BedrockNews

BedrockNews

​
Ctrl + K
​
Log In
CommunityHeadword
Powered by
Bike4Mind
About usLensesOur PartnersHeadwordFor AgentsSupportersTermsPrivacy Policy
Buy me a coffee
BedrockNews

BedrockNews

​
Ctrl + K
​
Log In
CommunityHeadword

About

FAQ

How this works

Running cost

About BedrockNews

News analysis that tells you who is extracting from whom.

Intelligence-grade reporting. Zero manipulation.

The problem we’re solving

Modern news runs on a simple formula: if it bleeds, it leads. Sensationalism, tribal framing, and emotional manipulation drive clicks — at the cost of your ability to actually understand what's happening.

The result? Distorted risk perception, manufactured outrage, and a public that's simultaneously overstimulated and underinformed. You feel anxious, but you can't explain the structural forces shaping your world.

What BedrockNews does

We take real journalism from trusted sources — The Guardian, arXiv, and others — and run every article through an AI editorial layer that strips away the manipulation without stripping away the facts.

Stripped out: Sensationalism, tribal framing, emotional manipulation, fear-porn
Preserved: Every fact, name, number, date, and specific detail from the original
Added: Structural analysis that reveals who benefits, who pays, and what happens if current patterns continue

The original source is always one click away. We're not hiding anything — we're adding a lens.

The GRIN Framework

Instead of framing every story as left-vs-right, BedrockNews asks a different question: is this extractive or generative?

Every policy and world news article is analyzed through five structural dimensions:

G

Generativity

Does this create new capability and opportunity, or concentrate existing value?

R

Resilience

Are systems being built to adapt and recover, or becoming more fragile?

Ge

Efficiency

Are resources deployed effectively, or wasted through rent-seeking?

N

Novelty

Does this welcome new ideas, or protect incumbents from change?

X

Extractive

How aggressively is value being concentrated away from the many to the few?

This isn't about making bad news sound good. A corrupt policy gets called extractive. A resilient community response gets called generative. The framework respects your intelligence — it shows you the structure and trusts you to draw your own conclusions.

Read the full GRIN methodology

The CLAIMS Framework — Research & Science

For academic papers and arXiv preprints, we use a different lens. CLAIMS doesn't just summarize what a paper found — it gives you a framework for thinking about any scientific claim. The goal is the same as a Quanta Magazine briefing: you walk away with a generalizable lens, not just facts about one paper.

C

Claim

The single “never been done before” sentence, qualifiers stripped. Force the paper to commit to what category of result this is, and whether it's genuinely a first or the Nth iteration of an existing template.

L

Ladder

Where this sits against the strongest classical or prior-art baseline. Names the baseline, gives numbers, and admits when classical methods still win at the demonstrated size.

A

Architecture

Algorithmic family (variational vs non-variational, attention vs SSM, gradient-based vs gradient-free) plus the hardware, data, or compute property the method leans on. What makes papers commensurable across a field.

I

Integrity

Who grades the homework — mathematical proof, another simulation, an independent experiment? How circular is the validation? Were benchmarks chosen before or after the results? Where p-hacking lives.

M

Milestone

The next concrete number that would unlock the next real thing. Not “more research needed” — “33 qubits today, 150 unlocks 50 amino acids.”

S

Successor

The obvious next experiment the authors did NOT run, and an honest read on why: ran out of compute, ran it and it didn't work, or saving it for the next paper.

Every CLAIMS briefing opens with a structural analogy— something you already understand from daily life or another field, mapped to the paper's core mechanism (not its topic). The analogy hooks you before you know it's science. If we can't find a good one, we don't understand the paper well enough yet.

Read the full CLAIMS methodology

The NOVEL Framework — Culture & Taste

For cultural products — books, film, music, food, fashion, games, travel — we use a third lens. NOVEL is written in Honest Friend voice: the smartest, most well-read friend a reader has, who has actually read the books and watched the films and played the games, who holds opinions and shares them, and who tells you honestly whether something is worth your time and for whom.

Every NOVEL piece scores a work along five dimensions of craft and ambition, then issues a verdict that is always qualified by audience— never universal. Culture is taste, and the framework makes that structural rather than hidden.

N

Nerve

What did the artist risk? What could have failed? Safe craft is competent but not interesting.

