Research Note

E098 Lab Artifact Research Note

E098 is supported by the canonical transcript and two recovered screen-recording files in the original Dropbox production folder. The filenames are `screenshare_hd-2026-1

Aug 4, 20263 min readBy Dalton Anderson
In this article

E098 Lab Artifact Research Note

E098 is supported by the canonical transcript and two recovered screen-recording files in the original Dropbox production folder. The filenames are screenshare_hd-2026-1-5__21-52-57-CFR.mp4 and screenshare_hd-2026-1-5__22-9-41-CFR.mp4. Their names establish January 5, 2026 as the capture date. The files run for approximately 20 minutes and 42 minutes.

The original 122-page requirements document, account receipts, exact plan names, model selections, and machine-readable run logs were not found in the episode folder or the immediate raw-assets folder. The comparison therefore cannot be reproduced from the preserved package alone.

What the recordings establish

The Base44 recording shows an editor and a generated application named Rrisco. The preview contains a risk-management dashboard, recent submissions, portfolio health, properties, claims, analytics, and a multi-step new-submission flow. The editor also records a correction to component import paths. This supports the bounded claim that Base44 generated a coherent, navigable, multi-screen demo in this run. It does not establish correct insurance logic, secure permissions, persistent data, or production readiness.

The Emergent recording shows a Risk OS landing experience and a limited preview. It also shows unsupported generated claims such as 95 percent time saved, 100 percent data accuracy, and sub-three-second enrichment. Those values are demo copy, not measured results. The recording supports Dalton's criticism that the output looked more like a product claim page than a complete working application.

The Lovable recording shows an editor, a generated interactive quote flow, carrier options, and a bind action. It supports the claim that Lovable produced a broader interactive demo than a static landing page in this run. It does not prove carrier connectivity or valid quote logic.

The Replit recording shows a temporary development preview with a submission workflow and generated classification results. The transcript separately records authentication, 404, and analytics failures before and during the usable preview. The screenshot proves that a working screen eventually appeared, not that the full application ran reliably.

The Firebase Studio recording shows the App Prototyping agent, generated Next.js interface work, a Gemini API-key requirement, a visible security-check recommendation, and several generated dashboard variants. It supports the claim that the interface appeared but required additional setup and contained unstable navigation. The recording should be attributed to Firebase Studio's App Prototyping agent, not to Genkit as a competing hosted builder.

Preserved stills

Five representative JPG stills were extracted into Assets. They remain first-hand episode media. Captions must state the January 5, 2026 capture date, the relevant platform, and the specific interface fact visible in the frame. A screenshot is not evidence that an integration, metric, classification, or transaction is real.

The raw videos remain outside the Obsidian vault. They should not be uploaded publicly without a separate media review for account details, private information, and distribution rights.

Publication boundary

The recovered recordings satisfy the evidence gate for a dated lab report, provided the page discloses the missing requirements file, plan details, costs, models, and reruns. They do not satisfy the stronger gate for a reproducible benchmark or a current vendor ranking.

Sources

Follow the evidence.

  1. docs.base44.com: githubdocs.base44.com
  2. web.dev: vitalsweb.dev
  3. docs.replit.com: replit appsdocs.replit.com
  4. docs.replit.com: build with agentdocs.replit.com
  5. csrc.nist.gov: finalcsrc.nist.gov
  6. help.emergent.sh: 272715 features and toolshelp.emergent.sh
  7. firebase.google.com: migrating projectfirebase.google.com
  8. owasp.org: www project application security verification standardowasp.org
  9. docs.base44.com: Quick start guidedocs.base44.com
  10. w3.org: WCAG22w3.org
  11. firebase.google.com: get started aifirebase.google.com
  12. help.emergent.sh: plans and creditshelp.emergent.sh
  13. docs.lovable.dev: githubdocs.lovable.dev
  14. docs.lovable.dev: getting starteddocs.lovable.dev
  15. firebase.google.com: overviewfirebase.google.com

From this episode

Two useful next steps.

Evergreen · 1 min

What One-Shot App Generation Actually Proves

A one-shot AI app build can prove initial instruction-following and visible interaction. It cannot prove security, correctness, scale, or demand.

Research Note · 1 min

Vibe Coding Benchmark Method Research Note

A useful AI app-builder benchmark must answer a decision rather than manufacture a universal leaderboard. The decision might be which tool best supports a team's internal

Return to the episode
E098 Lab Artifact Research Note