Quarantined evidence

Why each removed evidence row failed the grounding audit. Ambiguity excludes, it never includes. Judge the audit itself.

About this universe

110tracked
84with evidence
98discovered

rule72 tracks a hand-curated universe of 110 companies — the companies it understands today, built from attributed creator evidence. This is not market-wide coverage. 84 of them carry grounded creator evidence in the current dataset.

Beyond it, 98 more companies surface in creator transcripts but aren’t tracked yet — what rule72 may understand next. Discovery is not coverage: a candidate becomes tracked only after an operator-gated review. See discovered candidates →

Dataset & review context

dataset window: 2026-06-25
grounded evidence: 912
published: 0
with timestamp: 248/912
tracked companies: 110
companies with evidence: 84
companies with no grounded evidence: 26
creators with evidence: 59
registered channels with no grounded evidence: 39

Read-only. Nothing here publishes or mutates data — 0 published: nothing auto-publishes; every row is review-pending until a human review step sets it.

Audit summary

candidate evidence audited: 403
kept (grounded): 241
quarantined: 162 (40.2%)
— not grounded: 46
— bad attribution: 52
— stance mismatch: 100
— misresolved: 31
— invented number: 16
audit model: sonnet
method: full curated corpus audit

A row can fail more than one check, so the breakdown can sum past the quarantined total. This audit run covered 403 candidate rows (241 kept); the current dataset now has 912 grounded rows — the audit is a point-in-time run and may not cover the full current dataset.

Verdicts

showing 162 of 162 quarantined rows

ticker	stance	channel	failed checks	reason
SPY	negative	tastylive	stance mismatch	The quote ("risk reward was pretty decent in spy... started to nibble on this") does not itself convey a negative/bearish stance — the bearishness is only inferrable from surrounding context about selling call spreads; the extracted quote alone is directionally ambiguous and could describe a long entry, so the "negative" stance label is not grounded in the quote as written.
QQQ	negative	tastylive	not grounded, stance mismatch	The quote ("risk reward was pretty decent in spy and in the Qs… started to nibble on this") does not itself state a bearish/negative stance on QQQ — the direction (short via call spreads) is inferred from a prior speaker's remarks in the dialogue, and the quote's own language ("pretty decent risk reward," "nibble") reads as a cautiously constructive entry assessment; without the prior-speaker context the quote cannot independently support a negative stance label, so both grounded and stance_matches fail.
MU	mixed	Real Vision	bad attribution, stance mismatch	The quote's directional sentiment on Micron's earnings outcome is explicitly attributed to "a lot of the experts," not to the creator Chris — Chris himself states "I don't know what's going to happen, it's anybody's guess," making the recorded "mixed" stance a misattribution of hearsay expert opinion rather than the creator's own expressed view.
IWM	positive	Real Vision	misresolved	The creator references the Russell 2000 index as a macro indicator confirming the correction is contained — it is cited as an aside/data point in a macro analysis, not as the subject of expressed sentiment — and the creator never mentions IWM or any ETF; resolving a passing index reference used as a macro signal to a specific ETF ticker is a resolution error.
IWM	negative	tastylive	stance mismatch	The creator's IWM-specific commentary is observational and mixed — small caps are described as "still green," the shooting star is qualified as "not a particularly good one, but it's all right" — so the IWM-specific stance is neutral-to-mixed, not a clean negative; the negative label is inferred from the creator's general bear posture rather than expressed directly about IWM.
QQQ	negative	tastylive	stance mismatch	The stance label "negative" does not match the creator's actual expression: they call the QQQ chart "a mess" (mildly negative) but immediately follow with "so far, so good" (positive relief that it isn't rallying), making the expressed stance mixed rather than cleanly negative.
NVDA	mixed	42 Macro	not grounded	NVIDIA is used as a category exemplar to frame a broader chip-sector observation ahead of Micron's earnings — the creator is not expressing a targeted stance on NVIDIA itself; the primary subjects are Micron and the semiconductor sector, so grounded=false for NVIDIA as the intended subject.
IWM	positive	42 Macro	not grounded, misresolved, stance mismatch	The quote expresses a thematic view on AI adopters outperforming AI providers — a sector rotation thesis — and does not directly reference IWM or the Russell 2000; the transcript window's brief mention of "Russell outperformance" at the open is not sufficient to ground the quote to IWM specifically, making the ticker resolution and the "positive" stance attribution to IWM unsupported.
SPY	negative	Wealthion	bad attribution	The "disaster" view on the market is expressed by a named guest (addressed as "Mark" by the Wealthion host) who is being interviewed, not by the Wealthion channel creator; attribution_is_creator fails because this is guest hearsay, not the channel creator's own stance.
QQQ	negative	Wealthion	misresolved, bad attribution	Two defects: (1) attribution_is_creator is false — the bearish views are spoken by a guest (addressed as "Mark" by the Wealthion interviewer), not by the channel creator/host; (2) resolution_correct is false — the speaker never names QQQ or the Nasdaq 100; the discussion refers generically to "the indices," the S&P, and index funds broadly, making the QQQ resolution an unsupported inference rather than a named security.
ORCL	negative	Wealthion	bad attribution	The Oracle statement is made by a guest/interviewee on the Wealthion channel, not by the Wealthion creator/host — the transcript's >> dialogue markers show the host asking questions while the guest (an economist who references being Christian and owning bonds) makes the Oracle remark, so attribution_is_creator fails.
META	positive	tastylive	stance mismatch	The quote is an explicit joke about Zuckerberg's personal image change ("chains," "flowy hair") — not a genuine expression of positive investment sentiment toward Meta stock; labeling it "positive" stance misrepresents a comedic aside as a substantive creator view.
AVGO	positive	Bloomberg Technology	bad attribution	The quote is Ed Ludlow (the Bloomberg Technology anchor) paraphrasing what Broadcom's CEO told Bloomberg News — the bullish sentiment belongs to the Broadcom CEO, not to the channel creator; attribution_is_creator fails because Ed is reporting hearsay from a company executive, not expressing his own view on AVGO.
MU	positive	Bloomberg Technology	stance mismatch	ED's commentary is journalistic/anticipatory — "the market really wants a bullish signal" describes the market's expectations, not the creator's own positive view on Micron — so the positive stance label overstates what is effectively neutral observational framing around an earnings event.
PLTR	positive	a16z	stance mismatch	The creator invokes Palantir solely as an illustrative data point ("the only one growing, whatever 70%") to support a broader argument about the scarcity of high-growth public companies and why upcoming mega-IPOs are good for the market — not to express a positive view on PLTR itself; the stance is neutral/observational, not positive.
SPY	mixed	Monetary Matters	bad attribution	The transcript_window reveals a two-speaker dialogue where the mixed-stance SPY argument ("SPY is a good example but shouldn't be your entire portfolio") is made by the non->> speaker, while the >> speaker (who appears to be the Monetary Matters host, pushing back with "I don't see it as being that different than the SPY") holds a different view; the attributed stance belongs to a guest, not the channel creator.
