Find the first layer.
Follow citations, reposts, and recurring claims back to accessible primary documents. Distinguish the earliest source we can verify from the true origin we may not know.
Question: Where did this claim come from?Investment narrative research
Follow the idea.
Find the evidence.
Investment ideas spread faster than their evidence. We’re building a research platform to trace where a thesis comes from, how it gains momentum, and what actually supports it.
Early-stage concept. Designed for curious, skeptical researchers.
Example thesis · AI infrastructure
The next bottleneck isn’t compute.
It’s power.
Narrative diffusion
From technical constraint to investment thesis
Trace the evidence, not just the mentions
Primary sourcesFilings & technical reports
→
InterpretationAnalyst & expert framing
→
Social diffusionDiscussion on X & Threads
Hypothesis, not a conclusion
Power constraints may be turning an engineering problem into a market narrative. Research should distinguish regional bottlenecks from a broad investment claim.
CHECK AGAINSTEfficiency gains, capacity additions, project delays, and valuations that may already reflect the story.
Research built around
why an idea travels.
Beyond social sentiment
A popular idea is a starting point.
Not an investment thesis.
We want to connect social conversation to the underlying events, incentives, and evidence that give it meaning.
Explain why it travels.
Study how a narrative changes across communities, what events may have accelerated it, and whether new attention brings new evidence or simply repeats an old claim.
Question: Why is this gaining attention now?Build a defensible view.
Turn fragmented discussion into a source-backed research note: the claim, its supporting evidence, competing explanations, and the observations that could change the view.
Question: What would make this thesis wrong?Proposed research workflow
From scattered conversation
to structured research.
The goal is a reproducible research process, not a black-box score or a feed of confident predictions.
Start with a thesis
Identify a claim or theme in permitted social data, user-provided links, and documents.
Connect its sources
Extract claims and entities, reconstruct citation paths, and retrieve accessible primary evidence.
Test the explanation
Compare catalysts, incentives, timing, and counterevidence. Separate observations from hypotheses.
Review the research
Produce a cited brief with gaps, alternative interpretations, and a clear path for human verification.
Planned use of Claude: source-grounded claim extraction, comparison across documents, and generation of concise research explanations for human review. Social-data access will depend on authorized APIs, permitted integrations, and applicable platform terms. No X or Threads integration is currently live.
A look at the research experience
Make the reasoning inspectable.
Explore an illustrative note. These are research questions and hypothetical explanations, not findings from collected market data.
RESEARCH MAP / SOURCE TRAIL
The claim needs a chain of evidence.
“Power is the next AI bottleneck” contains several distinct claims. Each needs a source, a time frame, and a boundary on what it actually proves.
A
Check the physical constraint
Seek grid-connection timelines, utility disclosures, and local capacity data. Do not generalize one region’s problem to every market.
B
Check the economic transmission
Ask which businesses capture value, under what contracts, and over which investment horizon.
C
Check how the claim was reframed
Compare the original document with later commentary. Identify where context or caveats disappeared.
Illustrative interface and synthetic research outline. Decorative bars above are not measured attention or performance data.
Research, not reassurance
Confidence should come
from evidence.
Not repetition.
Not Just Buzz is intended for independent investment researchers, analysts, and research teams who want to understand a narrative before acting on it.
We are exploring how quantitative measures of diffusion can complement qualitative, source-grounded analysis. An attention metric is not a measure of truth, and an explanation of momentum is not proof of causality.
- Provenance before popularityKeep sources and the limits of coverage visible.
- Claims are not factsSeparate what a source says from what the evidence supports.
- Counterevidence is part of the outputShow what challenges the thesis, not only what confirms it.
- Human judgment stays in the loopAI helps organize research. Researchers verify and decide.
Frequently asked
A few important
distinctions.
Is Not Just Buzz available today?
Not yet. This page presents an early-stage product concept and an illustrative research experience. It is not a live analytics service. Product scope, data access, and implementation are still being developed.
Is this a sentiment dashboard or trading signal?
No. The proposed focus is narrative provenance, diffusion, evidence, and alternative explanations. We do not claim predictive returns, a validated quantitative model, or an automated trading capability.
How will X and Threads data be accessed?
We plan to evaluate authorized APIs, permitted integrations, and user-provided source material. Availability, coverage, and retention will depend on provider permissions and terms. No comprehensive monitoring or partnership with either platform is implied.
How does Claude fit into the product?
Our planned workflow uses Claude to help identify claims, compare source documents, and compose cited research explanations. Outputs would be treated as fallible analyses for human review, not independently verified facts. There is no live Claude integration in this concept preview.
What about privacy and investment advice?
This static preview has no signup form, analytics scripts, cookies, or data collection functionality. It does not offer personalized investment advice or buy/sell recommendations. Production privacy policies and data handling practices will be published when the service is ready.
The project
A better question than “what’s trending?”
Why is this idea spreading, where did it begin, and what would justify believing it? That is the research experience we want to build.
Contact: [email protected]