I can't believe this happened...
Credibility score: 39/100 — Low Credibility. High BS alert! Many claims lack evidence or are misleading.
BSmeter analyzed "I can't believe this happened..." and rated it 39/100 for credibility (a BS score of 61/100 — low credibility), on 2026-06-26. Its weakest claim — "Rhetorical question to justify government action — shifts blame." — scored 20/100 and was flagged as straw man. 18 claims were checked against the video transcript. Scores are produced by BSmeter's AI analysis of the transcript, not independent human verification.
Of 18 claims analyzed: 4 scored under 40, 14 between 40 and 69, and 0 at 70 or above.
Claims analyzed
Claiming "Daario" achieved regulatory capture, preventing access to the Frontier model. — Anonymous Authority (45/100)
At 0:00
Blaming "Daario" for "regulatory capture" without saying who that is or what they actually did. Very specific, very vague 💀
Why this score: The speaker uses a specific name, 'Daario,' and a strong accusation, 'regulatory capture,' but provides no context or evidence to support this claim. It's an assertion of a negative outcome attributed to an unnamed, unverified entity, making it hard to assess or verify.
Original quote: “This is really bad. Daario got his way. He has regulatory capture and now all of us don't get the Frontier model right when it drops.”
Claiming GPT 5.6 is out, but only for a select few, leading to disappointment. — Confidence Mismatch (45/100)
At 0:11
Talking about "GPT 5.6 sole preview" like it's a known thing, but OpenAI hasn't announced a 5.6. That's a bold claim with zero receipts 🚩
Why this score: The speaker confidently refers to 'GPT 5.6' as if it's a released or previewed model, yet there's no public information from OpenAI about a GPT 5.6. This creates a mismatch between the speaker's certainty and the verifiable facts, potentially misleading the audience about the current state of AI model releases.
Original quote: “Here is GPT 5.6 sole preview. Only a select group of companies get to use it right now. And the rest of us, we wait. We wait for the little morsels. And it's I'm I'm so disappointed right now. I really it's hard for me to convey in words what I'm”
Rhetorical question to justify government action — shifts blame. — Straw Man (20/100)
At 0:30
Asking 'what were they supposed to do?' as if there were no other options. That's a classic straw man to justify the government's actions. 🎭
Why this score: By posing the rhetorical question 'what were they supposed to do?', the speaker creates a false dilemma, implying that the government had no choice but to regulate aggressively due to Anthropic's actions. This simplifies a complex issue and deflects potential criticism from the government's role, setting up a straw man argument where the only alternative is inaction.
Original quote: “but what were they supposed to do?”
Declares a 'negative turning point' for AI industry — dramatic language for a prediction. — Loaded Language (45/100)
At 0:30
Calling it a 'negative turning point' is pure drama, not a verifiable fact. It's a vibe, not a metric 🎭
Why this score: The speaker uses emotionally charged language ('negative turning point') to describe a future event, presenting it as a certainty rather than a potential outcome or an opinion. This framing aims to evoke a strong reaction rather than provide objective analysis.
Original quote: “How this is a turning point, a negative turning point for the AI industry.”
Government told Anthropic to restrict access, so they just turned off the model. — Confidence Mismatch (45/100)
At 2:30
Claims the government 'said' to restrict access to non-US citizens, then Anthropic 'just turned it off' — but no source for this specific government directive. 🕵️♂️
Why this score: The speaker states with high confidence that the government issued a direct order to Dario Amodei (Anthropic's CEO) to restrict access to non-US citizens, leading to the model being 'turned off.' While government scrutiny of AI models is real, the specific directive and Anthropic's response as described lack direct, cited evidence within the video. This sounds like a confident interpretation of events without explicitly stating the source of the 'government said' part.
Original quote: “So the government stepped in and they said, "Hey Dario, go ahead and uh don't give that model to any non US citizens." How is anthropic supposed to do that? Well, they just turned it off. We lost access to the model after just a couple days.”
Claims models are 'incredibly dangerous' and 'only a select few should be using them' while also saying they're always hackable. — Volume Game (45/100)
At 2:30
He's quoting a hypothetical government concern about 'dangerous' models, then immediately undercutting it by saying they're 'always' hackable. The danger is real, but the 'select few' part is the real issue. 🤡
Why this score: The speaker sets up a straw man argument by attributing the 'incredibly dangerous' and 'select few' claims to the government, then immediately dismisses the idea of a '100% protected model' as non-existent. This creates a false sense of contradiction, implying the government's concern is invalid because models are inherently hackable, rather than addressing the core issue of who controls powerful AI. It's a classic volume game: loud claim, quiet retraction.
