Welcome to TruthX

A cleaner, ranked social feed with score-based discovery.

Create account Log in

Public feed

Ranked by score first
verified OSINT claim 65% verification - developing Track 39 establishing Signal 75
Verification details 65% developing
Verification 65%: Multiple sources confirm IDF announced strikes targeting two senior Hamas operatives in northern Gaza on May 27, 2026, closely matching the claim's timing and content. All sources rely on IDF statements without independent verification of operative identities or strike outcomes.

Multiple sources confirm IDF announced strikes targeting two senior Hamas operatives in northern Gaza on May 27, 2026, closely matching the claim's timing and content. All sources rely on IDF statements without independent verification of operative identities or strike outcomes.

Evidence 42 AI 67 Track prior 47 4 resolved OSINT posts 25.0% false-positive/noise rate Source reliability 42 Bias balance 57 0 trusted sources
  • Trusted-source evidence score: 42.0
  • OSINT track record prior: 47.0
  • Checked window: 2026-05-27 13:49 through 2026-05-28 14:49.
  • AI evidence review score: 67.5
The IDF says it struck two key Hamas operatives in northern Gaza.
OSINT quality 78 73-82
Scored for discipline, source tier, information density, confidence signaling, coherence, and redundancy.
Claude 73 ChatGPT 82
  • This is a straightforward relay of an IDF statement about a military strike. The attribution is clear ("IDF says"), which provides minimal epistemic transparency, but as OSINT content it lacks source links, timestamp, or verification pathway. The claim itself is Tier 2: the IDF making this statement is likely documentable, but the underlying factual claim (that the strike occurred and hit the stated targets) remains
  • The statement is narrowly framed as an attributed claim (“IDF says”), which is epistemically safer than asserting the strike as confirmed fact. However, it lacks basic grounding details (when/where, which IDF release, names/roles of operatives), limiting verifiability and OSINT usefulness.
Claude raw JSON 73
{
    "overall": 72,
    "content_type": "reported-fact",
    "rhetorical_intent": "inform",
    "dimensions": {
        "claim_honesty": 75,
        "logical_integrity": 85,
        "epistemic_transparency": 60,
        "communicative_effectiveness": 70,
        "grounding_adequacy": 65
    },
    "flags": [
        {
            "type": "warning",
            "text": "Attribution to IDF present but no independent verification or source link provided for OSINT content"
        },
        {
            "type": "info",
            "text": "Claim tier: Tier 2 (documented but contestable) - IDF statements are public record but represent one party's account in active conflict"
        },
        {
            "type": "caution",
            "text": "No epistemic hedging beyond attribution - reader cannot assess claim strength or verify independently"
        }
    ],
    "reasoning": "This is a straightforward relay of an IDF statement about a military strike. The attribution is clear (\"IDF says\"), which provides minimal epistemic transparency, but as OSINT content it lacks source links, timestamp, or verification pathway. The claim itself is Tier 2: the IDF making this statement is likely documentable, but the underlying factual claim (that the strike occurred and hit the stated targets) remains unverified and represents one combatant's account. For reported-fact content, especially OSINT, grounding adequacy suffers without source access."
}
ChatGPT raw JSON 82
{
    "overall": 78,
    "content_type": "reported-fact",
    "rhetorical_intent": "inform",
    "dimensions": {
        "claim_honesty": 85,
        "logical_integrity": 90,
        "epistemic_transparency": 80,
        "communicative_effectiveness": 70,
        "grounding_adequacy": 60
    },
    "flags": [
        {
            "type": "sourcing_absent",
            "text": "No link, date, or statement excerpt is provided to verify what the IDF said."
        },
        {
            "type": "attribution_limited",
            "text": "The claim is attributed to the IDF, but it does not clarify whether this is confirmed independently or only an official assertion."
        }
    ],
    "reasoning": "The statement is narrowly framed as an attributed claim (“IDF says”), which is epistemically safer than asserting the strike as confirmed fact. However, it lacks basic grounding details (when/where, which IDF release, names/roles of operatives), limiting verifiability and OSINT usefulness."
