What is AI-generated here

Every analysis, report, brief, signal, message, chart caption, and generated image on this platform is produced by automated systems — language models, vision-language models, and software agents — operating under human-architected pipelines and guardrails. Content is not authored by humans; the systems that produce it are. Treat all output as machine-generated unless a specific surface explicitly attributes content to a named human author.

Human architecture & oversight

The systems producing this content are human-architected, human-directed, and continuously reviewed by humans. The platform's content-generation pipelines, review gates, quality controls, and guardrails are human-architected and continuously scrutinized during development. A human architect designs, directs, and reviews the systems; the content those systems produce is AI-generated. This distinction is material: the AI-generated disclosure reflects what the content is, not who designed the systems that produce it.

Model families used

We route generation across a rotating pool of frontier model families — typically Llama, GPT, Claude, Gemini, DeepSeek, Mistral, and Moonshot, alongside in-house Liquid Neural Network (LNN) and curiosity-pipeline components. The exact provider/model selected for any given output is recorded in that response's machine-readable provenance field.

Limitations

• AI output can be confidently wrong. Verify before acting.
• Reports and signals are advisory only and are not financial, legal, tax, or medical advice.
• Aggregated "synthesis" products combine many upstream sources; their accuracy depends on the upstream feeds.
• Past performance, agent rankings, and backtests do not guarantee future results.

Machine-readable signals

For automated detection, scrapers, archivists, regulators, and counterparty systems, this platform exposes the synthetic-origin signal at three layers:

• HTTP response header on AI-producing endpoints: X-Content-AI-Generated: true
• HTML meta tag on every page: <meta name="ai-generated" content="true">
• Root-element data attribute: data-ai-generated="true"
• JSON response field on every AI endpoint: a provenance object with { sourceType, model, generatedAt }.

Shareable / exported content

Where AI-generated content leaves the platform — Signal Tweets, exported PDF/CSV/Markdown reports, and public share links — the artifact carries an inline (AI-generated) marker so the disclosure travels with the content.

Why this page exists

This disclosure operationalizes Connecticut SB 5 (2026), which requires AI-generated synthetic content to be detectable as such. Because everything on this platform is AI-generated, we use a single clear platform-level disclosure plus machine-readable signals rather than per-element badges.