MU	mixed	tastylive	stance mismatch	The transcript shows the creator discussing a neutral options strategy (iron condor, defined risk) around Micron's earnings — not expressing a mixed directional view on the stock itself. The stance "mixed" misrepresents what is a range-bound/neutral options trade preference, not a mixed bullish/bearish opinion on MU as a security.
QQQ	negative	Trade Brigade	not grounded	The transcript window does not contain the distinctive NASDAQ-specific language from the quote ("less well-rounded than the S&P 500... a little bit of a nastier look") — the window covers sector rotation and general risk-off commentary without the quoted comparative language that specifically identifies QQQ/NASDAQ as the subject, so the quote cannot be grounded in this excerpt.
NVDA	mixed	Trade Brigade	not grounded, stance mismatch	The quote's opening clause — "Nvidia is not immune. It's certainly pushing higher uh on some of this news as well" — does not appear in the transcript_window; the window only shows "You can thank Nvidia. Both really heavy today. Uh, horrible, horrible stuff," which is purely negative, making the quote partially ungrounded and the 'mixed' stance label unsupported by the available evidence.
META	negative	Trade Brigade	stance mismatch	The quote explicitly anticipates a flush lower AND a subsequent counter-trend bounce ("maybe we go counter trend"), making the sentiment mixed rather than purely negative; labeling it "negative" drops the bullish counter-trend component the creator is openly waiting for.
COIN	negative	Trade Brigade	not grounded, misresolved, stance mismatch	The quote references "sailor needs to get liquidated," which refers to Michael Saylor and MicroStrategy (MSTR), not Coinbase (COIN); the transcript window shows no expression of sentiment about Coinbase at all, making both the ticker resolution and the grounding wrong.
QQQ	mixed	Prof G Markets	not grounded, misresolved, stance mismatch	The transcript window shows the creator discussing SpaceX, not QQQ — the Nasdaq 100 index inclusion is cited only as a buying mechanism for SpaceX shares; QQQ is never the subject of any expressed stance, making resolution_correct and grounded both false.
TSLA	mixed	The Information	bad attribution	The views in the quote belong to Theo Wayt, a guest reporter introduced by the host as "our Elon Musk reporter here at The Information" — the attribution guard fails because this is a guest's opinion, not the creator/host's own stance.
TSLA	mixed	The Information	bad attribution	The Tesla sentiment is expressed by Theo Wayt, a journalist guest being interviewed on The Information's channel, not by the channel/show creator itself — the transcript explicitly identifies him as "our Elon Musk reporter" brought on as a subject-matter guest, so attributing this stance to the channel "The Information" as creator is a guest-hearsay attribution failure.
QQQ	negative	Brian Shannon - Alpha Trends	invented number	The quote retains "695-ish," a specific creator-stated downside price level that the numeric scrubber must remove before publication; surfacing it as a target violates the removal-only scrubbing rule.
AMZN	mixed	The Compound	misresolved, bad attribution, stance mismatch	The quote is a group-level derating comment about five hyperscalers (Amazon, Microsoft, Google, Meta, Oracle) simultaneously, not an Amazon-specific stance; additionally the quote stitches together two distinct speakers across a >> speaker-transition boundary, so neither the resolution to AMZN alone nor the attribution to a single creator holds.
NFLX	negative	The Compound	stance mismatch	The creator explicitly states they are LONG Netflix ("I'm longing uh Netflix with an average cost in the 90s") — their actual position is bullish; the "going to zero" remark is sardonic self-deprecation about an underwater trade, confirmed by the co-host's "Don't feel bad" response, so labeling the stance "negative" misrepresents what the creator expressed.
AAPL	negative	SpotGamma	stance mismatch	The creator is NOT expressing a negative directional stance on Apple as a security — they are saying Apple's options profile (persistently positive gamma near price) makes it a poor candidate for their SpotGamma/EquityHub signal methodology; this is a neutral methodological observation about tool applicability, not a bearish view on AAPL, so stance_matches fails.
MSFT	neutral	SpotGamma	not grounded, stance mismatch	The transcript shows the creator dismissing Microsoft because its options strike is too far out of the money for their specific trading strategy — this is a comment on options-structure suitability, not a sentiment stance on MSFT as a security; the window does not support extracting any directional or neutral creator sentiment on Microsoft, and "neutral" is a mislabel of what is actually a strategy-specific rejection.
META	positive	SpotGamma	invented number, stance mismatch	Two defects: (1) the quote retains specific price levels (565, 570, 575, 580, 585) presented as upside targets, which the numeric scrubber should have removed; (2) the stance is labeled "positive" but the creator explicitly states "This gives me an indication to potentially get short" — the described trade setup is bearish/short, not positive.
AMZN	mixed	SpotGamma	stance mismatch	The creator describes a straightforwardly bearish intraday setup on AMZN — watching price reject at 236-237 and calling it "a short entry" — with no bullish or offsetting element present in the transcript; the stance should be "negative," not "mixed."
QQQ	negative	SpotGamma	invented number	The quote retains specific numeric price levels ("720 is probably going to be your upper end resistance" and "we don't really get any support until we get down to the 715 level") that function as price targets/support-resistance facts; the numeric scrubber should have removed these creator-stated price numbers before surfacing.
TSLA	negative	Thoughtful Money	not grounded, misresolved, bad attribution, stance mismatch	The quote is explicitly flagged as not located in the transcript (quote_located_in_transcript=false), does not appear anywhere in the transcript window, and the window itself centers on SpaceX's valuation vs. Microsoft — not Tesla — making grounding, resolution, attribution, and stance all unverifiable.
SPY	negative	Thoughtful Money	stance mismatch	The creator expresses a relative preference for RSP over SPY due to concentration-risk concerns ("my recurring buy is not on the SPY... something that I am doing personally"), which is a cautious/mixed stance, not a clearly negative one — they explicitly acknowledge SPY remains valid for buy-and-hold investors and never assert SPY will fall; labeling this "negative" overstates the directional claim.
MSFT	neutral	Thoughtful Money	not grounded, stance mismatch	Microsoft is used solely as a size/profitability benchmark to critique SpaceX's valuation — the creator expresses no stance toward MSFT itself, so the transcript window does not support any sentiment extraction on Microsoft, and the "neutral" stance label is a mislabel of a purely comparative reference.
MU	positive	Bloomberg Technology	bad attribution	The quote is spoken by guest Tiffany Wade, a portfolio manager at Columbia Threadneedle being interviewed, not by the Bloomberg Technology channel host/creator; attribution_is_creator fails because this is a guest's view, not the creator's own stance.
MSFT	positive	Bloomberg Technology	not grounded, stance mismatch	Microsoft is mentioned only in passing as one item in a list ("Micron and Anthropic and then also around Microsoft and more data center spending in Texas") to support a broad AI infrastructure spending thesis — the creator is not expressing a specific positive stance on Microsoft as a security; the transcript window's focus is on semiconductor stocks and Micron earnings, making both grounded and stance_matches false.
MU	positive	Bloomberg Technology	bad attribution, stance mismatch	The quote is a Bloomberg TV anchor's factual real-time price observation ("stock is up 4.4%, bucking the broader trend"), not a named finance creator's own analytical sentiment; additionally, the stance label "positive" misrepresents what is neutral market reporting of a price move, not an expression of a bullish view on Micron.