}
What we measure
0
$
Scoring calls: $0.013709
AI Score In Out Total ms $
Claude 73 1166 372 1538 11915 0.009078
ChatGPT 82 1070 197 1267 3601 0.004631
~ estimated where historical API usage was unavailable
Verification calls: $0.023161
AI Provider In Out Total $
claude-sonnet-4-5 anthropic 1170 688 1858 0.013830
gpt-5.2 openai 1188 518 1706 0.009331
verified OSINT claim 64% verification - developing Track 39 establishing Signal 75
Verification details 64% developing
Verification 64%: IDF confirmed striking two senior Hamas operatives in northern Gaza on May 27, 2026, matching the claim's timeframe. Times of Israel and i24NEWS corroborate the core event within hours of the post. Specific identities as Northern Gaza Brigade commander and Gaza City deputy commander lack independent confirmation in provided sources.

IDF confirmed striking two senior Hamas operatives in northern Gaza on May 27, 2026, matching the claim's timeframe. Times of Israel and i24NEWS corroborate the core event within hours of the post. Specific identities as Northern Gaza Brigade commander and Gaza City deputy commander lack independent confirmation in provided sources.

Evidence 40 AI 65 Track prior 38 4 resolved OSINT posts 25.0% false-positive/noise rate Source reliability 42 Bias balance 50 0 trusted sources
  • Trusted-source evidence score: 40.0
  • OSINT track record prior: 38.0
  • Checked window: 2026-05-27 13:51 through 2026-05-28 14:51.
  • AI evidence review score: 64.7
IDF struck two senior Hamas figures in northern Gaza. Gaza sources identify them as the Northern Gaza Brigade commander and Gaza City deputy commander.
OSINT quality 71 70-72
Scored for discipline, source tier, information density, confidence signaling, coherence, and redundancy.
ChatGPT 70 Claude 72
  • The post makes concrete, news-like claims about a strike and the identities/roles of those targeted, but provides no sourcing and uses vague attribution (“Gaza sources”). It is internally coherent and clearly intended to inform, yet it lacks transparency about confidence and verification for a contested, high-stakes identification.
  • This OSINT post reports a military strike and identifies specific Hamas commanders but provides no verifiable sourcing beyond 'Gaza sources.' While the factual tone suggests Tier 2 claims, the lack of attribution transparency undermines epistemic honesty. For OSINT content, source identification is critical to grounding adequacy. The post is clear and concise but fails to signal uncertainty or provide readers means t
ChatGPT raw JSON 70
{
    "overall": 62,
    "content_type": "reported-fact",
    "rhetorical_intent": "inform",
    "dimensions": {
        "claim_honesty": 70,
        "logical_integrity": 85,
        "epistemic_transparency": 55,
        "communicative_effectiveness": 75,
        "grounding_adequacy": 30
    },
    "flags": [
        {
            "type": "unsourced_claim",
            "text": "No sources or links provided for the IDF strike or the identification of the individuals."
        },
        {
            "type": "attribution_vagueness",
            "text": "\"Gaza sources\" is vague and does not specify which outlets/officials or how the identification was made."
        },
        {
            "type": "verification_gap",
            "text": "Identities/roles of \"two senior Hamas figures\" are contestable and require stronger corroboration than provided."
        }
    ],
    "reasoning": "The post makes concrete, news-like claims about a strike and the identities/roles of those targeted, but provides no sourcing and uses vague attribution (“Gaza sources”). It is internally coherent and clearly intended to inform, yet it lacks transparency about confidence and verification for a contested, high-stakes identification."
}
Claude raw JSON 72
{
    "overall": 72,
    "content_type": "reported-fact",
    "rhetorical_intent": "inform",
    "dimensions": {
        "claim_honesty": 70,
        "logical_integrity": 85,
        "epistemic_transparency": 60,
        "communicative_effectiveness": 75,
        "grounding_adequacy": 65
    },
    "flags": [
        {
            "type": "warning",
            "text": "Attribution to 'Gaza sources' is vague and unverifiable without specific sourcing details"
        },
        {
            "type": "info",
            "text": "IDF strike claim presented as fact without independent confirmation or sourcing"
        },
        {
            "type": "caution",
            "text": "Identities and roles of targeted individuals stated definitively despite relying on unnamed sources"
        }
    ],
    "reasoning": "This OSINT post reports a military strike and identifies specific Hamas commanders but provides no verifiable sourcing beyond 'Gaza sources.' While the factual tone suggests Tier 2 claims, the lack of attribution transparency undermines epistemic honesty. For OSINT content, source identification is critical to grounding adequacy. The post is clear and concise but fails to signal uncertainty or provide readers means to verify the intelligence independently."
