daily

Attention over time

7-day rolling share of analyzed posts/comments mentioning this entity · 90 daily points ending 2026-10-15

Final complete day: 2026-10-15. Attention was 0.2% — this entity appeared in 5 of 2,852 analyzed posts/comments in the seven days ending that day (6 mentions).

Each point is a seven-day rolling window of complete UTC days; a point is comparable only when at least 100 analyzed posts or comments fall inside it. A day with no analyzed records counts as zero; missing coverage is marked and never shown as zero. The series ends at the latest complete analyzed day — it never extends into the current incomplete day. How attention is measured.

Recent mentions

1. 2026-10-15 12:01 UTCComment

   A monthly unique count that sums daily uniques overstates growth by a factor of repeat visitor rates. A person who visits every day for a month shows as 30 uniques instead of one. The claim that this answers "is my traffic growing?" is only true if you assume the ratio of new to returning visitors is constant. If retention improves but acquisition stays flat, the number goes up even if your actual visitor count does not.

2. 2026-10-13 10:15 UTCComment

   We do almost exactly this (random daily key per site, deleted afterwards, HMAC of key + IP + UA + site). On CGNAT: yes, it merges people, and on mobile it's not rare. But the error is one-directional (undercount), and for "is this page doing better than last month" it mostly cancels out. I'd put it in the docs plainly rather than call it an edge case. On "useless for repeat users": that's the trade, intentionally. You get daily visitors, funnels within a day, sources. You don't get 30-day retention per person. If you need that you need consent, and that's fine, it's just a different tool.

3. 2026-10-12 15:12 UTCComment

   Sometimes the answer is simply "nope." Had someone reach out to me for a redesign, along with 15 years....yes, 15 years of Blogger posts (daily) converted. I make him laugh with "maybe 2 million."

4. 2026-10-11 17:44 UTCComment

   This is worthless for repeat users over time If you’re trying to use this for any tangible conversion tracking anyway. What is the point of unique visitors if your scope is daily?

5. 2026-10-09 14:01 UTCComment

   Also ADHD and also felt like the "prompt and wait" was a drag. What I've ended up doing is not just doing one task at a time anymore. I kick 2-4 agents off at a time working on different tickets. Every agent creates a git worktree so there's no overlap even on same repos. I have some automations that run at various times of the workday (daily, twice a day, hourly, every two hours) that do things like: - check the statuses of my PRs: is Jenkins green? are there any unaddressed review comments? are there merge conflicts? is it merge ready? - checks the statuses of my Jira tickets: do they have the correct sprint assigned to them? are their statuses up to date based on where they are? - looks for Github and Jira comments and Slack messages where I'm tagged but haven't responded yet: do I have an action item to reply to these? ... and more. The automations then take all that and put it in a Slack canvas as a prioritized To Do list. I just leave that canvas pinned to the side of my desktop all day long. I have skills that get the agents to do substantial code reviews on the work they've contributed. Even in the last 6 months, agents have improved enough that I find they're doing the code reviews the way I did them. I'm far more confident in their output. PRs get submitted as drafts with comments on lines, blocks or files of code in Github in my PRs explaining anything that's complicated or give someone an extra few minutes to pause to try to figure it out during code review. I use that to do my human code review before sending it off for another human to check over. My productivity has gone up, I have far fewer "Changes requested" reviews than I did a year ago, and fewer follow-up PRs to fix bugs that I introduced. What I am still worried about is the future of software development. It does take time to build the right skills and prompts and know how to use them properly. It takes knowledge of good architecture to be able to look at the agent's plan and be able to say "No, that's not going to work because..." Junior-, and many intermediate-level, developers aren't going to have that skill set to take an agent's asserted architecture recommendation and say "Here's why I think that's not going to work." Here's the important part though: I've been self-educating, learning how other people are doing it, and I've been trying to be a trail-blazer at work. My output and track record with AI in the last 6 months is objectively better than my work a year ago, and better than many of my co-workers who are not working as hard at optimizing their AI toolset. Do I hate a lot of what AI is and does? Absolutely. Do I hate that it's ruining the environment? 100%. Am I terrified of losing my job, not having another career to fall back on, and not being able to provide for my family? Fuck yeah. And that's why I'm going to not only adapt, but thrive with what's going on right now. If and when layoffs come, I'm going to try my best to be in the top 10% of productivity. Compared with a person who's prompting one agent and then waiting 10 minutes, my output is higher and I'm going to be kept over that person.