting.

What does it actually cost?

The review burden comes first, because it is what ends most attempts. People accumulate thousands of hours and open almost none of it. Anything sold to you as "record now, review later" is asking you to do the work twice, and the second half never happens. So the promise has to run the other way round. Things arrive, or the whole exercise is a backup you never open.

There is a quieter cost, which is holding your own past. People who keep journals for years sometimes destroy them, and the reason is rarely embarrassment. They did not want to carry all of it, and some say afterwards that they wish they had kept the good parts. What I keep is deletable, and the raw audio expires by default whether I remember to act or not.

The strongest objection I have run into is confirmation bias. "Are you just deciding, after the fact, that it was a stressful week?" Against a chart that is a fair hit. The distinction I would draw is that an observation taken from what I actually said, surfaced before I went looking for it, is not the same act as reading a graph backwards.

Battery is the practical ceiling, not a session length. If a call or Siri takes the microphone, iOS will not hand it back on its own, so what I get is a notification to start it again.

Isn't recording all day a problem for everybody else?

Sometimes, and the usual answer is a bad one. The pattern is to tell people at first, find that it unsettles them, then quietly stop mentioning it. That solves your own awkwardness by moving the cost onto them.

Say it instead, and stop when somebody would rather you did not. In the US, 39 states and DC need one party to consent; 11 need everybody. Where several people are talking, speakers are separated as anonymous labels, never by name.

The form factor does more work here than any policy. Putting a recorder on the table changes a room. A watch is already on my wrist.

How I actually run it

One press on the watch face in the morning starts it. It does not start itself, and I would not want it to. It runs until I stop it, and I do not press anything again when a conversation begins, which is the entire reason the neighbour observation existed to be made.

Raw audio is deleted after processing, within seven days by default. What persists is text, and that stays until I delete it. Transcription covers 119 speech locales, and what comes back follows the language I was speaking. The word-for-word transcript lives on the web; my phone shows the summaries and the reminders.

That is the trade. All day is what makes the ordinary details available, and the ordinary details are where the thing you did not know about yourself is hiding.

Start with one day

Press once in the morning and leave it alone. What comes back that evening is the whole test: if nothing in it was worth keeping, you have lost a day and learned something useful about whether this is for you.

Start a day of capture

Related

Start with one day

Press once in the morning and leave it alone. What comes back that evening is the whole test: if nothing in it was worth keeping, you have lost a day and learned something useful about whether this is for you.

Start a day of capture