}
What we measure
0
$
Scoring calls: $0.013438
AI Score In Out Total ms $
ChatGPT 70 1082 225 1307 4121 0.005044
Claude 72 1178 324 1502 7458 0.008394
~ estimated where historical API usage was unavailable
Verification calls: $0.029203
AI Provider In Out Total $
claude-sonnet-4-5 anthropic 1409 910 2319 0.017877
gpt-5.2 openai 1408 633 2041 0.011326
unverified OSINT claim 0% verification - uncertain Track 39 establishing Signal 75
Verification details 0% uncertain
Verification 0%: No corroborating evidence provided for Khan performing Hajj around May 27, 2026. The claim includes a pointed observation about Mecca's non-Muslim entry restrictions. Without travel reports, official statements, or credible media coverage in the check window, the core claim remains unverified.

No corroborating evidence provided for Khan performing Hajj around May 27, 2026. The claim includes a pointed observation about Mecca's non-Muslim entry restrictions. Without travel reports, official statements, or credible media coverage in the check window, the core claim remains unverified.

Evidence 0 AI 0 Track prior 50 4 resolved OSINT posts 25.0% false-positive/noise rate Source reliability 0 Bias balance 38 0 trusted sources
  • Trusted-source evidence score: 0.0
  • OSINT track record prior: 50.0
  • Checked window: 2026-05-27 13:50 through 2026-05-28 14:50.
  • AI evidence review score: 0.0
London Mayor Sadiq Khan, champion of London’s inclusivity, in Mecca performing Hajj.

Non-Muslims are forbidden from entering.
OSINT quality 65 59-70
Scored for discipline, source tier, information density, confidence signaling, coherence, and redundancy.
ChatGPT 59 Claude 70
  • Promotional/branded ad-like content is downranked by policy.
  • The statement that non-Muslims are forbidden from entering Mecca is broadly accurate, but the key factual claim—Khan being in Mecca performing Hajj—is not substantiated by the provided image and may rely on an unverified/possibly manipulated photo. The critique is clear and rhetorically effective, but the implied conclusion about inclusivity vs. religious site restrictions is more insinuation than argued inference.
  • Image shows the Kaaba and surrounding mosque area consistent with Mecca during Hajj/Umrah, which is relevant to the post’s claim about being in Mecca performing Hajj.
  • The person in the foreground could be anyone; the image alone does not reliably verify the identity as Sadiq Khan.
ChatGPT raw JSON 59
{
    "overall": 55,
    "content_type": "polemic",
    "rhetorical_intent": "critique",
    "dimensions": {
        "claim_honesty": 58,
        "logical_integrity": 55,
        "epistemic_transparency": 60,
        "communicative_effectiveness": 62,
        "grounding_adequacy": 70
    },
    "flags": [
        {
            "type": "identity_unverified",
            "text": "The image does not reliably verify the person is Sadiq Khan; it could be anyone and may be edited/composited."
        },
        {
            "type": "implied_hypocrisy_framing",
            "text": "Juxtaposes 'champion of inclusivity' with Mecca access rules to imply hypocrisy without explicitly arguing the connection."
        }
    ],
    "reasoning": "The statement that non-Muslims are forbidden from entering Mecca is broadly accurate, but the key factual claim—Khan being in Mecca performing Hajj—is not substantiated by the provided image and may rely on an unverified/possibly manipulated photo. The critique is clear and rhetorically effective, but the implied conclusion about inclusivity vs. religious site restrictions is more insinuation than argued inference."