MSFT	neutral	Bloomberg Technology	not grounded	The transcript window shows only a bare price ticker call ("MICROSOFT SHARES, MODESTLY LOWER. MICROSOFT, DOWN 2%.") — a factual intraday price readout during a live news segment — not any expression of sentiment stance toward Microsoft as a security; a price flash is not creator sentiment, so grounded=false.
MSFT	mixed	Value Investing with Sven Carlin, Ph.D.	invented number	The quote retains "likely 10% plus return going forward" — a specific quantitative forward-return projection presented as a factual read on the stock's current pricing; this is a numeric claim that should have been removed by the scrubber before reaching review.
GOOGL	positive	Bianco Research	stance mismatch	The creator's "I'm so bullish on AI" is bullishness on AI as a theme, not on GOOGL stock; the Google-specific content frames Google's situation as existential risk — it must win the AI race or lose Gmail, Maps, and everything else — which is a conditional/mixed view on GOOGL, not a straightforwardly positive stance.
SPY	mixed	Bianco Research	not grounded, misresolved, stance mismatch	The transcript shows the creator discussing the AI trade, using the S&P 500 only as a rhetorical illustration ("If I use the S&P 500 as an example"); the volatility/thrill-ride sentiment is directed at AI stocks, not SPY, so resolution_correct and grounded both fail.
NVDA	positive	Real Vision	stance mismatch	The creator uses NVDA as a foil to argue that chips are NOT the bottleneck — "chips sitting in warehouses cuz there's nowhere to plug them in" conveys that Nvidia supply is irrelevant to the real constraint (power/infrastructure), which is neutral-to-negative context for NVDA, not a positive stance; labeling it "positive" misrepresents the creator's expressed view.
MU	negative	Morningstar, Inc.	misresolved	The creator's negative thesis is about memory stocks as a sector ("memory is a classic example..."); Micron is named only once as one of four interchangeable commodity suppliers ("SanDisk, Micron, Samsung, SKH Highix"), never singled out as the subject of the bearish view — resolving this to MU specifically misidentifies Micron as the primary security when the creator's stance targets the memory category broadly.
META	positive	The Investor's Podcast (We Study Billionaires)	stance mismatch	The quote is selectively truncated to capture only Tobias Carlisle's bullish opening remarks, but his full statement in the transcript window clearly expresses a mixed view — he raises material uncertainty about AI CapEx monetization, warns of potential multi-year earnings underperformance, and describes the stock as "probably reasonable value" with "some discount for uncertainty," making the actual stance mixed rather than positive.
MSFT	positive	The Investor's Podcast (We Study Billionaires)	not grounded, misresolved, stance mismatch	Microsoft is referenced only as a comparison ("just like Microsoft Office") to describe Adobe's bundled product suite — the entire transcript window is discussing Adobe (ADBE), not Microsoft; MSFT is misresolved and the transcript does not express any stance toward Microsoft as a security.
CRM	negative	20VC with Harry Stebbings	bad attribution	The Salesforce quote is spoken by guest Ryan Peterson (founder of Flexport), not by the creator/host Harry Stebbings — the transcript window explicitly identifies Peterson as "our guest today," and the quote appears in the pre-interview teaser highlights attributed to him, failing the attribution_is_creator check.
META	negative	Monetary Matters	bad attribution	The transcript window uses ">>" markers indicating two distinct speakers; the quoted critical takes on Meta's AI strategy ("I don't think Meta really has a strategy", "Mark Zuckerberg is a capricious man") appear in ">>" blocks attributed to a second voice, not clearly to the Monetary Matters channel creator — the attribution guard fails because the speaker identity is ambiguous and may be a guest rather than the creator.
NVDA	mixed	Thoughtful Money	stance mismatch	The creator's bearish warning about parabolic advances correcting "very hard" is directed explicitly at AMD, Micron, Marvel Technologies, and Qualcomm — not at Nvidia; Nvidia is mentioned only as a holding ("We own Nvidia and Broadcom"), which is a positive signal, so labeling the NVDA stance as "mixed" misattributes the bearish sentiment that belongs to the other named semiconductors.
AVGO	mixed	Thoughtful Money	bad attribution	The "We own Nvidia and Broadcom" statement is in a non-">>" block in an interview-format transcript; on Thoughtful Money the ">>" lines are the host's short interjections ("Sure.", "Yeah.", "Right?") while the long analytical passages belong to a guest speaker, so the stance is attributed to a guest, not the Thoughtful Money creator.
AMD	neutral	Daniel Pronk	not grounded, misresolved, stance mismatch	AMD appears only as a size-comparison aside ("larger than AMD today") — the entire quote and surrounding analysis are about Amazon's (AMZN) undervalued chips business and Amazon's share price, making the ticker misresolved and the stance on AMD entirely unsupported by the transcript window.
NVDA	negative	Thoughtful Money	stance mismatch	The creator cites NVIDIA's on-device chip development as corroborating evidence of a broader data-center-demand decline thesis — NVIDIA is framed neutrally/positively (adapting to the trend), not negatively; the negative stance applies to the semiconductor/data-center space broadly, so labeling the NVIDIA entry as "negative" is a stance mismatch.
AAPL	positive	Thoughtful Money	misresolved, stance mismatch	Apple/AAPL is not the subject of the creator's sentiment — the Mac mini mention is an incidental data point in a bearish data-center/semiconductor thesis, not an expression of positive sentiment toward Apple; additionally, resolution_correct fails because the creator is not discussing Apple as an investable subject at all, making ticker assignment unsupported.
MSFT	negative	Trade Brigade	stance mismatch	The creator's expressed stance is mixed, not purely negative: while they state "structural trends are bad across Microsoft" on the givebacks, they simultaneously say they "allow names like Microsoft to sort of regroup" and watch the daily 50 for potential stabilization — a wait-and-see posture that the negative label fails to capture.
QQQ	mixed	Brian Shannon - Alpha Trends	invented number	The quote retains the specific price level "720 area" — a creator-stated key technical level — which the numeric scrubber should have removed; a concrete number presented as a fact/level survived scrubbing, failing the no_invented_number gate.
AAPL	mixed	Brian Shannon - Alpha Trends	invented number	The quote retains the specific price level "290" ("I'd love to see it come right back down towards that 290"), which is a creator-stated pullback target that the numeric scrubber was required to remove — the scrub pass failed to strip it.
GOOGL	mixed	Brian Shannon - Alpha Trends	stance mismatch	The creator's expressed view on Google is cautiously positive — "shaping up a little bit better," holding the 50-day moving average, prior resistance held as support, "perhaps it could get going" — with no negative element stated; the stance should be 'positive' (or at minimum 'neutral'), not 'mixed'.
TSLA	positive	20VC with Harry Stebbings	bad attribution	The transcript is a multi-speaker panel on 20VC; the quoted passage ("you could say that about Tesla shareholders, too...") is spoken by a guest, not by Harry Stebbings — the ">>" speaker-change markers and the reference to "Rory" as a named third participant show this is a guest's statement, not the channel creator's own view.
