← Back to Dashboard

About Signals

A daily practice for scanning the horizon — surfacing unusual events, papers, projects, and announcements that hint at possible futures before they become trends.

"The future is already here, it's just unevenly distributed." — William Gibson

What This Is

Signals is a personal futures intelligence project. Every day, four AI systems independently scan the internet for weak signals — outliers, edge cases, and early indicators that suggest where things might be heading, typically 6–18 months before they become obvious trends.

The reports you see here are the unified output: a human-curated synthesis of what those AI scanners found, filtered for the weird, the surprising, and the potentially consequential.

What We're Looking For

🤖

AI & Agents

Unexpected capabilities, new frameworks, tools fighting AI side effects

⚛️

Quantum Computing

Error correction breakthroughs, new applications, accessibility milestones

🛠️

Indie Innovation

Garage projects, open source tools, one-person projects punching above their weight

🏛️

Policy & Security

Legislative shifts, digital sovereignty moves, infrastructure vulnerabilities

🧬

Bio-Digital

Health-tech crossovers, robotics, bio-computing interfaces

🎪

Delightfully Weird

Retro-computing revivals, bizarre experiments, things that make you go "wait, what?"

What We Skip

Routine product updates. Funding announcements (unless the structure is novel). Obvious hype cycles. Mainstream news that everyone already knows. If most of the signals come from Nature, TechCrunch, or Wired — we haven't dug deep enough.

The Pipeline

Each signal passes through a five-stage intelligence pipeline, from raw collection to actionable forecasting.

📡
Collect
Daily weak signal scanning
🔍
Process
Identify implications & patterns
🎯
Track
Define measurable thresholds
👁️
Monitor
Search at regular intervals
🔮
Predict
Forecast when futures arrive

📅 Signals Tab — Daily Reports

The raw collection. Every day, four AI systems scan independently and their outputs are unified into a single report. Cross-source signals (appearing in multiple AI scanners) are given extra attention since independent discovery suggests something real is happening.

🎯 Tracking Tab — Measurable Indicators

When a pattern of weak signals points toward a potential future, we ask: what specific, measurable thing would need to change for this future to arrive? That becomes a tracking signal — a concrete threshold we monitor over time. When a threshold is crossed, the signal "fires" and we know that future is closer than we thought.

🔮 Futures Tab — Possible Worlds

The synthesis layer. Eight axes map the landscape of possible futures, with positions based on accumulated evidence from daily signals and tracking data. These aren't predictions — they're a visualization of which directions the evidence currently points.

Four AI Sources

Each source uses a specialized prompt tuned for weak signal detection, with mandatory requirements for weird/delightful signals, counter-movements, and indie projects.

Grok Scheduled daily task with access to real-time X/Twitter data
Perplexity Private Signals space running daily digests with web search
ChatGPT Signals project with structured scanning across seven categories
Gemini Daily digest prompt scanning for discontinuities and weak signals

Evidence Standards

Not all evidence is created equal. We use a three-tier system to assess how much weight to give each data point — and actively watch for hype.

Tier Type Examples
Tier 1 Primary measurement Peer-reviewed paper, benchmark result, SEC filing, shipped product, enacted law
Tier 2 Official claim Company blog, press release, pre-print, announced partnership
Tier 3 Secondary report News article, analyst note, social media post, community discussion

Hype Detection

When a topic has lots of Tier 3 coverage but little Tier 1 evidence, that's a hype signal — many people talking, few verifiable results. Conversely, steady Tier 1 evidence accumulating quietly is a substance signal. We flag the difference.

Guiding Principles

The Unification Process

Each morning, a Claude agent visits all four AI sources, extracts their reports, and synthesizes them into a single unified document. The process prioritizes: