One feed for AI news, blogs, papers, repos, tools, and events — built it, and now questioning whether learning is the bigger problem

Keeping up with AI now means jumping across way too many places — X, GitHub, Product Hunt, YouTube, research papers, company blogs, newsletters, Reddit, events, random builder threads. I got tired of checking 10+ sources every day, so I built AgenticBrew to put it all in one place. What it does right now: \- Coverage — pulls from hundreds of sources daily: RSS feeds, headless scrapers, APIs, GitHub, Product Hunt, YouTube, X, AlphaXiv, Hugging Face, company/research blogs, newsletters, and event sources. \- Clustering — a model launch usually shows up as a blog post, gets discussed on X, lands in YouTube videos, triggers GitHub repos, and sparks Reddit/HN threads. AgenticBrew groups all of that into one story instead of 8 disconnected updates. \- Filtering & ranking — lightweight ranking to surface higher-signal items instead of generic AI noise. \- Categorization — everything sorted into AI news, research papers, technical/company blogs, tools/repos/products, events, and community signals. The source list evolves weekly and biases toward sources recommended by builders, researchers, KOLs, company blogs, and AI communities — most of the big names you'd expect are in there. ***Here's where I'd love this sub's input.*** While building it, the thing that became obvious is that the information layer is already crowded — there are tons of newsletters, aggregators, dashboards, and feeds. Just showing "more AI updates" isn't enough. The gap I keep hitting is learning. People come to AI from completely different roles and literacy levels — a designer, a PM, a backend engineer, a founder, and a marketer all need different things. Most don't need another feed; they need to actually get better at using AI for their own work. There's already a lot of genuinely good material out there (courses, tutorials, docs, talks), but it's scattered and not organized around who you are or what you already know. So instead of making yet another course from scratch, the direction I'm exploring is an aggregation + sequencing layer on top of existing courses: pull together the high-quality material that already exists, then assemble a personalized learning path based on your role and current AI literacy — not one generic "AI for everyone" course. Quick survey before I build more in that direction: 1. When you try to level up your AI skills, what's the most annoying part right now — finding good material, knowing the right order, not knowing what you don't know, needs more supporting materials/explanations/tutoring when learning a topic, or lacking practices aside from theoretical learning? 2. What's your role, and what would "getting better at AI" concretely look like for you? 3. Would you pay for a personalized path like this, and if so, what would it have to deliver to be worth it? Website: [https://agenticbrew.ai/](https://agenticbrew.ai/) Source list: [https://github.com/sunxiayi/awesome-ai-sources/blob/main/SOURCES.md](https://github.com/sunxiayi/awesome-ai-sources/blob/main/SOURCES.md)