MSFT	positive	Value Investing with Sven Carlin, Ph.D.	stance mismatch	The creator's comment on Microsoft is a heavily hedged relative comparison ("perhaps, I don't know, in a diversification strategy, it's better to add Microsoft than Adobe") made while rejecting Adobe — not a positive endorsement of MSFT; labeling the stance "positive" overstates what the creator expressed, which is closer to neutral/uncertain.
INTC	positive	Bloomberg Technology	bad attribution, stance mismatch	The Bloomberg anchor is neutrally reporting a news fact (Intel stock at record high due to a Trump Truth Social post) — this is broadcast news narration, not the creator's own expressed sentiment about Intel; labeling it "positive" stance misattributes an external price event as creator opinion.
AAPL	mixed	Bloomberg Technology	bad attribution	The Bloomberg Technology anchor is neutrally relaying factual market data (Apple pared its intraday gain) and a third-party CEO statement from a WSJ interview — the host expresses no personal sentiment toward AAPL, so the "mixed" stance cannot be attributed to the creator as their own view.
MSFT	positive	Bloomberg Technology	bad attribution, stance mismatch	The quote is a news anchor (Ed) reading a headline tease before handing off to a guest reporter (Matt Day) — this is straight news reporting, not the creator expressing their own bullish investment sentiment; additionally, the stance "positive" is wrong because reporting "Microsoft is cashing in on China" is a neutral factual description of business activity, not an endorsement or bullish view.
INTC	positive	No Priors	bad attribution, stance mismatch	The quote is spoken by Lip-Bu Tan (Intel's CEO, the guest/interviewee), not by the No Priors podcast hosts (the creators); attributing the guest's self-description of his own turnaround strategy as the creator's positive stance on INTC is a clear attribution failure.
NVDA	neutral	Option Alpha	bad attribution, stance mismatch	The quote is a pedagogical explanation of how a short put trade structure works (using NVIDIA as a live teaching example), not the creator's own market sentiment on NVIDIA — "we're not saying Nvidia is going up, we're saying it probably won't drop that far" describes what the trade mechanics imply, not an original view the creator holds; attribution_is_creator fails because we cannot confidently attribute this as genuine creator sentiment rather than instructional narration, and stance_matches also fails since no personal sentiment on NVIDIA is actually expressed.
META	mixed	Brian Feroldi	not grounded, misresolved, stance mismatch	Meta Platforms is invoked solely as a historical IPO analogy to illustrate post-IPO price discovery risk for SpaceX — it is not the security under discussion, so the transcript window does not support any current sentiment stance on META, and the ticker resolution is wrong (the subject security is SpaceX, not Meta).
TSLA	mixed	Brian Feroldi	not grounded, bad attribution, stance mismatch	The quoted text about the Tesla/SpaceX merger complexity is spoken by a co-presenter (Stouffer, indicated by the >> turn markers), not by Brian Feroldi, who contributes only brief "Yep" affirmations; additionally, the entire discussion is a speculative hypothetical ("What happens IF Tesla merges with SpaceX?") rather than a grounded current stance on Tesla, so neither attribution_is_creator nor grounded holds.
MSFT	negative	Value Investing with Sven Carlin, Ph.D.	not grounded, misresolved, stance mismatch	Microsoft appears only as one name in a passing list (Google, Microsoft, Meta, SpaceX, Anthropic, OpenAI) illustrating a general AI-bubble argument; the transcript window shows the creator's focused subject is SpaceX and the broad AI/IPO bubble, not Microsoft specifically — resolving this as a MSFT-targeted stance is a homonym/wrong-subject error, and the window does not support a Microsoft-specific negative view.
GOOGL	negative	Value Investing with Sven Carlin, Ph.D.	misresolved	GOOGL is named only as one item in a multi-company list ("Google, Microsoft, Meta, SpaceX, Anthropic, OpenAI") used as evidence for a broad AI-sector bubble argument; the "hugely overpriced" claim is directed at the overall market phenomenon of equity being given away freely, not at Alphabet specifically — resolution_correct fails because Google is an incidental example, not the subject of a distinct creator take on GOOGL as a security.
META	negative	Value Investing with Sven Carlin, Ph.D.	misresolved	META appears only as one name in a list of companies (Google, Microsoft, Meta, SpaceX, Anthropic, OpenAI) used to illustrate a general AI-sector bubble thesis; the creator expressed no stance specifically about Meta Platforms, so resolving this to ticker META is a misresolution.
AMZN	mixed	Value Investing with Sven Carlin, Ph.D.	misresolved, stance mismatch	Amazon is used solely as a comparison benchmark to argue SpaceX is in a bubble — the creator expresses no stance on Amazon as an investment subject; it is a passing reference, not the security under discussion, so resolution_correct and stance_matches both fail.
META	mixed	Brian Stoffel	not grounded, misresolved, stance mismatch	Meta is used solely as a historical IPO comparison to illustrate SpaceX price-discovery risk — the creator expresses no current stance on META as a security, so grounded=false, resolution_correct=false (Meta is an aside/comparison, not the subject), and the "mixed" stance label has no basis in the transcript window.
TSLA	negative	Brian Stoffel	bad attribution	The negative Tesla/Solar City take ("not a good move," "extremely disappointing") is voiced by the second speaker in the dialogue (marked ">>"), not by Brian Stoffel, who only responds "Yep." — making this hearsay from a co-host, not Brian Stoffel's own stated view.
NVDA	mixed	42 Macro	not grounded, stance mismatch	The quote "A key AI supplier, Nvidia" is a bare descriptor with no expressed stance; the transcript uses NVIDIA as a factual example in a macro circular-financing argument, and the creator's conclusion is that circular financing concerns are "mistimed" while global liquidity expands — there is no mixed (or any directional) stance on NVIDIA as a security in the transcript window.
GOOGL	mixed	42 Macro	not grounded, stance mismatch	The transcript window shows Alphabet mentioned only as a factual data point in a macro circular-financing discussion (e.g., "Alphabet bought a stake in SpaceX in 2015"); Darius Dale expresses no stance toward GOOGL as a security — grounded fails and the recorded stance of "mixed" has no basis in the window.
MSFT	mixed	42 Macro	not grounded, stance mismatch	The quote is a neutral factual data point ("Microsoft owns 27% of OpenAI") used as analytical context in a macro discussion about AI circular financing — the creator expresses no discernible stance on Microsoft as a security, so neither "mixed" nor any other directional stance is grounded in the transcript window.
IWM	positive	Trade Brigade	invented number	The quote retains the specific price level "288" ("a very clean higher low 288"), a creator-stated price number that the numeric scrubber should have removed; surfacing it in the output violates the removal-only scrubbing gate.
CRM	negative	Trade Brigade	not grounded, bad attribution, stance mismatch	The quote "CRM loser." is explicitly flagged as not located in the transcript (quote_located_in_transcript: false), and the transcript window contains zero mention of Salesforce or CRM — it covers FOMC/Fed commentary and generic broad-market technical chart analysis, so the quote is entirely ungrounded and cannot be attributed to the creator.
