What is “Healthy” News Consumption?

Concept:
In today’s world of limitless news sources and algorithmic feeds, “healthy” consumption means:

  • Staying informed without becoming overwhelmed, manipulated, or anxious.
  • Being intentional rather than passive (“engage” rather than “drift”).
  • Building awareness across perspectives, but protecting focus and well-being.

Engage vs. Abstain:

  • Engage = Selectively, actively seeking news, filtering sources, reflecting (critical thinking).
  • Abstain = Limiting or avoiding certain types of news/sources or reducing overall consumption to improve well-being.
  • “Healthy” likely means a dynamic balance — engage where it matters, abstain where it’s toxic or irrelevant. checking.

So WHY do this really?

Improve quality of news consumption > Improve accuracy and ”quality” of my own beliefs (unlearning expired and inaccurate beliefs) > enjoy and balance the good with the bad

StepPassive DriftHealthy Engagement
SourceAlgorithm feedCurated, diverse sources
FrequencyAll-day scrollScheduled, limited time
ConsumptionSwiping, auto-playIntentional, selective
Capture/ReflectNoneNote-taking, analyzing
SocialComments, reactsTrusted discussion groups
AbstainNeverRegular detox, breaks

0. Design Brief

  • drift
    • by YouTube algorithms, including being drawn by clickbait title
    • editors in sources followed
    • Email bulletins
    • Word of mouth
    • Distractions, biases and echo chambers from all of above (”information overload”)
  • design
    • based on my context and relevance (personalisation)
    • with AI as a “news and media consultant” (NMC)
      • Curate and summarize news.
      • Suggest trusted sources.
      • Highlight bias/conflicts.
    • workflow above with automation (n8n)
      • setup, design, feed, receive, log, analyse….
      • Automatically pull, filter, summarize news.
      • Capture key points to notes.
      • Remind to revisit, reflect, or share with a group.
  • related solutions
    • offload ALL email bulletins from my main email address
      • Use my gmail.com address to ”firewall” my custom domain?
    • customize my YouTube feed, and assert MY algorithm
      • ..or take a feed and then use another client to see new videos?
    • create an RSS feed solution that my new NMC will work and iterate / evolve?
    • get my NMC to create an audio news bulletin e.g. 15-30 min that I can listen to as I wake my cortisols up in the morning in the park?
      • publish here for anyone else interested
      • whose voice should i use? (If mine then that WOULD be an echo chamber!)

In 2026 I would like to start putting design to my entire workflow of news and how I - consume visually with video (youtube) and diagrams (tldr news) - capture, reflect and analyse via my notes tool - discuss with my tribes (assuming they want to)

News Source Audit and Selection

Pay for vs. Free:

  • Paid: Often (not always) higher quality, fewer ads, less clickbait. Supports independent/journalistic work.
  • Free: Wider access, but beware of biases, clickbait, ad-driven content.

Independent vs. Mainstream:

  • Independent: Greater diversity of voices, sometimes more depth or bias.
  • Mainstream: Established fact-checking, but may carry institutional or political bias.

With and Without Bias

1. Discover

What should my portfolio of sources be?

  • soon Use cross vault manager to scan my Obsidian vault and give me a starter for 10

    • Decide time window (e.g., once per day, max 30 mins).

Automate & Filter:

  • Use RSS, newsletters, or automation tools to collect headlines.
  • AI summarizes and scores for relevance, bias, urgency.

2. Consume and Engage

  • Choose what to dive into based on interest and importance.
  • Take your own notes, critically analyze.

Pause/Abstain

  • Set rules for disengagement (e.g., digital Sabbath, muting certain topics).
  • Inform NMC so it can handle when I unpause

3. Reflect, Xpress & Discuss:

  • Periodically review your news notes.
  • Discuss with thoughtful communities for context/fact-

A. Spinup Plan

Step 1: Clarify Your “Why”

  • Decide why you want to change your news diet. (e.g., Avoid overwhelm? Improve quality? Stay relevant?)

Step 2: Select Trusted Sources

  • List 3–5 trusted mainstream sources (BBC, NY Times, Reuters, etc.)
  • List 2–3 independent and/or international outlets
  • Choose 1–2 newsletters, podcasts, or YouTube channels for deep dives Need to Audit : How can I get an extract of my YouTube sources? Similarly, how do I plug my AI tool (mammouth.ai including mammouth CLI) to my inbox to see what my sources are?

