JARVISLabs

Research that shows where every sentence came from

Asked how the James Webb Space Telescope keeps its instruments cold, JARVIS read 3 of 3 sources from 3 independent origins in 19.9 s, captured each page as read, and tied every finding to its source. The spoken summary was weak.

LIVE-PROVEN What this run supports. Recorded 22 September 2026.

Asked how Webb keeps its instruments cold, JARVIS read 3 of 3 sources from 3 independent origins in 19.9 s and marked every finding with the source it rests on. ShowsThe research desk retrieves real pages, keeps a capture of each, and ties every finding to a source. Does not showThe spoken summary was thin: it mostly repeated a source title. Source count is not a calibrated confidence.
What happens, step by step
  1. 0:00 Asked: “How does the James Webb Space Telescope keep its instruments cold?”
  2. 0:04 3 of 3 sources answered · the research panel opens
  3. 0:09 Each finding names its source · pages captured as read

Captured at revision 893913a9c

The maximised research panel: numbered findings about how Webb stays cold, three NASA and ESA sources with retrieval times, and page captures as read.
Every finding is tagged with the source it rests on; each source keeps a capture of the page as it was read.

Captured at revision 893913a9c

Asked

What was asked

Research how the James Webb Space Telescope keeps its instruments cold, and give me a short cited summary.

Result

What happened

JARVIS called research.run with a sharpened question (authoritative sources, preferably NASA and ESA). The research desk fetched three pages (NASA Science, NASA and ESA), saved the synthesis as revision 1, and the HUD opened the research panel with numbered findings, per-source retrieval times and captures of each page as read. The spoken answer was: "To capture infrared light from cosmic objects, the James Webb Space Telescope will have to be kept as cold as possible. How Does NASA's Webb Telescope Stay Cool in Chamber A? That's from 3 sources, on the research panel."

Pipeline

Architecture path

  1. Typed request
  2. Hosted cognition
  3. research.run
  4. Three sources fetched and captured
  5. Findings tied to sources
  6. Research panel
  7. Spoken answer

What this proves

  • The research desk retrieves real pages on request and keeps a capture of each page as it was read.
  • Each finding on the panel is tagged with the source it rests on, and each source shows how many findings rest on it.
  • The panel states its method plainly: a count of sources and distinct hosts, not a calibrated confidence.

What it does not prove

  • The spoken summary was thin: it read out a finding and then a source title. The panel is better than the voice here.
  • Three agreeing sources are not independent verification of the physics.
  • Recorded on 2026-09-22; the sources may have changed since.

The run

The research desk ran from 22:49:07 to 22:49:15 UTC. The whole turn, including the spoken answer, took 19.9 s. The maximised panel shows the question, “saved · revision 1”, seven numbered findings and the three sources. Each source is listed with its host, the characters read and the retrieval time, and has a thumbnail of the page as read.

What to take from it

The panel is the product here. The answer can be audited sentence by sentence, and the page each sentence came from is kept. The voice answer is not yet as good as the panel: it quoted a finding and then read out a source title instead of summarising. That gap is visible in the footage, and it stays visible here.