}
Claude raw JSON 70
{
    "overall": 72,
    "content_type": "polemic",
    "rhetorical_intent": "challenge",
    "dimensions": {
        "claim_honesty": 65,
        "logical_integrity": 70,
        "epistemic_transparency": 60,
        "communicative_effectiveness": 90,
        "grounding_adequacy": 75
    },
    "flags": [
        {
            "type": "warning",
            "text": "Image assessment indicates possible manipulation/compositing with low authenticity score (0.42), undermining visual evidence for Khan's presence in Mecca"
        },
        {
            "type": "caution",
            "text": "Implicit hypocrisy claim (championing inclusivity while in exclusionary space) presented without explicit argument or acknowledgment of interpretive leap"
        },
        {
            "type": "info",
            "text": "Non-Muslim prohibition in Mecca is Tier 1 settled fact; no penalty for stating without evidence"
        }
    ],
    "reasoning": "The post makes a pointed juxtaposition between Khan's inclusivity advocacy and Mecca's access restrictions. The factual claim about non-Muslim prohibition is Tier 1 and accurate. However, the image evidence is questionable per the assessment's compositing concerns, and Khan's identity/presence is not reliably verified visually. The implied hypocrisy argument is never explicitly stated, reducing epistemic transparency. As a polemic, the punchy framing is appropriate and communicatively effective, but the reliance on potentially manipulated visual evidence and unstated logical connection between the two facts weakens claim honesty and grounding."
}
What we measure
0
$
Scoring calls: $0.016014
AI Score In Out Total ms $
ChatGPT 59 1353 229 1582 3635 0.005574
Claude 70 1485 399 1884 8759 0.010440
~ estimated where historical API usage was unavailable
Verification calls: $0.013716
AI Provider In Out Total $
claude-sonnet-4-5 anthropic 896 384 1280 0.008448
gpt-5.2 openai 938 259 1197 0.005268
verified OSINT claim 25% verification - uncertain Track 39 establishing Signal 75
Verification details 25% uncertain
Verification 25%: The claim asserts cutting deals with Iran is unwise but provides no specifics about which deal, when, or by whom. The single source discusses Iran policy broadly without confirming any actual deal negotiation or agreement during the check window. No corroborating evidence supports a specific deal event.

The claim asserts cutting deals with Iran is unwise but provides no specifics about which deal, when, or by whom. The single source discusses Iran policy broadly without confirming any actual deal negotiation or agreement during the check window. No corroborating evidence supports a specific deal event.

Evidence 20 AI 0 Track prior 58 4 resolved OSINT posts 25.0% false-positive/noise rate Source reliability 42 Bias balance 35 0 trusted sources
  • Trusted-source evidence score: 20.0
  • OSINT track record prior: 58.0
  • Checked window: 2026-05-27 13:48 through 2026-05-28 14:48.
  • AI evidence review score: 0.0
Not the best idea to be cutting deals with Iran.
OSINT quality 53 41-65
Scored for discipline, source tier, information density, confidence signaling, coherence, and redundancy.
Claude 41 ChatGPT 65
  • Promotional/branded ad-like content is downranked by policy.
  • The comment asserts as fact that deals are being cut with Iran, but provides no evidence, sourcing, or epistemic hedging. The attached image shows a formal portrait with no visible connection to negotiations or agreements. The normative critique is valid opinion, but it's built on an ungrounded factual claim presented with false certainty. For a standalone post flagged as OSINT, this lacks the discipline and transpar
  • Image appears to be a formal portrait of a Middle Eastern leader/official seated on an ornate throne-like chair, suggesting a real-world political context.
  • The post text references Iran and political deal-making; the image loosely supports a geopolitical theme but does not specifically show Iran or any deal/meeting.
Claude raw JSON 41
{
    "overall": 42,
    "content_type": "polemic",
    "rhetorical_intent": "critique",
    "dimensions": {
        "claim_honesty": 35,
        "logical_integrity": 40,
        "epistemic_transparency": 30,
        "communicative_effectiveness": 65,
        "grounding_adequacy": 50
    },
    "flags": [
        {
            "type": "critical",
            "text": "Core claim ('cutting deals with Iran') presented as settled fact (Tier 1) without evidence that any deal exists or is being negotiated."
        },
        {
            "type": "critical",
            "text": "Zero epistemic transparency: no signal that this is speculation, interpretation, or opinion about a potential future action."
        },
        {
            "type": "warning",
            "text": "Image shows a Middle Eastern official but provides no specific evidence of Iran or any deal-making activity; visual context does not substantiate the claim."
        },
        {
            "type": "info",
            "text": "Normative judgment ('not the best idea') is legitimate opinion, but rests on an unestablished factual premise."
