Burnout and Pace in a Fast-Moving Field
5 min read · updated August 3, 2026
The tiredness people describe in this field is usually not overwork. It is the specific exhaustion of an obligation that cannot be discharged — a queue that refills faster than it drains, attached to a fear that the un-read item was the important one.
What the fatigue actually is
Three distinguishable things get called the same name, and they have different remedies, so separating them is the first useful move.
- Information obligation. The sense that you should have read the thing everyone is discussing. This is the one that responds to a rule, and most of this page is about it.
- Obsolescence anxiety. The fear that your skills are being invalidated. This does not respond to reading more — reading more usually feeds it — and it responds instead to looking honestly at what your skills are attached to, which is generally more reassuring than the anxiety predicts.
- Ordinary overwork, wearing a field-specific costume. Deadlines, understaffing, too many contexts. This is a management problem and no amount of triaging your reading will fix it. It is worth checking first, because the field-flavoured version is easier to talk about and it can hide the real one for months.
The first two share a mechanism: an implied standard of “current” that has no definition and therefore can never be met. An obligation with no completion condition is a reliable way to manufacture guilt, and the guilt is not evidence of a deficiency.
Keeping up was never on the table
Worth doing explicitly, because the feeling of falling behind assumes a baseline that does not exist. Use your own numbers; these are illustrative and deliberately conservative.
Say you follow, modestly: 40 papers/week you might plausibly care about 15 substantial blog posts or write-ups 10 release notes and changelogs 5 tools worth an hour of evaluation Reading a paper properly ~2 h A write-up, with the linked source ~0.5 h Changelogs, skimmed ~0.1 h Evaluating one tool ~1 h 40*2 + 15*0.5 + 10*0.1 + 5*1 = 93.5 h/week Available: whatever is left after doing your job.
The point of the arithmetic is not the number. It is that the ratio is not close, and it has not been close for years, which means everyone you compare yourself to is also ignoring almost all of it. The people who appear current are not reading more than you; they are ignoring differently, and usually more decisively.
Two corollaries follow, and both are permissions rather than advice. Selection is compulsory, so choosing what to ignore is not a concession — it is the only available strategy. And because everyone is selecting, the person confidently discussing a paper in a meeting has read a summary of it, as have you, and the asymmetry you are imagining is mostly not there.
A triage rule
Concrete, so it survives a bad week. Everything that arrives goes into one of three buckets, decided in about five seconds.
| Bucket | Description |
|---|---|
| Act | It changes something I am deciding now, or it affects a dependency I run. Read it today, properly. This bucket is small — usually one or two items a week, often zero. |
| Park | Might matter later. One line in a note with a link, and nothing else. The note is not a backlog to clear; it is a search index for the moment the question becomes live. Reviewing it monthly and deleting most of it is the correct maintenance. |
| Drop | Everything else, which is the great majority. Not 'later' — dropped. Anything genuinely important will reach you again through a different route, and that redundancy is what makes dropping safe. |
The rule that makes it work is that Park is not a queue. The most common failure is turning it into a reading list, at which point it becomes a second source of obligation and reproduces the original problem in a tidier form. Park is a search index. You are allowed to never read any of it.
One extra safeguard against the fear behind the fatigue: subscribe directly to the changelogs of the things you depend on, so the category of news that could genuinely hurt you arrives by a channel that does not require vigilance. Once that is in place, the cost of dropping everything else is much lower than it feels, and the wider filtering method becomes mostly optional.
Depth is the antidote, not discipline
The instinctive response to feeling behind is to consume more, which reliably makes it worse: more input, more awareness of what you have not read, more guilt. The response that works is counterintuitive and consists of going deeper on less.
The reason is that breadth-anxiety is fed by not having a stable frame to place new information into. Someone who understands why generation is sequential and what that implies for latency and price can classify most new announcements in seconds — this is a variation on a mechanism I understand, or it is genuinely new. Someone without that frame experiences every announcement as an undifferentiated new thing to learn, which is both exhausting and endless.
So the practical move is to spend the reading time you were going to spend anyway on one durable thing rather than ten current ones. A week spent properly understanding how retrieval fails, or what an evaluation can and cannot detect, permanently reduces the volume of material that feels like it requires attention — because most of it is now recognisable as an instance of something.
There is a second-order benefit worth stating. Depth is also what makes you hard to replace, since it accumulates at the layers that do not turn over, so the anxious impulse and the career-optimal move point in opposite directions. Following the anxiety is the expensive option in every sense.
When it is not fatigue
A short section, because a page about pace that ignores this would be irresponsible. Triage rules and reading habits address information load. They do not address the following, and treating the following as an information problem delays getting the actual help:
- Tiredness that does not lift on a proper holiday.
- Loss of interest in the parts of the work that used to be engaging, rather than in the discourse around it.
- Cynicism about the work or the people, arriving where it was not before.
- Difficulty starting tasks you know how to do.
- Effects showing up in sleep, health or relationships.
Those are burnout signals rather than pace signals, and the response is a conversation with a manager, a doctor or a professional — not a better reading system. It is worth saying plainly because the fast-moving nature of the field supplies an unusually convincing explanation for symptoms that have a different cause, and a convincing wrong explanation is the thing most likely to keep somebody stuck.
For the ordinary version — the tiredness of an unfinishable queue — the whole of the remedy is in one sentence: nobody is keeping up, the obligation you are failing does not exist, and the time you get back by dropping most of it is better spent understanding one thing properly.