August 16, 2026 · 5 min read
Blocklists lied to you
You installed a focus app. You banned YouTube, Twitter, Reddit. For about three days it worked. Then you needed a lecture that happened to live on YouTube, and the wall you built to protect yourself became the thing in your way. So you added an exception. Then another. Within a week the blocklist was a colander, and you were back to surfacing forty minutes deep in a video essay about medieval siege weapons.
This isn't a willpower problem. It's a model problem — the mental model every mainstream focus tool is built on is wrong.
The blocklist fails in both directions
Domain blocklists — Freedom, Cold Turkey, BlockSite, StayFocusd — all ask the same question: which website is this? That question is too blunt and too leaky at the same time.
- Too blunt. YouTube contains the lecture you need. Reddit contains the exact answer to the bug you're chasing. Banning the domain bans your own work.
- Too leaky. Distraction lives everywhere — a news site, a shopping cart, a random blog you'd never think to list. None of them are on your blocklist, and each one quietly costs you twenty minutes.
You can't list your way out of this. The set of "sites I sometimes need" and the set of "sites that sometimes distract me" overlap almost completely. A blocklist forces one rule onto both.
The better question
The question that actually matters is never which website is this? It's:
Does this page belong to what I'm doing right now?
That's a question about meaning, not addresses. And once you ask it that way, the two-directional failure disappears. The chemistry lecture on YouTube belongs to "organic chemistry exam prep." The trending video on the same domain does not. Same site, opposite verdicts — because the page is judged by what it's about, not where it lives.
This is the whole idea behind Solidmind. You declare your topic in plain words, and a small AI model reads each page's meaning and compares it to that topic. On-topic pages are left completely alone. Off-topic pages get a soft blur and a plain-English reason.
Why blur instead of block
Notice the second word: it blurs, it doesn't block.
A hard block page feels like punishment — a locked door with a stern message. A blurred page with a short explanation feels like a coach's hand on your shoulder: "hey, this doesn't look like what you sat down to do." You can see there's something there, you understand why it faded, and there's always a way out sized to how serious you told it to be.
That difference matters because the system isn't perfect, and it doesn't pretend to be. On genuinely ambiguous pages it will be wrong sometimes. A block that's wrong is infuriating. A blur that's wrong is a half-second shrug — you dismiss it and move on. Recoverable beats strict-but-brittle every time.
"So it reads every page I open?"
Yes. That's what makes it work — and it's worth being direct about where that reading happens.
Analyzed on your device by default; nothing is sent anywhere unless you explicitly enable a clearly-labeled cloud feature.
The model runs inside the extension. Understanding a page requires seeing the page; the important part is that seeing it doesn't mean shipping it somewhere. By default, page content doesn't leave your browser.
The point isn't focus. It's finishing.
Nobody wants "focus" as an end in itself. You want the exam grade, the shipped project, the report that was due. Focus is just the road there — and a blocklist keeps rerouting you off it.
Stop maintaining a list of enemies. Say what you're working on, and let the pages that don't belong quietly fade while you finish the thing you actually sat down to do.