The June 21, 2026 episode of This Dum Week opens on a genuinely strange note: Alex's week was made briefly intelligent by the Anthropic Mythos/Fable saga, then "made dumb again" by what he calls doctor-driven ultrasound gate — a viral medical establishment freakout over Midjourney (the image-generation company, not the AI model) announcing a new solid-state ultrasound device that can do a full-body scan in about a minute while the patient is submerged in water. The hosts spend the opening fifteen minutes of the episode dissecting why doctors, in numbers, immediately took to social media to declare the technology dangerous and unworkable, with Alex cataloguing the responses with the weary precision of someone who has seen this before: overgeneralized mammography reasoning, legal liability evasion dressed up as epistemology, and condescension that tracks almost perfectly to COVID-era dynamics. RollerGator attempts, heroically and unsuccessfully, to construct a charitable version of the doctors' argument. From there the show plows through a stack of shorter segments — the DoorDash delivery robot that rolled into an Arizona SWAT operation and refused commands to leave; D-list celebrities in legal trouble (an Elf cast member surviving on royalties and living out of his car, rapper Mystikal sentenced to 20 years for rape, Carlos Mencia's $1.2 million federal tax lien); a ropeless bungee jump death in Brazil; a Chicago mob gambling bust with an FBI undercover mole; the AI-powered Bullfrog drone-killing machine; and Cuba's sweeping free market reforms, attributed by Alex to American pressure following recent diplomatic activity including John Ratcliffe's Havana visit. The episode's spine is formed by three substantive segments that arrive in the second half. Tulsi Gabbard's departure as Director of National Intelligence, on Father's Day, June 21st, which is also Juneteenth, generates her exit video — a directed broadside against Anthony Fauci accusing him of funded gain-of-function research at the Wuhan Institute of Virology, manipulating intelligence assessments, lying under oath to Congress in 2024, and using his proximity to the intelligence community to suppress dissenting analysts. RollerGator reads at length from Larry Sanger's (Wikipedia co-founder) Claude-assisted analysis of the four PDFs Gabbard released, identifying two buried bombshells: the IC Inspector General's classified whistleblower complaint — routed by Biden-era DNI Avril Haines not to an independent HHS Inspector General but to Fauci's own boss, Javier Becerra — and what the ODNI briefer's internal email describes as the FBI either withholding information from the IC or providing outright inaccurate testimony to Congress about its COVID origins interview sequence. Alex, characteristically, is not satisfied: he observes that the targets of Gabbard's release are narrowly and conveniently Democratic officials (Becerra, who is now running for California governor), that the release stops short of documenting the deeper intelligence-bioweapons nexus he considers the real story, and that declaring the lab leak hypothesis confirmed — as Gabbard does definitively in her exit video — is asserting more certainty than the evidence actually provides. The episode closes with two interlocking segments on AI, both of which are really about the same question: who controls AI capability and on what basis? The "Okay, Sure, Whatever" UAP segment features the third Pentagon batch of declassified UFO files (72 documents, 1940s to present, increasingly iPhone video rather than military camera), David Grusch's congressional conference testimony about recovered non-human craft and what he calls "sentient plasmoid life," a Chris Cuomo outdoor rant about not knowing what to do with the information, and an ElevenLabs-dramatized exchange between Eric Weinstein and professional UFO debunker Mick West. RollerGator then pivots — via a leak of Peter Thiel's invitation-only "Dialogue" society (222 members including the NATO Supreme Allied Commander, two US senators, Treasury Secretary Scott Bessant, Ted Cruz, and a former Middle East intelligence chief, all registered on personal rather than government email, all storing their data in an unsecured Airtable database) — to what he calls the most important story on the docket: Alex's Twitter thread constructing a complete timeline of the Anthropic Mythos-5/Fable-5 situation. Alex's central revelation is that the "ban" is considerably more nuanced than it appears: Mythos Preview — a different model designation — is still being used by over 200 companies in Project Glasswing and by the NSA itself for offensive cyber operations against adversaries, while the export control order applies only to Mythos 5 and Fable 5; Scott Bessent's otherwise puzzling involvement is explained by his role convening financial infrastructure CEOs to plan for AI-enabled cyberattacks on the banking system; and what Vice