Shayan argues from personal experience that the same devices and browsers used for coding are optimized to fracture attention, creating an inescapable collision between work tools and engagement-maximizing surfaces. He describes concrete moments — alt-tabbing to check a flag and ending up on Reddit, reaching for his phone mid-paragraph — as evidence that concentration itself now requires deliberate training.
The editorial pushes back on the lazy 'blame TikTok' framing, arguing instead that the cumulative tuning of every interface for engagement has raised the baseline cost of concentration. It notes that when the professional class building these systems describes itself in identical language as its casualties, the mechanism must be structural rather than app-specific.
The editorial highlights that hundreds of engineers in the HN thread — the professionals who ship attention-economy products — are describing their own reading, video-watching, and code-review capacity in near-identical terms of decline. This convergence of self-reports from insiders is treated as stronger signal than any external critique because these are the people who understand the mechanics.
A blog post titled *It's getting harder to focus every day* by developer Shayan (glyphack.com) climbed to 660 points on Hacker News, with a comment thread full of engineers describing the same slow-motion collapse of their ability to sit with a problem. The essay is not a research paper. It's a first-person account of a working developer noticing that the deep-work stretches that used to be routine now feel like something he has to train for.
The specifics land because they're mundane. Opening the terminal, alt-tabbing to a browser to look up a flag, seeing a red notification badge, ending up 15 minutes into a Reddit thread about mechanical keyboards. Sitting down to read documentation and finding, three paragraphs in, that his hand has already reached for the phone. The essay's central claim is that the tools we use to write software have quietly become the same tools optimized to fracture our attention, and we've stopped noticing the collision.
The HN comments are the real signal. Hundreds of developers — the professional class that supposedly *builds* this stuff — describing the same pattern in near-identical language. Reading getting harder. Long-form videos getting harder. Books requiring a warm-up period. Code review sessions punctuated by phantom-notification checks. When the people who ship the attention economy are themselves its most articulate casualties, that is worth pausing on.
There's a lazy version of this discussion that blames TikTok and moves on. The essay, and the thread underneath it, is more useful than that. The mechanism people keep describing is not that any single app is uniquely addictive — it's that the *aggregate* pressure of every surface being tuned for engagement has raised the ambient cost of concentration to a level where deep work now requires deliberate defense.
Compare the working conditions. A senior engineer in 2010 had a text editor, a terminal, a browser with maybe five tabs, and email that pinged every few minutes. A senior engineer in 2026 has an IDE that surfaces AI suggestions inline, a terminal with a copilot, Slack with @-mentions from four workspaces, GitHub notifications, Linear updates, a phone within arm's reach with 40+ apps each individually A/B-tested against a control group of millions, and a browser where every content site now interrupts scroll with a modal. The environment changed. Blaming the individual for not being able to focus in it is like blaming the fish for the water.
The engineering angle sharpens the point. We understand — because we build them — how these systems work. Recommendation feeds are optimized against a proxy metric (dwell time, session length, next-video-clicked) that has no term for the user's stated goals. Push notifications are batched and timed against models of your reopening probability. The infinite scroll isn't a UI choice, it's a specific counter-affordance to the natural stopping cue of a page break. Every one of these is a technical decision made by an engineering team, and every one of them is aimed, with real skill and real budget, at the exact cognitive faculty you need to do your own job.
Community reaction split usefully. One camp: this is a cop-out, focus is a discipline, put the phone in a drawer. The other camp: individual willpower against a trillion-dollar attention supply chain is not a fair fight, and pretending it is lets the supply chain off the hook. Both are right about something. The individual can absolutely install stricter defaults. The industry has absolutely built a machine that makes those defaults necessary in the first place.
Treat focus like you treat any other constrained resource: measure it, budget it, and remove contention. Concrete moves that showed up repeatedly in the thread and are worth trying:
Kill notifications at the OS layer, not the app layer. App-level notification settings are designed to be re-enabled — every update nudges you back toward defaults. macOS Focus modes and Android's Do Not Disturb, configured once as an allow-list (calls from three people, nothing else), are harder to erode. The rule that actually works is not "turn off notifications" but "nothing may interrupt me except a small explicit list of humans."
Batch, don't stream. Slack and email checked at fixed times (e.g., 10am, 1pm, 4pm) recovers more cognitive budget than any other single change. The pushback is always "but what if something's urgent" — in almost every engineering org, the actual urgent path is a phone call or a page, not a Slack DM. If it isn't, fix that instead of surrendering your attention to hedge against it.
Reintroduce boredom deliberately. Several commenters described a version of the same experiment: leave the phone in another room for a week, and by day four the ability to sit with a problem returns. The mechanism isn't mystical — it's that the compulsive check-and-scroll loop is a habit, and habits die when the cue is removed. If you can't do a week, do an afternoon.
Audit your IDE. AI autocomplete that fires on every keystroke is a productivity win *and* an attention tax — you're now context-switching between writing and evaluating suggestions dozens of times per minute. Some developers in the thread reported turning off inline suggestions and only invoking them on demand, and getting back a quality of thinking they hadn't realized they'd lost. Worth an experiment.
The honest read is that this gets worse before it gets better. The economic incentives point one direction — every consumer software company's growth model depends on capturing more of your session — and no regulatory or market force is currently pointing the other way. The near-term response has to be individual and defensive: fewer surfaces, stricter defaults, and a working assumption that anything designed to be engaging is designed at your expense. The longer-term response probably looks like a market for tools that explicitly refuse to optimize against you, and a professional culture that treats focus the way it treats sleep — as a non-negotiable input to good work, not a personal virtue. We build the machine. We can also decide what we let it do to us.
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.