        }
    ],
    "reasoning": "The comment asserts as fact that deals are being cut with Iran, but provides no evidence, sourcing, or epistemic hedging. The attached image shows a formal portrait with no visible connection to negotiations or agreements. The normative critique is valid opinion, but it's built on an ungrounded factual claim presented with false certainty. For a standalone post flagged as OSINT, this lacks the discipline and transparency expected."
}
ChatGPT raw JSON 65
{
    "overall": 62,
    "content_type": "polemic",
    "rhetorical_intent": "critique",
    "dimensions": {
        "claim_honesty": 70,
        "logical_integrity": 55,
        "epistemic_transparency": 75,
        "communicative_effectiveness": 60,
        "grounding_adequacy": 50
    },
    "flags": [
        {
            "type": "vagueness",
            "text": "The comment warns against 'cutting deals with Iran' but gives no specifics about which deal, actors, timeframe, or why it is a bad idea."
        },
        {
            "type": "unsupported_assertion",
            "text": "Implies a negative judgment about deal-making with Iran without providing evidence or reasoning."
        },
        {
            "type": "image_context_gap",
            "text": "Attached image appears to be a formal portrait of a Middle Eastern official, but it does not substantiate any specific claim about Iran or negotiations."
        }
    ],
    "reasoning": "This is a broad, opinionated warning rather than a verifiable factual claim, and it does not misstate concrete facts. However, it offers no argument or context to support the critique, and the image only loosely signals geopolitics without grounding the implied claim about a specific deal."
}
What we measure
0
$
Scoring calls: $0.015721
AI Score In Out Total ms $
Claude 41 1414 396 1810 9598 0.010182
ChatGPT 65 1293 234 1527 3729 0.005539
~ estimated where historical API usage was unavailable
Verification calls: $0.016260
AI Provider In Out Total $
claude-sonnet-4-5 anthropic 993 439 1432 0.009564
gpt-5.2 openai 1026 350 1376 0.006696
Elon Musk has created the most succesful electric car company, autonomous driving, robots, rocket ships, nueralink and he's only in his low 50s. If you're still living in your parents' basement and have no history of producing value you probably shouldn't be calling him dumb.
Quality 68 64-72
Ranked by average quality score first, then freshness.
ChatGPT 64 Claude 72
  • The comment defends Musk by listing achievements, but several are Tier 2/3 contestable (credit attribution, "most successful," and implying he personally created multiple technologies). It also relies on a personal insult to delegitimize critics, weakening the logical link between Musk’s accomplishments and whether calling him dumb is justified.
  • This polemic defends Musk by citing his accomplishments (Tier 1/2 claims, largely accurate) but commits a logical error: it suggests critics lacking comparable achievements shouldn't call him 'dumb.' This conflates domain expertise with general correctness - entrepreneurial success doesn't validate all statements. The ad hominem rule doesn't apply here because the argument dismisses critique based on the critic's lif
ChatGPT raw JSON 64
{
    "overall": 63,
    "content_type": "polemic",
    "rhetorical_intent": "defend",
    "dimensions": {
        "claim_honesty": 62,
        "logical_integrity": 55,
        "epistemic_transparency": 70,
        "communicative_effectiveness": 78,
        "grounding_adequacy": 60
    },
    "flags": [
        {
            "type": "overstatement",
            "text": "Claims Musk \"has created\" autonomous driving, robots, rocket ships, Neuralink, and \"the most successful electric car company\" are contestable and oversimplify corporate/collective achievements."
        },
        {
            "type": "ad_hominem",
            "text": "Dismisses critics as \"living in your parents' basement\" and \"no history of producing value\" rather than addressing their specific claims about Musk."
        }
    ],
    "reasoning": "The comment defends Musk by listing achievements, but several are Tier 2/3 contestable (credit attribution, \"most successful,\" and implying he personally created multiple technologies). It also relies on a personal insult to delegitimize critics, weakening the logical link between Musk’s accomplishments and whether calling him dumb is justified."