President Vance saw in classified briefings apparently flipped him from pro-deregulation to convening AI CEOs and issuing soft-regulation executive orders. The episode closes with Father's Day wishes and a parting academic gift: a preprint paper proving that if LLMs have human-like attributes, then so does Age of Empires II — because someone has implemented a fully functional NAND-gate logic network inside the game using goats as signal carriers. Detailed Outline Opening: Doctor-Driven Ultrasound Gate (00:00:00 - 00:18:30) Main Topic: Midjourney's Full-Body Ultrasound Device and the Medical Establishment's Response Alex's week was made "intelligent" by the Fable/Mythos saga and then "dumb again" by what he calls doctor-driven ultrasound gate Midjourney — the company best known for its Discord-based image generator — announced a medical imaging device: Arrays of thousands of extremely tiny solid-state ultrasound chips arranged in a circular chamber Patient submerges in water for approximately one minute Claims to produce high-fidelity full-body 3D imaging faster and cheaper than MRI Alex's framing: the device is cheap, fast, and pleasant — the opposite of modern medical imaging Medical establishment's reaction: immediate and sweeping dismissal, COVID-era condescension register "More tests bad" as the unstated foundational premise Overgeneralizing from legitimate overdiagnosis concerns (mammography false positive problem) to "any new diagnostic information = bad" One doctor's argument: sound waves cannot penetrate bone — promptly answered with videos and studies on compute-enabled bone penetration, which the doctor then dismissed on procedural grounds Legal liability as actual motivation: "if I'm exposed to the information that you have these masses in your body... I will then be legally liable for not having done anything about them" Alex's parallel to mandated reporters avoiding asking about molestation: "this is why you don't ask children about any potential molestation at home, because if you're a mandated reporter and you get to be informed of it, then you have to do something. And that's like, you might have to spend the weekend filling out forms" RollerGator attempts steelmanning: relearning how to read higher-quality imaging; false-positive pipeline under cost/speed conditions; potential need for years of research to extract value from higher-fidelity data Alex: "I can see your brain frantically trying to compose an argument here, and are unfortunately doing way too good a job because this is not what they're doing" Alex's limiting principle question: if more information is dangerous, what are the odds the current exact level of information is perfect? "The question is not knowing what happens from this device. Yeah, not having any information, in which case carcinoma can still not be ruled out, technically." RollerGator's Twitter exchange: posts asking what other diagnostic information doctors have been refusing; gets response about tech bros not understanding sound wave physics Both hosts note the response is not "if it worked we would use it" but categorically "more tests bad" Resolution: more research needed; possible coverage next week Key Quote: "They're literally saying, I don't want to know. I do not — do not tell me, do not show me this. I don't want to know. There isn't a more sophisticated argument behind this. This is like, if I knew, I would have to order an autopsy, and that could kill you. We literally suck at this, guys. Please don't make us do more." — Alex, summarizing the revealed logic of the doctors' position Notable Detail: At least one doctor acknowledged he misunderstood the device after being corrected about the false-positive pipeline concern — the cheap/fast/repeat-scan feature eliminates the primary incidentaloma risk that underpins the mammography argument. Hosts' Analysis: Both hosts treat this as a direct replay of COVID-era epistemic dynamics: credentialed authority declaring a question settled, refusing engagement with the actual technical question being raised, and defaulting to condescension as an argument. Alex's sharpest observation is structural — the legal liability explanation, offered more openly by some physicians, reveals an institutional incentive that is the actual driver, and it is an incentive not to know rather than an incentive to protect patients. Quick Segments: Celebrities, SWAT Robots, Bungee Deaths, and the Chicago Mob (00:18:30 - 00:54:00) Main Topic: A Stack of Shorter Stories From the Week DoorDash Robot in SWAT Situation A DoorDash delivery robot named Dot rolled into an active weapons-related SWAT operation in Chandler, Arizona Refused commands to leave from Chandler officers; had to be manually retrieved by a DoorDash technician RollerGator's observation: law enforcement regularly uses robots in standoffs — but those robots are invited; Dot was not "Unless someone on the SWAT team ordered a pizza, Dot the B