SPY	mixed	Bianco Research	not grounded, misresolved	The creator never mentions SPY by name or expresses a stance on the SPDR S&P 500 ETF as a security; his discussion is about "non-AI stocks" as a broad market category within the S&P 500, making the resolution to SPY a category-level misattribution — SPY tracks the full S&P 500 (both AI and non-AI), not the non-AI subset the creator is actually discussing.
IWM	positive	EarningsBeats	stance mismatch	The creator explicitly notes IWM "lagged yesterday" while other indices (NDX, semiconductors) surged over 3%; the all-time high mention is a factual observation, not an expression of positive sentiment — the overall framing is neutral/factual, making the "positive" stance label a mismatch.
NVDA	neutral	The Julia La Roche Show	bad attribution	The quote is spoken by guest Ted Oakley (founder of Oxbow Advisors), not by the channel creator Julia La Roche — attribution_is_creator fails because this is a guest's observation, not the host's own expressed view.
QQQ	negative	The Julia La Roche Show	not grounded, misresolved, stance mismatch	The creator never names QQQ — they say "buy the NASDAQ" as one item in a broad list of passive ETF behaviors; the negative warning is directed at the general strategy of passive index investing, not at QQQ as a specific security, making both the ticker resolution and the grounding of a QQQ-specific negative stance unsupported by the transcript window.
SPY	negative	The Julia La Roche Show	bad attribution	The negative ETF sentiment ("those are the ones that will get you") is expressed by the guest being interviewed, not by Julia La Roche (the channel creator/host), who plays the interviewer role throughout the transcript — making this a guest's view, not the creator's own stance.
QQQ	positive	Bravos Research	stance mismatch	The quote is a product-advertisement description of the creator's quantitative signal model mechanics ("Aggressive means 2x leverage long on the QQQ tech ETF") — it defines what the model's signals entail, not an expressed current bullish stance on QQQ itself; stance "positive" is therefore a mislabel, since the model definition includes a fully-out-of-market "cash" signal and the quote is promotional copy selling a subscription service rather than the creator voicing a directional view on QQQ.
CRM	negative	SaaStr AI	not grounded, bad attribution, stance mismatch	The quote is a guest entrepreneur's product-UX complaint about Salesforce's dashboard UI during a Replit workflow discussion, not the channel creator's financial sentiment on CRM as a security; the transcript window shows no investment stance toward CRM, so grounded=false, attribution_is_creator=false, and stance_matches=false (labeling a tool-friction remark as a negative financial stance is a category error).
NVDA	positive	Monetary Matters	bad attribution	The quote about Nvidia's massive margins and chip design monopoly is spoken by a guest/interviewee, not the Monetary Matters channel creator; the transcript shows two speakers delimited by ">>" where the host asks questions ("Tell us...", "Yes. Yeah. I I think you're you're absolutely right...") while the guest delivers the analytical content including this quote — attribution_is_creator fails.
MSFT	negative	Monetary Matters	stance mismatch	The creator frames Microsoft's capex pullback as historically rational conservatism ("Microsoft historically has always been more conservative") while noting issues with the OpenAI exclusivity loss and Copilot adoption; the overall characterization is mixed/explanatory rather than clearly negative, making the "negative" stance label a mismatch.
TSLA	negative	Monetary Matters	bad attribution	The Tesla robo taxi comment is made by a guest interviewee (a former Bank of America macro strategist), not the channel creator — the transcript's ">>" markers consistently mark the host's brief interjections while the long monologue containing the quote belongs to the guest, so attribution_is_creator fails.
GOOGL	neutral	Patrick Boyle	not grounded, stance mismatch	The transcript window shows Alphabet mentioned only as a factual data point (it did a large equity offering) within a macro critique of AI capital-raising; the creator expresses no stance toward GOOGL as a security, so grounded=false and stance_matches=false — labeling this as "neutral" sentiment on GOOGL misrepresents a passing factual aside as a creator view on the stock.
NVDA	neutral	Patrick Boyle	not grounded, stance mismatch	The transcript window shows NVIDIA mentioned only as a factual aside — private companies "passing the hat to pay for Nvidia chips" — with no evaluative stance from the creator toward NVIDIA as a security; the recorded "neutral" stance is unsupported because Boyle expresses no sentiment about NVIDIA at all, only uses it as a descriptive backdrop for the SpaceX IPO discussion.
SMCI	negative	EarningsBeats	stance mismatch	The creator is neutrally recapping weekly S&P 500 performance data — stating SMCI "lost 27%" as a factual sorted-table data point — not expressing a negative opinion or bearish sentiment toward SMCI; the stance should be neutral, not negative.
TSLA	mixed	The Plain Bagel	stance mismatch	The stance is labeled "mixed" but the quote and surrounding context are uniformly skeptical — Tesla's "rich valuation" despite "falling earnings and slowing revenue growth" is offered as a cautionary example of Musk-associated hype, with no positive dimension expressed; the correct stance is negative.
GOOGL	neutral	The Plain Bagel	not grounded, stance mismatch	Google/GOOGL appears only as a passing valuation comparison to illustrate SpaceX's elevated P/S ratio; the creator expresses no stance on Alphabet as a security — it is a comparison aside in a SpaceX IPO video, not a view on GOOGL itself, so grounded=false and stance_matches=false.
AAPL	mixed	Thoughtful Money	bad attribution	The Apple commentary ("Apple's never the first... my guess is they're laying in the weeds...") is spoken by a guest interviewee, not the Thoughtful Money host/creator — the >> speaker-turn markers in the transcript show the host's lines are the short acknowledgments ("Right"), while the extended Apple analysis belongs to the guest.
GOOGL	mixed	Thoughtful Money	bad attribution	The views in the quote ("highly valued," equity above intrinsic value) are expressed by a guest speaker on the Thoughtful Money channel, not by the channel creator (Adam); the transcript's `>>` speaker-change markers show the analytical voice making these statements is the interviewee, not the host, so attribution_is_creator fails.
SPY	negative	Thoughtful Money	bad attribution	The cautionary stance ("this is a time to be cautionary... let's be cautious here") is expressed by a GUEST being interviewed, not by the Thoughtful Money channel creator — the host Michael is asking "what's your bias?" and the guest answers, so attribution_is_creator fails.
GOOGL	negative	Capital Allocators with Ted Seides	bad attribution	The negative views on Google's margins and depreciation policy are expressed by a guest speaker, not Ted Seides (the channel creator); Seides is the interviewer asking questions, making this hearsay from a guest rather than the creator's own stance.
MU	mixed	Chip Stock Investor	stance mismatch	The creator is conducting a teaching exercise on business-cycle classification, labeling Micron "short cyclical" as a business-type descriptor — there is no mixed investment sentiment expressed; a neutral label fits the transcript, not mixed.
NVDA	mixed	Forward Guidance	not grounded, bad attribution, stance mismatch	The quote "Oh my gosh, Nvidia and these companies are so cheap. Like" is not the creator's own view — the creator explicitly frames it as a claim "pedled a lot" on Twitter by others, which the creator then argues against; the creator's actual expressed stance on NVDA is negative/bearish, making both attribution_is_creator and stance_matches false.
