Noctel
NOCTELbeta
← back to archive
archived signalJun 12, 2026 · 39d ago

signal-2026-06-12-001

! Warning3 sourcesFresh 0m$0.002Score 0.62

The core Japan policy claim is plausibly supported by the two Defiant headlines, but the evidence is thin because the cited sources are only headlines and no primary source is provided. The added ETF/tax implications are suggestive rather than firmly established, and one cited source is clearly off-topic, so this is not a cleanly supported signal.

This signal's full body wasn't persisted in the archive snapshot. Run metadata is shown below.

why this matters

things to watch

    provenance

    sources cited
    80
    model
    gpt-5.4-mini
    cost
    $0.1190

    Full source quotes for archived signals are not preserved in the historical index. The agent run log below shows the pipeline trace.

    AGENT RUN LOG

    Pipeline trace · 6 steps · run run-2026-06-12-071

    ! Warningsee all runs →
    Total cost
    $0.119
    Sources
    80
    Steps
    6
    1. 01fetch-tweetsok
      05:13·87.0s·$0.1080
      Input: 40
      Model: n/a
      Output: cache/tweets-2026-06-12.jsonl

      Fetched 720 tweets from 40/40 curated X handles.

    2. 02narrative-trackerok
      05:15·27.4s·$0.0045
      Input: 514
      Model: gpt-4o-mini
      Output: outputs/narratives.json

      Clustered 514 tweets into 5 narratives.

    3. 03digestok
      05:15·46.7s·$0.0019
      Input: 50
      Model: gpt-4o-mini
      Output: outputs/long-form.json

      Ranked 50 candidates from 14/20 feeds, kept top 8.

    4. 04on-chain-monitorok
      05:16·47.6s·$0.0011
      Input: 293
      Model: gpt-4o-mini
      Output: outputs/on-chain.json

      Scanned 15 wallets, 293 transfers, surfaced 7 notable events.

    5. 05morning-briefok
      05:17·4.0s·$0.0019
      Input: 20
      Model: gpt-5.4-mini
      Output: outputs/today.json

      Composed Today's Signal from 5 narratives + 8 long-form + 7 on-chain. It has the cleanest narrative arc and broadest second-order impact, linking a rising AI theme to con

    6. 06signal-verdictwarn
      20:37·3.0s·$0.0017
      Input: 3
      Model: gpt-5.4-mini
      Output: outputs/signal-verdicts.json

      WARNING score 0.62: The core Japan policy claim is plausibly supported by the two Defiant headlines, but the evidence is thin because the ci