Hank Green vs. Luis Elizondo, absent summaries, and the AI rabbit hole

jdog

Senior Member.
YouTube science communicator Hank Green stirred up a hornet's nest on Bluesky this week at https://bsky.app/profile/hankgreen.bsky.social/post/3lfdibahtsk26 by posting the cover of Luis Elizondo's book "Imminent" with a Gimbal photo and declaring it "Just so, so ballsy to put an image of something that we 100% know is the heat signature of an airplane on the cover of your book about UFOs." This started the predictable flame war in the responses as well as a Green-bashing thread on /r/ufos.
2025-01-10_16-42-05.jpg

In a later Bluesky post, Green backed off the assertion, but brings up Mick West's analysis of the rotation: "Look, I apologize for being hyperbolic and also dismissive...but I have never seen anyone make a convincing case that Mick West is wrong here. They ad hominem him, they say that he doesn't understand things...but his case is simple and well-supported. The object doesn't rotate, the gimbal does."

I didn't want to attach my professional account to the issue, so I didn't comment there, but it struck me how so much of the defense of the Gimbal object as anomalous depended on the belief in corroborating evidence (eyewitness accounts, multi-sensor recordings) that doesn't exist, as well as confusion between Ryan Graves and David Fravor and the Gimbal and Tic Tac incidents. I also didn't want to fall into making imprecise claims myself.

I know the cases have been thoroughly discussed in threads here, but I couldn't recall if the Gimbal object had been seen on radar before the sighting. And these threads can be... a lot. I wanted a summary of the current understanding. Wikipedia was less than helpful. When I searched "was gimbal object on radar" I got this dismaying "AI Overview" answer from Google:
1736556829004.png

The main source for this conclusion seems to be an AOL version of a New York Post story about the Nov. 18, 2024 oversight hearing of AARO. (For some reason the author describes the still images as coming from "infrared radar" and "FLIR radar.") Google also links to a Live Science blog post summarizing the 2019 New York Times story about the three famous Navy UFO videos -- none of which actually say the Gimbal pilots saw their target on radar, let alone "with their own eyes." (The secondary answers for AI Overview were worse and the regular search results below were barely relevant.)

For comparison, I tried the Perplexity AI system. I got a much more nuanced response that I admit seems to fit my prior expectations: "The GIMBAL UFO object was not definitively confirmed to have been seen on radar. While radar detections were reported in some UFO incidents, the search results do not specifically mention radar confirmation for the GIMBAL video." Though its sources turn out to be one 2020 CBC interview with Mick and a blogger's 2021 summary of Mick's "Explained: New Navy UFO Videos." Not particularly definitive. :rolleyes:

I read some transcript where Graves more recently claimed to have seen the Gimbal object on the aircraft carrier's radar, but it's not clear if he saw the specific target or just the "fleet of objects" radar returns at the same time and concluded that one of them was what the pilots were looking at. (Given the need for Mick and Peings & von Rennenkampff to try to (separately) recreate the potential flight path and distance of the Gimble object's location relative to the F/A-18 recording it, it seems more likely no one knows definitively that the carrier had it on radar.)

I digress, but it's interesting at the start of 2025:
  • How Hank Green slid into claiming 100% certainty without some resource to point to.
  • How many believers imagine there's corroborating evidence of the Gimbal target's nature.
  • How badly the Google AI Overview system mangles its presentation of even simple facts.
  • How hard it remains to find relatively definitive summaries of these cases.
 
How Hank Green slid into claiming 100% certainty without some resource to point to.
I overall didn't like the tweet (or however BlueSky posts are called) because big dismissive statements generally just strengthen the conspiracy. But I assume he was just in a frustrated mood seeing UFO conspiracies around and then the absurdity of Elizondo using Gimbal after their repeated dismissal of the glare hypothesis, so he decided to lash out and mock it. I wish he hadn't done it but I understand why he did it since I've felt the same frustration reading misinformation myself.

How badly the Google AI Overview system mangles its presentation of even simple facts.
I think AI struggles already with well-established truths and should be taken with a gigantic grain of salt for anything you actually care to know the truth about. It does not surprise me at all that it struggles with more muddy terms like UFO cases where most of the references an AI would find are extremely vague and filled with speculation.

How hard it remains to find relatively definitive summaries of these cases.
I agree with this, even on Metabunk, I was trying to find some stuff about Gimbal and I found myself needing to jump through multiple threads and still not being sure if there was a part of the conversation I missed because it happened somewhere else or if something was proven/disproven in another thread. Maybe some Megathreads for big enough stuff can be created to have easily accessible links to relevant threads about them, though they do have the disadvantage of needing to be updated when new conversations come up.
 
How badly the Google AI Overview system mangles its presentation of even simple facts.

The major free-access AI systems have a poor accuracy record, and shouldn't be relied upon for serious purposes.
It's a shame that the major providers don't state this each time a query is made.