META	negative	Bravos Research	stance mismatch	The creator is making a macro market-structure observation — that large-cap tech stocks including META are experiencing selling pressure ahead of the SpaceX IPO — not expressing a negative view on META as a security; META is one of four stocks used as incidental examples of a broader thesis, so labeling the stance as "negative" on META misrepresents what the creator actually expressed.
AMZN	negative	Bravos Research	stance mismatch	The quote is a neutral/observational statement that Amazon (along with Meta, Microsoft, and Broadcom) has been experiencing selling pressure due to IPO dynamics — the creator is not expressing a negative view on Amazon as a security; labeling the stance "negative" misrepresents a descriptive macro observation as a bearish take on AMZN specifically.
MSFT	negative	Bravos Research	stance mismatch	Microsoft is mentioned in passing as one of four example stocks illustrating a macro IPO-liquidity thesis; the creator makes no directional call on MSFT itself, so the stance is observational/neutral, not "negative."
AVGO	negative	Bravos Research	stance mismatch	The creator uses Broadcom as one of several example large-cap tech stocks to illustrate a macro flow argument about IPO-driven selling pressure, explicitly saying the reason for selling is unknown; this is a neutral macro observation, not the creator expressing a negative stance on AVGO specifically, so the recorded stance of "negative" is mislabeled.
QQQ	negative	Macro Voices	bad attribution	The bearish Nasdaq 100 view (valuation concerns, rotation to value/energy) belongs to the guest "Larry," not the Macro Voices host (Eric); the transcript shows Eric asking questions while Larry provides the analysis, so attribution_is_creator fails.
GOOGL	negative	Macro Voices	not grounded, invented number	The quote contains the phrase "the secondary offering from Google last week of shares, 80 billion" but this specific claim does not appear in the transcript window — the window only shows "you got Google dumping stock" as the Google-specific text; additionally, "80 billion" is a specific number surfaced in the quote as a stated fact/size that was not scrubbed.
SPY	negative	Forward Guidance	misresolved, bad attribution	Two defects: (1) the bearish macro view is expressed by a guest speaker (the >> dialogue markers show a two-speaker interview format on the Forward Guidance channel), not the channel creator — attribution_is_creator fails; (2) the creator says "bad for stocks" as a generic asset-class statement without mentioning SPY or any specific ETF, making the resolution to SPY an unsupported inference — resolution_correct fails.
CRM	positive	SaaStr AI	not grounded, misresolved, stance mismatch	The transcript window does not express positive sentiment toward Salesforce Inc. (CRM); the creator uses "a crummy clone of Salesforce from 2006" as a dismissive negative reference point to contrast with new AI CRM startups — the positive stance is directed at those unnamed startup vendors, not at Salesforce the company, making both the resolution and the stance label incorrect.
MSFT	negative	Brian Stoffel	not grounded, misresolved, bad attribution, stance mismatch	Multiple defects: (1) The quote "I haven't been impressed with Microsoft" is spoken by the guest (Justin, identified by ">>" speaker prefix), not by Brian Stoffel the channel host — attribution_is_creator fails. (2) The video's subject is Coreweave (CRWV), not MSFT; Microsoft is mentioned only as Coreweave's dominant customer (67% concentration risk), making MSFT a contextual aside rather than the security under analysis — resolution_correct fails. (3) Because it is a guest's aside about a company that is not the video's subject, the transcript window does not ground a standalone MSFT sentiment claim by the named creator — grounded fails.
NVDA	positive	Chip Stock Investor	not grounded, misresolved, stance mismatch	The transcript window is a video about Coherent Corp (COHR), not NVIDIA — NVIDIA appears only as a cash-rich customer/investor providing context for Coherent's balance sheet situation; the creator expresses no discernible stance on NVDA as a security, and the ticker resolution is wrong because NVIDIA is a supporting reference, not the subject of the creator's analysis.
IWM	mixed	Trade Brigade	invented number	The quote retains the specific price level "280" paired with a directional expectation ("I would expect a break through 280 based on how weak this action is up here"), which surfaces a creator-stated numeric level with directional implication — the numeric scrubber should have removed this number before the row was cleared.
GOOGL	positive	Trade Brigade	stance mismatch	The creator's quote is a relative preference within a bearish context — immediately followed by "it's kind of ugly. It's really ugly" and framed as Google "not helping" the sector — making the actual stance mixed, not positive; labeling it positive overstates the creator's expressed view.
AAPL	mixed	Cestrian Capital Research	invented number	The quote retains "up to 10% downsi[de]" — a specific numeric downside magnitude the creator stated — which the numeric scrubber failed to remove (it was only truncated mid-word, not excised), so a creator-stated price-move figure is surfaced in the output, violating the no_invented_number gate.
QQQ	negative	EarningsBeats	stance mismatch	The creator explicitly rejects a bearish label — "This isn't bearish. This isn't bearish momentum. It's slowing bullish momentum" and "I still look at this as though we are in a bullish environment but the momentum is slowing" — making the recorded stance of "negative" a mislabel; the actual expressed stance is mixed/cautious (slowing bullish momentum with potential short-term weakness).
SPY	positive	TraderLion	bad attribution	The "our core position is 30% in the S&P 500" statement belongs to a guest being interviewed (describing their own firm's — Growction/River — strategy), not the TraderLion channel creator's own view; attribution_is_creator fails.
PLTR	mixed	Ticker Symbol: YOU	not grounded, stance mismatch	Palantir appears only as a valuation benchmark to contextualize SpaceX's price-to-sales ratio; the creator never expresses a stance on Palantir as an investment subject — it is a passing comparison in a SpaceX analysis, so grounded=false and stance_matches=false.
MSFT	negative	Capital Allocators with Ted Seides	not grounded, misresolved	The negative capex/FCF thesis in the transcript window is directed at the Mag 7 group broadly (Google, Nvidia, Amazon) and not specifically at Microsoft — Microsoft appears only as a passing aside ("more than half open AI"), so the window does not ground the negative stance specifically on MSFT, and resolving this group critique to the MSFT ticker is a misresolution.
NVDA	positive	Ticker Symbol: YOU	not grounded, misresolved, stance mismatch	NVIDIA is mentioned only as a valuation benchmark ("SpaceX four times more expensive than Nvidia") to argue against buying SpaceX stock — the creator expresses no stance on NVIDIA itself; it is a comparison aside, not the subject of the segment.
TSM	positive	Asianometry	not grounded, misresolved, stance mismatch	The positive sentiment ("so far it looks good") in the quote refers to Rapidus, a Japanese domestic foundry, not to TSMC/TSM; TSMC appears only as the lead of JASM cited as context for a model EU chip policy, and the transcript window does not show the creator expressing any positive stance toward TSM as a security.
MSFT	positive	Joseph Carlson	stance mismatch	The creator says "I remain optimistic" about the hyperscalers including Microsoft, but the transcript window shows him immediately and extensively presenting Steve Eisman's bearish counter-arguments (AI is a commodity with no moat, companies transforming from asset-light to capex-heavy in a negative way), making the expressed stance mixed rather than positive — labeling it "positive" misrepresents the creator's qualified, both-sides framing.
GOOGL	positive	Invest Like The Best	bad attribution	The quote originates from a guest investor named Alex, not from the "Invest Like The Best" channel creator/host — the transcript explicitly distinguishes "Alex, you were saying..." as a guest speaking, so the sentiment cannot be attributed to the creator.