}
Claude raw JSON 72
{
    "overall": 72,
    "content_type": "polemic",
    "rhetorical_intent": "defend",
    "dimensions": {
        "claim_honesty": 75,
        "logical_integrity": 60,
        "epistemic_transparency": 70,
        "communicative_effectiveness": 85,
        "grounding_adequacy": 80
    },
    "flags": [
        {
            "type": "logical_flaw",
            "text": "Credibility challenge conflates accomplishment with correctness on specific claims - success in one domain doesn't immunize statements in another from critique"
        },
        {
            "type": "info",
            "text": "Tier 1 claims about Musk's ventures are broadly accurate (Tesla, SpaceX, Neuralink exist and have achieved notable success)"
        }
    ],
    "reasoning": "This polemic defends Musk by citing his accomplishments (Tier 1/2 claims, largely accurate) but commits a logical error: it suggests critics lacking comparable achievements shouldn't call him 'dumb.' This conflates domain expertise with general correctness - entrepreneurial success doesn't validate all statements. The ad hominem rule doesn't apply here because the argument dismisses critique based on the critic's life circumstances rather than addressing any specific claim. Effective as a rhetorical defense but logically incomplete."
}
What we measure
0
$
Scoring calls: $0.013728
AI Score In Out Total ms $
ChatGPT 64 1112 224 1336 4144 0.005082
Claude 72 1217 333 1550 8536 0.008646
~ estimated where historical API usage was unavailable
You know what's odd and completely illogical, when people who are jobless, no spouse or children, no history of productive living post online that they think Elon Musk is dumb or stupid. Criticizing Elon about something specific is ok. Claiming he's dumb is absurd.
Quality 63 63-63
Ranked by average quality score first, then freshness.
Claude 63 ChatGPT 63
  • This polemic attempts to defend Musk's intelligence by attacking critics' personal circumstances rather than their arguments. While the final distinction (specific criticism vs blanket dismissal) is valid, the core reasoning commits ad hominem by suggesting critics lack standing based on life circumstances unrelated to their claims' validity. The argument structure replaces substantive rebuttal with credibility attac
  • The comment mainly defends Musk by attacking the presumed personal status of his critics, which weakens the logic and relies on an unsupported generalization. It is clear about its opinion (that calling Musk dumb is absurd) and communicates its stance effectively, but provides little grounding beyond rhetorical assertion.
Claude raw JSON 63
{
    "overall": 62,
    "content_type": "polemic",
    "rhetorical_intent": "challenge",
    "dimensions": {
        "claim_honesty": 70,
        "logical_integrity": 45,
        "epistemic_transparency": 60,
        "communicative_effectiveness": 75,
        "grounding_adequacy": 70
    },
    "flags": [
        {
            "type": "logical_flaw",
            "text": "Ad hominem: dismisses critics' claims based solely on their personal circumstances (jobless, no spouse/children, no productive history) rather than addressing the substance of their arguments."
        },
        {
            "type": "info",
            "text": "The distinction between 'criticizing something specific' vs 'claiming he's dumb' is reasonable but undermined by the preceding personal attack on critics' standing."
        }
    ],
    "reasoning": "This polemic attempts to defend Musk's intelligence by attacking critics' personal circumstances rather than their arguments. While the final distinction (specific criticism vs blanket dismissal) is valid, the core reasoning commits ad hominem by suggesting critics lack standing based on life circumstances unrelated to their claims' validity. The argument structure replaces substantive rebuttal with credibility attacks."
}
ChatGPT raw JSON 63
{
    "overall": 62,
    "content_type": "polemic",
    "rhetorical_intent": "defend",
    "dimensions": {
        "claim_honesty": 60,
        "logical_integrity": 55,
        "epistemic_transparency": 70,
        "communicative_effectiveness": 75,
        "grounding_adequacy": 60
    },
    "flags": [
        {
            "type": "ad_hominem_risk",
            "text": "Dismisses critics by asserting they are jobless/without family/unproductive rather than addressing their claims."
        },
        {
            "type": "unsupported_generalization",
            "text": "Claims that people criticizing Musk as dumb are typically jobless/unproductive without evidence."
        }
    ],
    "reasoning": "The comment mainly defends Musk by attacking the presumed personal status of his critics, which weakens the logic and relies on an unsupported generalization. It is clear about its opinion (that calling Musk dumb is absurd) and communicates its stance effectively, but provides little grounding beyond rhetorical assertion."
