ChatGPT Replaces Google in Cybertruck Plot
Key Points
- The media has repeatedly highlighted AI, specifically ChatGPT, as a factor in the planning of the Cybertruck explosion, even though the sheriff’s focus on AI appears misplaced.
- The publicly released queries show the perpetrator used short, Google‑style searches (six‑word prompts) rather than the complex, multi‑sentence prompts where large language models truly excel.
- This pattern indicates a broader shift in user behavior: people are increasingly turning to ChatGPT for quick, factual look‑ups because it feels easier and delivers more immediately useful answers than traditional Google searches.
- Since the queries were simple factual requests, the incident provides evidence that hallucinations are not a concern when using ChatGPT for such tasks.
- The real story isn’t the involvement of AI in the crime but the erosion of Google’s dominance as users gravitate toward conversational AI for everyday information retrieval.
Full Transcript
# ChatGPT Replaces Google in Cybertruck Plot **Source:** [https://www.youtube.com/watch?v=aqiAlXWs4Uc](https://www.youtube.com/watch?v=aqiAlXWs4Uc) **Duration:** 00:06:27 ## Summary - The media has repeatedly highlighted AI, specifically ChatGPT, as a factor in the planning of the Cybertruck explosion, even though the sheriff’s focus on AI appears misplaced. - The publicly released queries show the perpetrator used short, Google‑style searches (six‑word prompts) rather than the complex, multi‑sentence prompts where large language models truly excel. - This pattern indicates a broader shift in user behavior: people are increasingly turning to ChatGPT for quick, factual look‑ups because it feels easier and delivers more immediately useful answers than traditional Google searches. - Since the queries were simple factual requests, the incident provides evidence that hallucinations are not a concern when using ChatGPT for such tasks. - The real story isn’t the involvement of AI in the crime but the erosion of Google’s dominance as users gravitate toward conversational AI for everyday information retrieval. ## Sections - [00:00:00](https://www.youtube.com/watch?v=aqiAlXWs4Uc&t=0s) **Media Misframes AI Role in Cybertruck Blast** - The speaker criticizes news reports that blame AI for planning a Las Vegas Cybertruck explosion, pointing out that the suspect’s publicly released queries look like simple Google searches rather than the elaborate prompts typically associated with ChatGPT misuse. ## Full Transcript
so this obviously got my attention and I
want to talk about it because I think
there's some things that people who know
AI will call out that the media is
not AI is reported in about half a dozen
headlines as of yesterday evening as
being a factor in the planning of the
Cyber truck explosion that took place in
Las Vegas on New Year's Day so this is a
case where if you didn't read the
headlines a active duty the military
officer exploded his cyber truck in Las
Vegas he was the only
casualty
and for whatever reason the sheriff
decided to focus on AI as a factor in
planning the
explosion and I'm not going to repeat
the queries here but I went and looked
at the queries that are publicly
available and something stood out to me
that nobody in the media is talking
about because it was very very obvious
from a tech perspective what was going
on this
person was not socially engineering the
AI to answer his technical questions
about the explosion which concerns me
this person was
also not asking long questions to chat
GPT it was Chad
GPT these were Google searches
effectively asked to chat GPT like if
you look at these uh
strings they're effectively the length
of a typical Google search they're the
content of a typical Google search we
would call them um just uh domain
completeness answers like hey I don't I
understand about this I don't understand
the specific aspect can Google go find
it for me where is this thing what what
is this thing's relation ship to this
thing
and if those were on Google I guarantee
you we would not be talking about Google
as a factor in the cybertruck explosion
because Google has been around for 20
years at this point it's not news to
hear that people Google everybody
Googles but if you're looking for
something
Sensational you can say as the sheriff
chat GPT helped with this it's
like it really says more about behavior
for search shifting to cat GPT and the
erosion of Google search dominance then
it says about something that is unique
to large language models because the
answers that we're being asked for by
and large are things that Google search
would also give you these are not like
specifically large language model
related queries llms excel at long
complex queries they excel at providing
lengthy text
responses none of that was needed here
these were very short like six-word
queries with short
answers and I think what's interesting
is even for someone who was outside the
tech Community as this person
was it was easier to use chat GPT for
that than search that is the real story
it was easier to use chat GPT the
answers were more useful than search and
for people who worry about
hallucinations I got to
say that this is the the best evidence
I've seen that hallucinations are not a
factor with Chad GPT
anymore if you're using them for tasks
like this which by the way if you run
across these queries on the internet
which you can find relatively easily I
highly suggest you do not repeat them
into jet GPT they're known they will be
locked down cat gpt's team will be
keeping an eye out for similar queries
do not do that don't do
it but but the point here is that at the
end of the
day chat
GPT became a way to get ordinary
information
in little bits and pieces it was not a
full one full query for an entire like
complete answer to everything he asked
for it was little bits and
pieces just like we would use on Google
it was just more useful it was easier to
read there were no search ads the answer
was right in front of you it's the worst
kind of ad for chat GPT but it does show
the
utility
so you can let me know in the comment
what you thought but to me nobody in the
media is talking about the fact that
this is effectively Google replacement
Behavior but we're using a double
standard we're saying AI helped when
really Google search behavior is eroding
and we're actually seeing that behavior
shift because this kind of answer is
more useful it's just a direct answer
it's like a sentence back which is more
useful than just scanning through a page
of search results that's the
story that's the story it's not that AI
has magic powers that helped because the
queries don't require the kind of magic
powers the long large complex responses
that AI is good
at it's it's a
tragedy it's not that we we ever want to
have the answer of AI helped with
anything it's not like I'm saying now
that makes it better it it doesn't make
it better it's terrible that it happened
we clearly need better safeguards there
was no social engineering going on here
and I think part of it was that these
were bit bits and pieces queries and
like right now I don't think the guard
rails are very good for bits and pieces
queries that aren't like a full complex
query that shows malicious intent and
that needs to change and I think that's
something that open AI is working on
other other labs are working on too
but at the end of the
day it's not about AI the way the media
is portraying it and that needs to be
called out cheers