YouTube’s recommendation algorithm doesn’t always get it right — the homepage is cluttered with videos you don’t care about, and the topic-based feed tabs can be surprisingly limiting. But YouTube’s new custom feed capabilities let you build your own algorithm with personalized tabs tuned to your actual interests. Here are five ways to make it better than the default.
YouTube app (Credits: How-To Geek)
1. Build a News Tab That Actually Matches Your Interests
The default news tab serves up world news you may not care about. Instead, create a custom tab with a prompt asking for recent, informative videos on specific topics — like mobile devices, gaming tech, and automotive technology — while excluding reaction videos, outrage content, and generative AI slop. The result: a news feed focused on longer-form, well-sourced videos you actually want.
2. Make an Educational Tab for Skills You Want to Learn
YouTube’s suggested education tabs often skew clickbait or off-topic. Build one around what you’re really learning — for example, vehicle maintenance and modification — asking for educational content on repairs and engineering while filtering out exaggerated claims, hype, and product shills. Teach yourself something instead of doom-scrolling.
3. Filter Out Creators and Games You Don’t Like
If your gaming feed is dominated by creators you find toxic or games you don’t play, a custom entertainment tab lets you name what to exclude. Ask YouTube for gaming entertainment filtered to your favorite titles, with unwanted creators and unrelated recommendations removed. You’re no longer hostage to whoever the algorithm is pushing this week.
4. Curate Hobby Tabs and 5. Prune the Clutter Regularly
Apply the same idea to hobbies — build tabs around niche interests the default algorithm never quite nails. And revisit your prompts over time: refine the wording, add new exclusions, and keep each tab sharp. A few minutes of tuning turns YouTube from an algorithmic slot machine into a tool you control.
Also Read: TinyFish Breaks Down Production-Ready Agentic Search Architecture for AI Agents
Source: How-To Geek — read the full story there.