Original quote: “Okay, well, you're telling us these models are incredibly dangerous. You're telling us that only a select few should be using them, but you just released it and now we're getting reports that you can prompt hack the model." By the way, you can always prompt hack a model. There is no such thing as…”
Predicting GPT 5.6 release and calling it 'effectively AI regulation' with no official source. — Confidence Mismatch (45/100)
At 4:30
Calling a limited release 'effectively AI regulation' is a big leap — that's a strong claim for something that's just a company's rollout strategy. 😬
Why this score: The speaker is presenting a company's internal release strategy as a form of 'AI regulation' without any evidence that it's government-mandated or a formal regulatory framework. It's a confident interpretation of a private company's actions as a broader policy.
Original quote: “the release of GPT 5.6. First, it will go to a handful of companies that deserve to use it, that are trustworthy enough to use it. They get to test it, and then maybe hopefully in a few weeks, maybe more, maybe less, we don't actually know, the general public will get it. This is effectively AI…”
Predicts GPT 5.6 release and limited access, then calls it 'effectively AI regulation.' — Confidence Mismatch (45/100)
At 4:30
Goes from 'maybe hopefully' to 'effectively AI regulation' real quick. The leap from speculation to certainty is Olympic-level. 🤸♀️
Why this score: The speaker starts by admitting uncertainty about the timeline ('maybe hopefully in a few weeks, maybe more, maybe less, we don't actually know') but then immediately jumps to a definitive conclusion, stating 'This is effectively AI regulation.' This rapid shift from speculation to confident assertion without new evidence is a confidence mismatch.
Original quote: “the release of GPT 5.6. First, it will go to a handful of companies that deserve to use it, that are trustworthy enough to use it. They get to test it, and then maybe hopefully in a few weeks, maybe more, maybe less, we don't actually know, the general public will get it. This is effectively AI…”
Claims only two companies are at the AI frontier, creating an insurmountable lead for startups. — False Dilemma (20/100)
At 6:30
Paints a picture of an AI duopoly, ignoring the massive open-source community and other major players. It's not just two 🤖
Why this score: The speaker frames the AI landscape as dominated by only two companies (OpenAI and Anthropic), implying an impossible hurdle for startups. This ignores the significant contributions and rapid advancements from open-source AI projects (like Hugging Face, Meta's Llama models) and other large tech companies (Google, Microsoft, Amazon) that also have frontier AI capabilities. It creates a false dilemma that only these two can innovate, which isn't the full picture of the competitive landscape.
Original quote: “release, there are effectively two companies, Open AI and Enthropic that have artificial intelligence at the frontier. And that AI is building the next AI. And now that these companies need to hold back their frontier even longer, they get to build even more with them. And then when they finally…”
Claims open-source is 'behind' and the rich get richer with locked-up AI 🚩 — False Dilemma (20/100)
At 8:30
Paints a picture where open-source is inherently 'behind' and only a few companies benefit, ignoring collaborative innovation. 🙄
Why this score: The speaker presents a false dilemma by suggesting that open-source is inherently 'behind the frontier' and that the benefits of advanced AI will be 'locked up within just a handful of companies.' This overlooks the significant advancements and collaborative power of open-source communities, which often catch up or even surpass proprietary models over time. It frames the situation as a zero-sum game where only a few can win, ignoring the potential for distributed innovation and the eventual democratization of technology.
Original quote: “But open-source is behind the frontier. And this actually might make it worse because although the broad economy won't get the benefits of the absolute frontier because it's going to be locked up within just a handful of companies, those companies are still accelerating. They are still using Fable…”
US slowing AI, China not, creating a 'race' narrative 🚩 — Emotional Button (45/100)
At 8:30
Pushes the 'race with China' button hard, making a 'few weeks' delay sound like a national crisis. The urgency is doing heavy lifting here. 🚨
Why this score: The speaker uses the 'race with China' narrative to amplify the perceived negative impact of a 'minor slowdown' in AI development. This taps into nationalistic fears and competition, making a small delay seem catastrophic, which is a classic emotional appeal to heighten concern without necessarily providing proportional evidence of the actual threat. It's all about making 'every minute counts' feel like a matter of national security.
Original quote: “While the US government is slowing down AI in the United States, you know who's not? China. China's certainly not slowing down. This might seem like a minor slowdown. This might be just a few weeks before the rest of us get the GPT 5.6 model, but guess what? It is a slowdown. Every single minute…”
Aaron Levy claims de facto AI regulation exists, leading to government review of models. — Confidence Mismatch (45/100)
At 10:30
Calling it 'de facto' regulation and then immediately saying it's 'not obvious why' it won't be reviewed. Pick a lane, Aaron! 🤡
Why this score: The speaker quotes Aaron Levy, who states 'we now have de facto AI regulation' but then immediately follows with a speculative question about why models 'won't have to be reviewed by the government.' This creates a contradiction, presenting a strong claim ('de facto regulation') with an immediate lack of clarity or certainty about its implications, suggesting a confidence mismatch in the assertion.
Original quote: “Aaron Levy. We now have de facto AI
regulation. It's not obvious why from
here on out models that have certain
levels of capability or are trained on
certain compute sizes won't have to be
reviewed by the government before
release.”