Saying that, this week Mark Zuckerberg has decided to effectively end fact-checking on Facebook, and nine days ago Elon Musk, owner of X, tweeted "Free Tommy Robinson". Robinson (one of several names he uses/ has used) is an Anglo-Irish extreme-right troublemaker with convictions for assault occasioning actual bodily harm to a police officer, assault with intent to resist arrest, an assault conviction against a member of his own extremist group, possession of cocaine with intent to supply, mortgage fraud, entering the USA illegally on a false passport, and stalking (amongst other things). https://en.wikipedia.org/wiki/Tommy_Robinson
So January 2025 has not reinforced my high opinion of tech mogul's commitments to truth or reason (Here endeth the rant).

Apple Intelligence managed to mangle news stories received from the BBC a few days ago, including a surprise for Rafael Nadal:

External Quote:

A news summary from Apple falsely claimed darts player Luke Littler had won the PDC World Championship - before he even played in the final.
The incorrect summary was written by artificial intelligence (AI) and is based on a BBC story about Littler winning the tournament semi-final on Thursday night.
Within hours on Friday, another AI notification summary falsely told some BBC Sport app users that Tennis great Rafael Nadal had come out as gay.
Apple, which has been contacted for a response, previously declined to comment on similar examples of false AI alerts about news stories.
"Apple AI alert falsely claimed Luke Littler had already won darts final", BBC News, Technology, 03 January 2025, Imran Rahman-Jones https://www.bbc.co.uk/news/articles/cx27zwp7jpxo

Capture.JPG


External Quote:

A BBC spokesperson said: "It is essential that Apple fixes this problem urgently - as this has happened multiple times.
... ...
The BBC previously complained to Apple about the Apple Intelligence feature when it generated a false headline about a high-profile alleged murder in the US.
 
Last edited by a moderator:
I digress, but it's interesting at the start of 2025:
  • How Hank Green slid into claiming 100% certainty without some resource to point to.
I'm 100% certain it's the heat signature of an airplane. Analysis by @Edward Current showed the data matches an aircraft cruising at 30 nm distance. It's the only sensible hypothesis we have that explains the video. Is it hyperbolic to put the certainty at 100%? Maybe a little.
  • How many believers imagine there's corroborating evidence of the Gimbal target's nature.
They follow the witnesses. The witnesses expected a group of radar targets, saw GIMBAL, and then assumed it was one of the objects they were looking for. It's a human reaction, but fallible.
  • How badly the Google AI Overview system mangles its presentation of even simple facts.
LLMs are like the students who cram off other people 10 minutes before the assignment is due, they do not understand anything, and when they're fed contradictory information, they may have a problem generating a coherent summary.
  • How hard it remains to find relatively definitive summaries of these cases.
I thought Mick's videos, and his OPs on the relevant threads here, work fairly well. It's a shame if they're hard to find.
 
Very true, unfortunately HI (human intelligence) very often struggles the same
(A quick confession, I noticed my post #3 had been edited by a mod- I had claimed Mark Wahlberg had ended fact-checking on Facebook...)
 
I agree with this, even on Metabunk, I was trying to find some stuff about Gimbal and I found myself needing to jump through multiple threads and still not being sure if there was a part of the conversation I missed because it happened somewhere else or if something was proven/disproven in another thread. Maybe some Megathreads for big enough stuff can be created to have easily accessible links to relevant threads about them, though they do have the disadvantage of needing to be updated when new conversations come up.
Metabunk needs its own wiki for summarizing cases.

Hank Green is awesome and he made a simple mistake in this case. Would love to hear a conversation between Hank and Mick on UFOs, drones, and skepticism overall.
 
I think it is a great idea, but does anybody have the time/energy/know-how to do it, and keep it up to date?
 
I think it is a great idea, but does anybody have the time/energy/know-how to do it, and keep it up to date?
There are three aspects to this:

1) sysadmin: set up a Mediawiki server, configure extensions, determine if the xenforo user database can be accessed to prevent account spoofing

2) bureaucrat: administer the wiki editorially, i.e. set & enforce policies, support editors, recruit administrators, design the content structure. Basically, support the group effort of wiki editing.

3) legal: determine a copyright policy for the wiki (everything needs to be editable), and figure out if/how we can copy content from the forum, i.e. how can we set it up so the existing rights cover it, and how can forum members give explicit permission (e.g. via their profile)

I have been a bureaucrat for a mid-sized wiki before.

(And this is absolutely the wrong thread in the wrong subforum.)
 
I think it is a great idea, but does anybody have the time/energy/know-how to do it, and keep it up to date?
I was hoping it could just be added as a new section of metabunk itself, rather than a separate website. That would hopefully be easier for Mick to setup and maintain, while also allowing us to use our current accounts. Looks like there are some 3rd party add-ons for XenForo platforms such as xencarta.

(And this is absolutely the wrong thread in the wrong subforum.)
You're definitely right on this. Maybe @Mick West or @Landru can move the wiki comments to a new thread under Site Feedback & News, please! (I think only admins can do that)
 
Last edited:
Back
Top