Step 3: Set Up Discovery

  • Use RSS reader (Feedly, Inoreader, or a personal reader with n8n)
  • Subscribe to selected newsletters
  • Use YouTube subscriptions (not “Up Next” recommendations!)

Step 4: Automate Filtering (w/ AI, n8n, or simple rules)

  • Collect headlines and summaries automatically each morning
  • Apply filters (topics, keywords, source credibility, etc.)
  • Label urgency (e.g., “must read”, “interesting”, “skip”)

Step 5: Mindful Consumption

  • Block time: Only “consume” news at set times (e.g., lunch break, 6–6:30pm)
  • Use visual & text media intentionally (watch, read, don’t scroll endlessly)
  • Capture insights/notes on key stories in your notes tool

Step 6: Reflect and Analyze

  • Once a week, review news notes. What themes? What changed your mind?
  • Use AI to generate quick summaries, trends, or bias checks

Step 7: Discuss and Debate

  • Share select stories or notes with trusted communities/friends
  • Keep discussion respectful, curious, evidence-based

Step 8: Schedule “Abstain” Days

  • E.g., one digital Sabbath per week; no news feeds, no social scroll

Step 9: Adjust and Review

  • Monthly: Audit your sources and routines. Are they serving your goals? Refine!

B. Workflow Diagram (Text/ASCII)

Here’s a concise, logical model you could plug into a tool like n8n, Obsidian, Notion, or even draw:

                        +----------------------+
                        |   Trusted Sources    |
                        |(RSS/Newsletter/Video)|
                        +-----------+----------+
                                    |
                                    v
            +--------------------------------------------+
            |    Discovery & Collection Automation       |
            |  (n8n, RSS Reader, AI Summaries, Folders)  |
            +-------------------+------------------------+
                                |
                     +---------------------+
                     v                     v
+--------------------------+   +------------------------+
|    Filtering & Labeling  |   |     Mindful Queue      |
|(AI tags, Topics, Urgency)|   |   ("Read Later", etc)  |
+------------+-------------+   +---------+--------------+
             |                           |
             +--------------+------------+
                            v
          +-------------------------------+
          |   Scheduled Consumption Block  |
          | (Read/Watch + Note Insights)   |
          +----------------+--------------+
                            |
                            v
           +-------------------------------+
           |  Reflection & Synthesis       |
           | (Review Themes, Bias, Impact) |
           +-------------+-----------------+
                         |
                         v
          +-------------------------------+
          |    Discuss/Share Selectively  |
          | (Tribe/Community/Notes)       |
          +-------------------------------+
                         |
                         v
             +------------------------+
             |   Rest/Abstain Period  |
             +------------------------+
                         |
                         v
                (Review, Repeat, Adjust)

Workflow (Short Version for Automation)

  1. Sources ➔ Discovery (RSS/newsletters) ➔ AI Filtering (topic, bias, urgency) ➔ Consumption window ➔ Capture to notes ➔ Reflect & summarize ➔ Share ➔ Abstain ➔ Audit.

C. SpinUp Step 2 - Sources

Okay, i have consolidated/ modified my chat above distinguishing between my in-life workflow and the spinup plan, of which I did step 1. I have yet to fire up and learn n8n, but wondering how I can take an audit of my current YouTube subscriptions and email bulletins to start taming that beast (as per my design brief)?


role::assistant (gpt-4.1)

Absolutely, here’s a structured approach to auditing your current YouTube subscriptions and email bulletins—an important step before automating or reorganizing your information flows.


1. YouTube Subscription Audit

A. Get a List of Current Subscriptions

used Google Takeout service and subscriptions available as csv

B. Analyze and Categorize

  • Create columns in your spreadsheet:

    • Category (e.g. education, entertainment, tech, etc.),
    • Frequency watched,
    • Value (High/Med/Low),
    • Action (Keep/Drop/Review).
  • Go through each subscription:

    • Unsubscribe right away if you’re sure you no longer want it.
    • Mark others for review based on your categories and value ranking.
  • Google Takeout can export your YouTube watch history and subscriptions, including:

    • Watch time per channel
    • Frequency of viewing
    • Video-level engagement

    👉 How to get it:

    1. Go to Google Takeout
    2. Select YouTube and YouTube Music
    3. Choose Subscription management and History
    4. Export and download the data (JSON or HTML)
    5. Use tools like Google Apps Script or Python to parse and analyze watch frequency. DONE - over here

C. Decide What to Keep/Optimize

  • Aim to reduce noise by removing channels that no longer serve your goals (as per your design brief).
  • For channels you love, consider grouping them in YouTube folders (using browser extensions like PocketTube).