Claiming hindered monetization will 'completely' affect the US economy and 'entire economy' with no specifics. 🚩 — Confidence Mismatch (45/100)
At 12:30
Went from 'hindered' to 'completely affect the entire economy' real quick. That's a leap of faith, not data. 🤸♀️
Why this score: The speaker makes a very strong, definitive statement about the economic impact ('completely hindered', 'affect the entire economy') without providing any specific data, models, or evidence to back up such a broad and absolute claim. It's a confident assertion without the receipts.
Original quote: “and now their ability to monetize their models that they created is completely hindered and that is going to affect the US economy. It's going to have these second and third order effects that will affect the non-investors, people who are not at the AI frontier lab that's going to affect the entire…”
Claiming hindered monetization will 'completely' affect the US economy, using loaded language. — Loaded Language (45/100)
At 12:30
Saying 'completely hindered' and 'affect the entire economy' is a bit dramatic without specific numbers or timelines. That's some serious economic doom-scrolling. 📉
Why this score: The speaker uses strong, definitive language ('completely hindered', 'entire economy') to describe the impact of monetization issues on the US economy. While there might be an impact, the use of such absolute terms without quantifiable evidence or a timeline exaggerates the potential effect, leaning on emotional impact rather than precise economic analysis.
Original quote: “and now their ability to monetize their models that they created is completely hindered and that is going to affect the US economy. It's going to have these second and third order effects that will affect the non-investors, people who are not at the AI frontier lab that's going to affect the entire…”
Speculates about GPU regulation and power concentration, then states who will control it with certainty. — Confidence Mismatch (45/100)
At 14:30
Goes from 'maybe' to 'I will tell you' about who controls GPUs. That's a leap of faith, not a prediction. 🔮
Why this score: The speaker starts with hypothetical questions about future regulation ('maybe you're going to start locking down GPUs') but then pivots to a definitive statement about who 'will' control access ('It's going to be Daario and company. It's probably going to be Open AI and company. and it's going to be the US government.'). This shift from speculation to certainty without presenting any new evidence or logical steps demonstrates a confidence mismatch.
Original quote: “One interesting question would be how regulation eventually extends to open models, which would have its own set of long-term consequences. Yeah, I mean, it can't. Maybe you're going to start locking down GPUs. Maybe you're going to start saying, "Okay, only certain people can have GPUs." We kind…”
Questions the AGI's capabilities by comparing it to simpler tasks it allegedly can't do, using a False Equivalence. — False Equivalence (20/100)
At 16:30
Comparing 'curing cancer' to 'sniffing illicit distillation' as if they're on the same difficulty scale for an AGI. That's a wild leap in logic 🤡
Why this score: The speaker creates a false equivalence between highly complex, real-world problems like curing cancer and the specific technical challenge of preventing data extraction or 'distillation attacks.' While both are difficult, the nature of the problems and the AGI's capabilities in each domain are not directly comparable in the way presented, making the 'why not just use AGI' argument a rhetorical trick rather than a logical one. It assumes AGI's general intelligence applies equally to all specific, real-world, and often human-driven security challenges.
Original quote: “If you're on the verge of AGI or ASI, artificial general intelligence, artificial super intelligence, why isn't your model smart enough to recognize espionage distillation in real time? You say cure cancer in a few years? Isn't sniffing elicit distillation quite a bit easier than curing cancer? Why…”
Speaker questions AGI's capabilities, highlighting a perceived contradiction. — Confidence Mismatch (45/100)
At 16:30
He's pointing out the 'cure cancer' claim versus the 'can't stop data theft' reality. The confidence in AGI's power doesn't match its reported vulnerabilities. 🤡
Why this score: The speaker is using a rhetorical question to highlight a perceived inconsistency in the claims about AGI's capabilities. If AGI is so advanced it can cure cancer, why can't it handle simpler security issues? This suggests a mismatch between the stated confidence in AGI's potential and its current, or implied, limitations. It's a classic 'if it's so smart, why can't it do X?' argument.
Original quote: “If you're on the verge of AGI or ASI, artificial general intelligence, artificial super intelligence, why isn't your model smart enough to recognize espionage distillation in real time? You say cure cancer in a few years? Isn't sniffing elicit distillation quite a bit easier than curing cancer? Why…”
Pushes open source as the solution, then immediately admits it's not as good. The classic 'but maybe one day' cop-out. 🤡 — Volume Game (45/100)
At 18:30
Tells everyone to go support open source, then immediately undercuts it by saying it's not as good. That's a quick pivot. 📉
Why this score: The speaker strongly advocates for supporting open-source models as a solution, but then immediately walks back the enthusiasm by admitting they are currently inferior to proprietary models like Fable or GPT 5.6. The 'maybe one day' is a quiet retraction after a loud call to action, creating a volume game.
Original quote: “So there it is. We no longer get the best models. Some do, not us. So there it is. I don't know what to tell you at this point to be honest. I think there are two things. One, go support open source like you never have before. Go test open source models. Download them and run them locally. Figure…”
See the full analysis with timestamps →