AVGO	positive	Invest Like The Best	not grounded, misresolved, stance mismatch	Broadcom (AVGO) appears only as a company Celestica works closely with — the entire creator pitch is about Celestica's competitive advantages; the transcript window never expresses a positive stance toward Broadcom itself, making both grounded and resolution_correct false.
QCOM	positive	Invest Like The Best	not grounded, stance mismatch	Qualcomm is cited in past tense as a historical textbook example of an IP moat ("what Qualcomm had") within a generic taxonomy of competitive advantages — the transcript window shows no forward-looking or current positive stance on QCOM as a security, so grounded=false and the "positive" stance label is unsupported.
TSM	mixed	Monetary Matters	bad attribution	The TSMC commentary is made by a guest journalist being interviewed, not the host/creator of the Monetary Matters channel — the transcript shows two distinct speakers separated by ">>" markers, with the host asking questions and the guest (self-described "journalist") making the TSMC bottleneck/bull-market claim; attribution_is_creator fails.
NVDA	mixed	Capital Allocators with Ted Seides	bad attribution	The views on NVIDIA are expressed by a GQG Partners guest/interviewee (not Ted Seides, the channel creator and host), as evidenced by references to "GQG's history" and the interview format with Ted asking follow-up questions — attribution_is_creator fails.
AMZN	negative	Capital Allocators with Ted Seides	bad attribution	Capital Allocators with Ted Seides is an interview-format podcast; the negative views on Amazon's AI data center economics ("Our view is that this is a powerful technology but the economics are really bad") are expressed by a guest/interviewee (a fund manager), not by Ted Seides the channel creator/host — attribution_is_creator is false.
AAPL	mixed	Capital Allocators with Ted Seides	not grounded, stance mismatch	The transcript window shows the creator explicitly deferring Apple from analysis ("Apple is a unique animal. So, let's leave Apple aside.") — this is a non-stance, not a mixed view; the window does not support the creator expressing any actual sentiment toward Apple, so both grounded and stance_matches fail.
MU	negative	Capital Allocators with Ted Seides	misresolved	The transcript window never mentions Micron Technology or the ticker MU — the creator's skepticism is directed at "memory prices" and the memory sector broadly; resolving generic sector commentary to the specific company MU is a misresolution.
IWM	mixed	Trade Brigade	invented number	The quote retains the specific price level "270s" — a number the creator states as an actionable watch level ("I would really be paying attention to about 270s previous resistance offering support here") — which the numeric scrubber should have removed before this row was eligible for review; surfacing it violates the removal-only scrubbing gate.
NVDA	negative	The Julia La Roche Show	bad attribution	The negative NVIDIA view ("I would say no") belongs to a guest named Chris, not to the channel creator Julia La Roche — the quote is hearsay from an interview guest, not the creator's own expressed stance.
MU	positive	The Julia La Roche Show	bad attribution	The Micron sentiment is expressed by a guest on The Julia La Roche Show, not by Julia La Roche (the channel's creator/host) — the transcript shows Julia asking questions while the unnamed guest makes the MU claim, so this is hearsay from a guest, not the creator's own stance.
GOOGL	positive	The Julia La Roche Show	bad attribution, stance mismatch	The creator (Chris Whalen) is describing crowd/market sentiment toward Google ("reckoned to be," "flavor of the week"), not expressing his own positive view — the statement is hearsay crowd-reporting, so attribution_is_creator fails; and the framing is skeptical/neutral about momentum-driven crowd behavior, not genuinely positive, so stance_matches also fails.
AMD	mixed	The Julia La Roche Show	stance mismatch	The creator expresses a straightforward profit-taking exit from AMD (and ARM together) after large gains, with no expressed ambivalence about AMD itself — the stance is a clean exit/neutral, not "mixed"; labeling it "mixed" misrepresents the creator's unambiguous sell-and-move-on statement.
ARM	mixed	The Julia La Roche Show	bad attribution	The quoted text blends two speakers: "I got out of arm" is Julia La Roche's (the host's) own line (marked with >> in the transcript), while the guest then echoes "AMD, arm, yeah" in confirmation — the quote cannot be cleanly attributed solely to the creator/guest, violating the attribution guard.
TSLA	positive	Aswath Damodaran	not grounded, misresolved, bad attribution, stance mismatch	Tesla is invoked only as a historical comparator in a SpaceX IPO discussion — Damodaran is noting that "other people" who bet on Tesla/Musk did well, not expressing his own current positive stance on TSLA; the window does not show Damodaran analyzing or taking a position on Tesla, so grounded, resolution_correct, attribution_is_creator, and stance_matches all fail.
GOOGL	positive	FAST Graphs	stance mismatch	The creator is using Alphabet as a teaching example to illustrate that the P/E=growth rate formula works well — they make no positive investment claim about GOOGL; labeling the stance "positive" misrepresents a neutral methodological illustration as a bullish take on the stock.
META	positive	FAST Graphs	stance mismatch	The creator uses Meta Platforms solely as a teaching example to illustrate a P/E-equals-growth-rate valuation framework — stating "would have been a very rational valuation" as a historical observation — and expresses no bullish sentiment about META; the recorded stance of "positive" is a mislabel of what is an analytical/neutral educational mention.
CRM	negative	Jussi Askola, CFA	not grounded	Salesforce appears only as one item in a generic list of SaaS companies used to illustrate a sector-level AI-disruption thesis; the creator never expresses a stance specific to Salesforce/CRM, so the transcript window does not ground a Salesforce-specific evidence row.
PLTR	mixed	Brian Stoffel	bad attribution	The transcript window shows a two-person conversation with an unnamed co-host, and the quoted summary line ("in absolute terms, this company is expensive...") is not clearly marked as Brian Stoffel's speech versus the co-host's — the attribution guard fails closed on this ambiguity.
TSLA	negative	Wall Street Millennial	not grounded, stance mismatch	Tesla (TSLA) is referenced only as a comparison example to illustrate deceptive teleoperation practices in humanoid robotics, not as the subject of an investment-sentiment discussion; the transcript window does not show the creator expressing a stance on Tesla as a security, making both grounded and stance_matches false.
SPY	positive	EarningsBeats	stance mismatch	The creator explicitly frames SPY as a passive benchmark to beat ("I could just put my money in the spider and just do whatever I want and get this return... we want to beat that"), expressing a neutral-to-negative stance toward SPY as an investment — not a positive one; the "positive" label is a mismatch.
QCOM	positive	High Yield	stance mismatch	The stance is recorded as "positive" but the creator's view is mixed/neutral: they note Snapdragon X2 has higher CPU performance than RTX Spark on Windows on Arm, but also that Windows on Arm is poor for gaming and the video closes dismissively ("not a new era for PC"). The quote reflects a narrow competitive-comparison point, not a broadly bullish stance on Qualcomm.
SPY	positive	SpotGamma	invented number	The quote contains "the clear target developing at SPX 6830 above" — a specific numeric level presented as a trade target/destination that the creator acted on, which violates the no-invented-number rule (the scrubber must remove target numbers, not merely creator-stated price levels used as actionable targets).