2. Email Bulletin Audit

A. Gather All Subscriptions

Manual Sweep

  • Search your inbox for terms like “unsubscribe”, “newsletter”, “subscription”, or “bulletin”.
  • As you find relevant emails, make a list of senders/brands in a spreadsheet.

Inbox Management Tools

  • Services like Unroll.Me or Leave Me Alone can scan your inbox and present a list of newsletters/subscriptions, letting you unsubscribe in bulk.
  • Use with caution if privacy is a concern.

B. Analyze and Categorize

  • Same approach as with YouTube.
  • For each newsletter:
    • Decide: Do you read it regularly? Is it adding value? Is there overlap with another source?
    • Keep, Unsubscribe, or Move to a review list.

C. Take Action

  • Unsubscribe from low/no-value senders (use links at bottom of emails or tools mentioned above).
  • Set up labels in your email client for newsletters you want to keep, filtering them away from your main inbox.

3. Document Your Insights

  • Note which sources produce the highest value.
  • Note any subject areas you’re oversubscribed in.
  • This insight will directly inform your n8n automation (e.g., which feeds you want to process, summarize, or deprioritize).

4. Prepare for Automation

  • Once you have your curated list:
    • For YouTube, you can use RSS feeds (most channels still provide these) for automation.
    • For bulletins, use filtered/forwarded emails or newsletter-to-RSS services (like Kill the Newsletter!).
  • This clean slate will make your n8n flows more purposeful and reduce the initial chaos.

D. Optimise Youtube quality over time

you’re right to be annoyed — and also right that channel-level filtering alone won’t solve this. good creators drift, incentives change, titles get worse. if you don’t account for that, your system will slowly degrade.

the good news:
✅ you can bring in quality signals
❗ but “easy” comes with trade-offs


handling clickbait practices

YouTube does have rich signals:

  • views
  • likes
  • subscriber count
  • watch time (not accessible)
  • engagement ratios

👉 but:

  • RSS feeds do NOT include most of these
  • the official YouTube API does — but adds complexity (API key, quotas, etc.)

so the question becomes:

how far do you want to go vs how clean you want to keep your system?


 ✅ 3 practical levels (pick your lane)

 🟢 Level 1 — “good enough, low friction” (start here)

stay within RSS + AI

you already have:

  • title
  • channel
  • publish date

👉 add AI scoring in n8n

example prompt:

“Score this video for signal vs clickbait. Penalise emotional language, exaggeration, vague claims. Reward specificity, neutrality, and depth.”

you can also:

  • flag words like: “shocking”, “you won’t believe”, “collapse”, “insane”

✅ pros:

  • super easy
  • no API needed
  • aligns with your “NMC” idea

❗ cons:

  • no hard engagement data

 🟡 Level 2 — “light enrichment” (sweet spot)

augment selectively with YouTube API

for each video, pull:

  • view count
  • like count
  • channel subscriber count

then compute simple signals:

 example metrics

  • Like ratio [ \text{likes} / \text{views} ]

  • Velocity [ \text{views} / \text{hours since publish} ]

  • Channel size normalisation (small channel with high engagement = gold)


👉 then combine with AI:

“High engagement + low clickbait score = prioritise”

✅ pros:

  • much smarter ranking
  • still manageable

❗ cons:

  • requires API setup in n8n

 🔴 Level 3 — “your own recommendation engine”

this is where you’re heading long-term:

  • track what you actually consume
  • track what you skip
  • track what you save / reflect on

👉 then weight sources dynamically

this becomes:

your personal YouTube algorithm — but aligned with your values


 ⚙️ what I’d recommend for YOU (based on your design brief)

you’re not trying to optimise for scale
you’re trying to optimise for:

clarity, intentionality, and signal

so:

✅ combine these 3 layers

1. channel baseline (your sheet)

