How I use Bubbles to find the quirky stuff 🎲
It always warms my heart to read kind blog posts with interesting ideas for making Bubbles even better. Of course I can't just keep adding features: Bubbles is already packed with stuff, and I want to keep it minimalistic and intuitive.
The main purpose of Bubbles is to find interesting things to read, and what counts as "interesting" is very subjective. To find posts they like, readers either search for blogs and topics, or they filter out the things they don't want to see. There is no single view where every reader magically sees only the posts they find interesting. I definitely don't want to build the next engaging, addictive TikTok algorithm. I want to keep the views simple to understand and the same for everybody (except "My", which shows new posts from the blogs you follow).
Scope creep can be a real problem. More users mean more ways to use a tool like Bubbles. More use cases mean more feature requests. More features mean more menus, documentation, distractions and changes. I always try to balance how many features I add, and which ones. If I can see myself using a feature regularly and I think it improves Bubbles for many users, I'm more likely to add it.
I built Bubbles to stumble across weird, curious, niche posts from personal blogs whose authors aren't blogging to become influencers. So I thought I'd share how I use Bubbles myself, and which features I've added recently:
- "Random" view: In response to Gordon's blog post I added the "Random" view, which gives older posts from the archive a fair chance to show up again. "Top" shows the posts that received votes recently, "New" shows all newly published posts, and "My" also shows the newest posts, but only from the blogs you follow. "Random" doesn't simply pick 30 random posts from the database, because that would favour bloggers who post more often. Instead it first draws 30 random blogs and then picks a random post from each of them. That way every blog has the same chance to show up. "Show more" draws a new set of 30 blogs.
- Filters: Hakkerman's blog post suggested a "Fresh" page that leaves out blogs which are on Top all the time. Instead of adding another view (remember scope creep), I added a small filter option, "Hide top blogs", which has a similar effect. It hides blogs whose posts received votes on 10 or more of the last 30 days, spread over at least 3 different posts. Their posts then disappear from Top, New, Hot, Random and the search, so the blogs that show up over and over again make room for others. If you're signed in, you can also hide individual blogs completely. I also added a word filter that can be switched on and off. My word list currently contains "ai", "llm\*", "generative", "anthropic", "openai", "claude", "codex", "chatgpt", "amodei", "altman" and "zuckerberg".
On top of that I filter out the "Tech" category, set the vote filter to "to 0" (so I only see posts nobody has voted for yet) and tick "Longreads only (5+ min)", which only shows posts with 1,000 words or more. By the way, you can click the "5 min read" label on any post to switch between minutes and words. As I don't scrape any websites, I can only calculate the reading time from the feed content. So posts from blogs without full-text feeds never get the label, and the Longreads filter hides them.
For me, the "Random" view combined with these filters (non-tech longreads without certain keywords, from blogs that haven't been on Top much lately) works very well for discovery. It's certainly not the best setup to find something specific or trending, but it's excellent for the quirky, niche stuff I'm mostly looking for when I browse Bubbles.