MU	mixed	The Julia La Roche Show	bad attribution	The Micron sentiment is expressed by a guest analyst (Chris, the interviewee who refers to "we wrote this week Julia"), not by Julia La Roche who hosts The Julia La Roche Show — attribution_is_creator fails because the view belongs to a guest, not the channel creator.
ASML	mixed	Anastasi In Tech	stance mismatch	The extracted quote ("ASML will continue dominating because its machines already exist and the entire semiconductor ecosystem already knows how to build the manufacturing recipe around them") is unambiguously positive/bullish, but the stance is labeled "mixed"; the surrounding transcript context discusses competitive alternatives (China's FEL approach) that inform broader video tone, but the quote as attributed expresses clear positive sentiment, making the "mixed" label a stance mismatch.
IWM	mixed	Trade Brigade	invented number	The quote retains "28675 is going to be the key watch," a specific creator-stated price level that the numeric scrubber should have removed — surfacing a concrete number as a watch target violates the removal-only scrubber rule.
MSFT	negative	Trade Brigade	invented number	The quote retains the specific price level "430 breakout point" — a concrete number presented as a fact/target — which the numeric scrubber should have removed before this row was surfaced.
GOOGL	negative	Trade Brigade	stance mismatch	Stance is labeled "negative" but the creator immediately qualifies the bearish description by warning against being "short in the hole," noting a possible 50 SMA bounce, and explicitly saying "I don't want to get overly pessimistic" while pointing to a bullish weekly higher-low structure — making the actual expressed stance mixed, not negative.
ORCL	mixed	Trade Brigade	not grounded, misresolved, stance mismatch	The extracted quote "Those things come to mind for our IGV" is a closing transitional sentence about the IGV ETF, not a statement about Oracle specifically; the creator's actual Oracle comment ("I don't think Oracle's broken by any stretch of the imagination") is a passing neutral remark, making the ticker resolution wrong (the quote is about IGV, not ORCL) and the stance label "mixed" unsupported by the brief neutral aside.
JPM	negative	Trade Brigade	stance mismatch	The transcript window explicitly states "JPM technically up over here" and the creator describes big banks as "trying to hang in there," which is neutral-to-mildly positive; the quote "hanging out on the lows" is a passing chart observation, not a negative thesis, making the negative stance label a mismatch.
MU	negative	SpotGamma	not grounded, stance mismatch	The transcript window shows the creator using Micron only as an example of extreme call-option volume in a macro call-froth analysis; the creator never expresses a negative view on MU itself and explicitly hedges any mean-reversion implication with "who knows exactly what that even means" — the "negative" stance label is not grounded in the window.
NVDA	positive	FAST Graphs	stance mismatch	The transcript shows NVIDIA used as a historical backtested illustration of the FAST Graphs valuation methodology, not as an expression of current bullish conviction; the creator is demonstrating that a PE compression from 56 to 35 still yielded 65% annualized returns historically, which is pedagogical rather than a positive stance on NVDA as a current investment.
NVDA	positive	Invest Like The Best	bad attribution	The NVIDIA sentiment is expressed by a guest (a hedge fund portfolio manager being interviewed), not by the host/creator of "Invest Like The Best" — the transcript's ">>" markers show a clear host-asks/guest-answers structure, and the bullish NVDA commentary comes entirely from the guest's turn, not the host's own view.
MU	mixed	Invest Like The Best	bad attribution	The Micron commentary is made by a guest interviewee on the "Invest Like The Best" podcast, not by the channel creator/host — the transcript shows a back-and-forth interview format with ">>" speaker switches, and the Micron example is delivered by the responding guest, so the view cannot be attributed to the channel creator.
ASML	positive	Invest Like The Best	stance mismatch	The quote merely states ASML is a current holding ("we're invested in a couple businesses there, Rolls-Royce, uh ASM, ASML") with no forward-looking bullish expression; the surrounding context frames European markets as "just tough right now" with companies there "more challenged," making the actual stance neutral at best — labeling it "positive" overstates what the creator expressed.
MSFT	mixed	SpotGamma	stance mismatch	The creator uses Microsoft exclusively as a pedagogical example to explain hedge wall and gamma strike mechanics — they narrate historical price action without expressing any personal sentiment (bullish, bearish, or mixed) toward MSFT; labeling the stance "mixed" misrepresents an analytical/educational discussion as a sentiment call.
INTC	positive	Monetary Matters	bad attribution, stance mismatch	The quote ("I mean, look at Intel. Intel, you know, there are so many examples already now.") is spoken by a second speaker (indicated by the ">>" prefix in the transcript), not the Monetary Matters channel creator — it is a guest/interviewee agreeing and adding examples, so attribution_is_creator fails; additionally, the mention of Intel is a passing illustrative example supporting a broader basket-of-Trump-touched-stocks thesis, not an expressed positive stance on Intel specifically, so stance_matches also fails.
GOOGL	positive	The Julia La Roche Show	stance mismatch	The creator mentions owning Google but explicitly frames it as an underperformer — "Google's up a little bit more" is offered as the best case in a group that is broadly lagging ("Microsoft's way down, Meta down"), while the creator's actual bullish call is on energy; labeling this stance as "positive" toward GOOGL misrepresents the creator's message, which is neutral-to-negative on big tech relative to energy.
MSFT	negative	The Julia La Roche Show	stance mismatch	The creator mentions "Microsoft's way down" only as a passing comparative aside to illustrate investor rotation into semiconductors/energy — it is a descriptive price-performance observation, not an expressed negative stance on MSFT; labeling it "negative" misrepresents the creator's actual subject (energy bullishness) and overstates what was said about Microsoft.
META	negative	The Julia La Roche Show	stance mismatch	The creator says "Meta down" as a fleeting factual aside in a list of Mag-7 stocks that haven't recovered since November/December — it is a price-level observation used to support his energy thesis, not an expression of negative investment sentiment toward Meta; labeling this as a negative stance on Meta misrepresents the creator's actual intent.
NVDA	positive	Trade Brigade	stance mismatch	The extracted quote characterizes NVIDIA as a "value play" rather than a momentum play — a reframing, not an expression of bullishness — so the quote itself does not support a "positive" stance label; the bullish content lives in surrounding context outside the quote, not in the quoted text.
AAPL	mixed	Trade Brigade	invented number	The quote retains two specific price-level numbers (30850 and 30325) that the creator explicitly presents as short-trigger and long-entry targets for Apple; the numeric scrubber failed to remove them, so the quote surfaces creator-stated price targets in violation of the no_invented_number gate.
GOOGL	mixed	Trade Brigade	invented number	The quote retains specific price levels ("381" as a short trigger, and the truncated "squeeze them straight ba[ck into this 400]") that function as explicit price targets; the numeric scrubber failed to remove these numbers, violating the no_invented_number rule.
COIN	negative	Trade Brigade	stance mismatch	The creator explicitly says they would NOT go out and short Bitcoin names, then speculatively notes "Maybe Coinbase breaking a trend line" — this is a tentative, hedged observation, not an expression of negative sentiment; labeling it "negative" overstates the creator's actual stance, which is neutral to cautious at most.