  • category
  • description
  • priority

2. video-level AI filter

  • clickbait detection
  • topic relevance
  • tone (reactive vs analytical)

3. optional: light engagement signal

  • only if needed later

 💡 a surprisingly powerful trick (low effort, high impact)

create a column:

CLICKBAIT_TOLERANCE

values like:

  • Low
  • Medium
  • High

then in n8n:

👉 apply stricter filtering to “Low” channels


 🧠 deeper point (this is the real game)

you said:

“genuine good content producers that stay that way”

that’s not static — it’s time-based trust

so eventually you want:

  • channels that earn trust over time
  • channels that lose ranking when they drift

👉 this is where your system becomes alive


 🚀 simple next step (don’t overbuild yet)

for now:

  1. ✅ use your sheet
  2. ✅ use RSS
  3. ✅ add AI scoring (clickbait + relevance)

ignore API for the first version


 if you want next

I can give you:

  • a ready-to-drop n8n AI prompt for scoring videos
  • or a scoring model (0–100) combining:
    • clickbait
    • relevance
    • depth

that’s the piece that turns this from a feed… into your NMC brain

✅ multiple RSS feeds for channel quality monitoring and management

What you’re describing is less “feed consumption” and more feed-driven review workflows — which is exactly where RSS + your sheet + n8n shines.

You can absolutely create multiple RSS-driven views from the same channel list, each serving a different purpose.


🧠 Concept: One Source → Multiple Feeds

Your Google Sheet becomes the control layer, and you slice it into different “virtual feeds”:

  • Daily Digest Feed → your core, high-value channels
  • Reconnection Feed → channels you haven’t evaluated recently
  • Experimental Feed → new or uncertain channels
  • Low Priority Feed → occasional browsing only

👉 same raw data, different intent


✅ How to structure this in your sheet

Add a couple of columns:

1. STATUS

  • Active
  • Review
  • Reconnect
  • Drop

2. LAST_REVIEWED

  • date you last intentionally checked the channel

3. (optional but powerful) LAST_VIDEO_SEEN

  • lets you track drift over time later

⚙️ How this becomes multiple RSS flows in n8n

Instead of one big pipeline:

Flow 1 — Daily Digest

  • Filter: STATUS = Active
  • Pull latest videos
  • Score + summarise
  • Output: your daily digest

Flow 2 — Reconnection Feed

  • Filter:

    • STATUS = Reconnect
      OR
    • LAST_REVIEWED > 30–90 days ago
  • Pull:

    • last 3–5 videos per channel

👉 output becomes:

“Here’s what this channel has been like recently”

perfect for deciding:

  • keep
  • downgrade
  • drop

Flow 3 — Review Queue

  • Filter: STATUS = Review
  • Lower frequency (weekly)

💡 Key trick (this is what makes it work)

For reconnection, don’t just pull the latest video — pull a small sample

In n8n:

  • loop per channel
  • fetch 3–5 recent items from RSS

👉 this gives you pattern recognition, not a one-off impression


🚀 What your reconnection output could look like

Instead of raw links:

Channel: X (AI commentary, previously high-value)
Recent videos:

  • “Why AI Will Replace Everything”
  • “This Changes Everything (Again)”
  • “The Truth They Won’t Tell You”

AI summary: Shift toward sensational framing, less technical depth

👉 Recommendation: Review / downgrade


🧠 Why this is powerful

You’re solving a real problem:

creators drift, but subscriptions don’t self-correct

This system gives you:

  • periodic re-evaluation
  • resistance to clickbait creep
  • a way to prune your inputs over time

⚡ Simple starting version (don’t overbuild)

You don’t need full automation yet:

  1. Add:

    • STATUS
    • LAST_REVIEWED
  2. Manually filter in Google Sheets:

    • “Reconnect” channels
  3. Open 2–3 videos per channel

👉 then later: automate this exact behaviour in n8n


✅ Bottom line

  • Yes, you can create multiple RSS-driven “feeds”
  • They don’t need separate infrastructure — just filters on your sheet
  • Your reconnection idea is not extra — it’s core to keeping the system clean over time

If you want, next I can sketch:

  • a clean n8n reconnection workflow
  • or a scoring + downgrade logic so channels automatically move toward “Reconnect” when quality drops