}
What we measure
0
$
Scoring calls: $0.012889
AI Score In Out Total ms $
Claude 63 1217 311 1528 10758 0.008316
ChatGPT 63 1109 188 1297 3466 0.004573
~ estimated where historical API usage was unavailable
Tiny product wish: let me sort by thoughtful disagreement. The best replies are often the ones that push back without turning the room into a food fight.
Quality 95 93-96
Ranked by average quality score first, then freshness.
ChatGPT 93 Claude 96
  • The comment makes a subjective product request and clearly frames its key claim (“best replies”) as preference rather than objective fact. The reasoning is coherent: prioritizing thoughtful disagreement is presented as a way to reduce unproductive conflict. No contested factual assertions are made, so grounding demands are minimal.
  • This is a straightforward personal wish about a product feature, expressing a subjective value judgment about reply quality. The claim that 'best replies often push back without turning the room into a food fight' is clearly framed as personal opinion ('tiny product wish', 'let me'). No factual claims requiring verification are made. The metaphor is clear and the communicative intent is achieved efficiently.
ChatGPT raw JSON 93
{
    "overall": 92,
    "content_type": "polemic",
    "rhetorical_intent": "persuade",
    "dimensions": {
        "claim_honesty": 95,
        "logical_integrity": 90,
        "epistemic_transparency": 95,
        "communicative_effectiveness": 90,
        "grounding_adequacy": 100
    },
    "flags": [],
    "reasoning": "The comment makes a subjective product request and clearly frames its key claim (“best replies”) as preference rather than objective fact. The reasoning is coherent: prioritizing thoughtful disagreement is presented as a way to reduce unproductive conflict. No contested factual assertions are made, so grounding demands are minimal."
}
Claude raw JSON 96
{
    "overall": 92,
    "content_type": "personal-narrative",
    "rhetorical_intent": "celebrate",
    "dimensions": {
        "claim_honesty": 100,
        "logical_integrity": 95,
        "epistemic_transparency": 95,
        "communicative_effectiveness": 85,
        "grounding_adequacy": 100
    },
    "flags": [
        {
            "type": "info",
            "text": "Personal product feature request expressing subjective preference for civil disagreement in replies"
        }
    ],
    "reasoning": "This is a straightforward personal wish about a product feature, expressing a subjective value judgment about reply quality. The claim that 'best replies often push back without turning the room into a food fight' is clearly framed as personal opinion ('tiny product wish', 'let me'). No factual claims requiring verification are made. The metaphor is clear and the communicative intent is achieved efficiently."
}
What we measure
0
$
Scoring calls: $0.010850
AI Score In Out Total ms $
ChatGPT 93 1084 131 1215 2476 0.003731
Claude 96 1183 238 1421 5189 0.007119
~ estimated where historical API usage was unavailable
One test I keep coming back to: would I still share this if the headline supported the other side? If not, I probably need to slow down.
Quality 95 93-96
Ranked by average quality score first, then freshness.
ChatGPT 93 Claude 96
  • The post offers a personal heuristic for avoiding partisan sharing and does not present contested factual claims as settled. The conditional reasoning is coherent and clearly framed as a self-check rather than an objective rule. No sourcing is needed because it’s advice, not a claim about external events.
  • This is a reflective heuristic for evaluating one's own epistemic hygiene when sharing content. It makes no factual claims requiring verification (Tier 1 auto-score). The logical structure is sound: partisan bias can distort judgment, so testing reversibility helps detect motivated reasoning. Transparency is high—it's clearly framed as personal practice ('I keep coming back to') rather than universal law. Slightly ab
ChatGPT raw JSON 93
{
    "overall": 92,
    "content_type": "educational",
    "rhetorical_intent": "inform",
    "dimensions": {
        "claim_honesty": 95,
        "logical_integrity": 90,
        "epistemic_transparency": 95,
        "communicative_effectiveness": 90,
        "grounding_adequacy": 100
    },
    "flags": [],
    "reasoning": "The post offers a personal heuristic for avoiding partisan sharing and does not present contested factual claims as settled. The conditional reasoning is coherent and clearly framed as a self-check rather than an objective rule. No sourcing is needed because it’s advice, not a claim about external events."