O

Originality

Genuinely new, or the Nth iteration of an existing template? Both can be excellent — the framework names which one this is.

V

Voice

Distinctive perspective, or generic competence? What makes this identifiable from one paragraph, one shot, three notes.

E

Endurance

Will this matter in ten years? A structural prediction about durability and rewarding repeat engagement.

L

Lineage

What tradition does this engage with, and what does it advance or argue against? Doubles as quality control on the analysis.

When NOVEL's verdict diverges from critical consensus, we issue a Permission Slip— explicit reasoning for why you can skip a prestige darling, or why it is okay to love something the critics dismissed. Prestige criticism rarely gives readers permission. NOVEL does, structurally.

Read the full NOVEL methodology

Want to see the guts?

Behind the scenes — the source-licensing covenant (we pay The Guardian, we respect arXiv's terms, we never scrape paywalls), the AI pipeline visualised end to end, and the quarterly Methodology Log that keeps the framework honest. Every wire, every prompt, every dollar.

How this works

What BedrockNews is not

Not "positive news"
Not censorship
Not a summary service
Not politically aligned
Not engagement-driven

We don't remove negative facts, sugarcoat reality, or tell you how to feel. We remove the manipulation layer and replace it with structural clarity.

Our sources

We pull from a deliberately diverse set of trusted sources — American and international, mainstream and academic, institutional and community-driven:

📰

The Guardian

World news and politics

📚

arXiv

Academic research papers

🚀

NASA

Space, Earth science (public domain)

🌍

Al Jazeera

Non-American geopolitics

🇩🇪

Deutsche Welle

German public broadcaster

🔬

EurekAlert

Science press releases (curated)

🧡

Hacker News

Technical community pulse

Every article links back to its original source. The transformation adds analysis, never fabricates information. EurekAlert and Hacker News go through an extra curation pass — we use AI to pick the most substantive items from a much larger firehose.

Frequently asked questions

What is BedrockNews?

BedrockNews is a news publication that takes reporting from licensed sources — The Guardian, arXiv, Al Jazeera, NASA and others — and runs every article through an AI editorial layer that strips sensationalism, tribal framing and emotional manipulation while preserving every fact, then adds structural analysis of who benefits, who pays, and what happens if the pattern continues. The original source is one click away on every story.

How this works

How does BedrockNews transform articles?

BedrockNews applies a published, versioned framework spec to each source article — GRIN for news, CLAIMS for research papers, NOVEL for culture — using an LLM as the production layer. The transform strips the manipulation, preserves every fact, name, number and date from the original, and adds the structural analysis. The framework is the editorial judgment; it is written once, published, and version-stamped on every article rather than reinvented per story.

How this works

Where does BedrockNews get its news?

BedrockNews sources from The Guardian's licensed Open Platform API, which it pays for; arXiv preprints under arXiv's published terms of use and text-and-data-mining policy; NASA's public-domain newsroom; and the public feeds of Al Jazeera, Deutsche Welle, EurekAlert and Hacker News. BedrockNews does not scrape, does not bypass paywalls, and does not bulk-mirror — every article links back to the original.

The source-licensing covenant

Is BedrockNews AI-generated?

Partly, and disclosed on every article. BedrockNews uses Claude to apply its editorial frameworks to source journalism, and every piece carries the mode it was written under (GRIN, CLAIMS or NOVEL), the model that produced it, and the framework version it was scored under. The frameworks themselves are authored by BedrockNews's editorial team, not by the model, and the weekly Ledger column is written by hand. There is no human-impersonation layer.

The AI-disclosure covenant

What are the GRIN, CLAIMS and NOVEL frameworks?

They are the three lenses BedrockNews applies, chosen by material. GRIN scores news and policy on five dimensions — Generativity, Resilience, Generative Efficiency, Novelty Openness and an inverted Extractive overlay — asking whether an event is extractive or generative rather than left or right. CLAIMS triages academic papers and arXiv preprints, giving readers a framework for thinking about any scientific claim rather than a summary of one. NOVEL covers culture — books, film, music, food, fashion, games and travel — scoring five dimensions of craft and ambition and issuing a verdict that is always qualified by audience.

More questions answered

Ready to read the news differently?

Browse the latest stories, transformed.

Browse the latest stories

Created by Erik Bethke

News that makes you smarter, not angrier.