}
Claude raw JSON 96
{
    "overall": 92,
    "content_type": "educational",
    "rhetorical_intent": "inform",
    "dimensions": {
        "claim_honesty": 100,
        "logical_integrity": 95,
        "epistemic_transparency": 95,
        "communicative_effectiveness": 85,
        "grounding_adequacy": 100
    },
    "flags": [
        {
            "type": "info",
            "text": "Prescriptive meta-commentary on information consumption rather than substantive claim"
        }
    ],
    "reasoning": "This is a reflective heuristic for evaluating one's own epistemic hygiene when sharing content. It makes no factual claims requiring verification (Tier 1 auto-score). The logical structure is sound: partisan bias can distort judgment, so testing reversibility helps detect motivated reasoning. Transparency is high—it's clearly framed as personal practice ('I keep coming back to') rather than universal law. Slightly abstract for maximum communicative punch, but effective as a memorable decision rule."
}
What we measure
0
$
Scoring calls: $0.011091
AI Score In Out Total ms $
ChatGPT 93 1085 126 1211 3143 0.003663
Claude 96 1181 259 1440 7242 0.007428
~ estimated where historical API usage was unavailable
I do not need every post to be breaking news. Sometimes the useful post is just someone saying, 'here is what changed and here is the source.'
Quality 70 61-78
Ranked by average quality score first, then freshness.
ChatGPT 61 Claude 78
  • Clear preference for informative updates over sensational “breaking news.”
  • No concrete examples or sources are provided, but the claim is framed as personal preference rather than a factual assertion.
  • Reasoning is coherent and encourages a norm of citing sources for changes.
  • Constructive, non-hostile tone that can support healthier discussion norms.
What we measure
0
$
Scoring calls: $0.018764
AI Score In Out Total ms $
ChatGPT 61 1119 366 1485 8080 0.007082
Claude 78 1254 528 1782 7012 0.011682
~ estimated where historical API usage was unavailable
Small local thing I wish more cities posted clearly: which streets are closed for weekend events before you're already stuck in the detour.
Quality 70 64-76
Ranked by average quality score first, then freshness.
ChatGPT 64 Claude 76
  • Clear, specific request about better communication of weekend street closures.
  • No evidence is needed for the personal preference, but it also provides no concrete examples or suggestions for implementation.
  • Reasoning is straightforward (advance notice prevents getting stuck in detours).
  • Tone is polite and constructive.
What we measure
0
$
Scoring calls: $0.018635
AI Score In Out Total ms $
ChatGPT 64 1113 350 1463 5765 0.006848
Claude 76 1249 536 1785 8113 0.011787
~ estimated where historical API usage was unavailable
The best version of a truth score probably explains what it knows and what it doesn't. A plain number without the reasoning would just create new arguments.
Quality 68 67-69
Ranked by average quality score first, then freshness.
ChatGPT 67 Claude 69
  • Makes a clear normative point about how a truth score should be presented (with explanations and uncertainty).
  • Gives a plausible rationale that a bare number could provoke disputes, though it’s not backed by examples or evidence.
  • Reasoning is coherent and appropriately framed as a preference/probability rather than a hard factual claim.
  • Constructive, non-hostile tone that invites better design and discussion.
What we measure
1
$
Scoring calls: $0.021900
AI Score In Out Total ms $
ChatGPT 67 1116 459 1575 7313 0.008379
Claude 69 1252 651 1903 9432 0.013521
~ estimated where historical API usage was unavailable
Trying this place out: I like the idea of a feed where the strongest comments float up, but I hope the score still leaves room for jokes and half-formed thoughts.
Quality 66 60-72
Ranked by average quality score first, then freshness.
ChatGPT 60 Claude 72
  • Clearly states a preference for ranking strong comments while expressing a concern about preserving humor and rough drafts.
  • Makes no empirical claims and appropriately frames the point as personal hope/opinion.
  • Adds some discussion value by implicitly raising a product/design question about what the scoring system should reward.
  • Lacks concrete examples or suggestions for how to balance quality ranking with jokes/early thoughts.
What we measure
1
$
Scoring calls: $0.019151
AI Score In Out Total ms $
ChatGPT 60 1122 417 1539 5733 0.007802
Claude 72 1258 505 1763 7750 0.011349
~ estimated where historical API usage was unavailable
Version 1.2 - 2026-05-20 17:30 ET