Latest / Mini Scroll / Drag Race AI film scandal, Anthropic restricts its latest model + the rise of “dopamine sites”

Transcript

Drag Race AI film scandal, Anthropic restricts its latest model + the rise of “dopamine sites” · Mini Scroll

• 0:00
  Coming up today on. Mini tech scroll the.
• 0:02
  AI allegations around RuPaul's new film, the US government
• 0:06
  shuts down anthropic's latest model and the dopamine websites
• 0:10
  going viral in Korea. Hey everyone, and welcome to
• 0:13
  Mini Scroll, the daily Internet culture podcast bringing you the
• 0:16
  biggest stories from social media, the creator economy and
• 0:19
  the digital space every Monday through Thursday and sometimes
• 0:22
  Fridays. My name is Annabelle and I'm
• 0:24
  your Wednesday Mini Tech Scroll host and also.
• 0:27
  The deputy editor of. Australian youth digital media
• 0:30
  grand Centennial world welcome back to mini tech scroll for
• 0:34
  another week and. Also a special.
• 0:36
  Welcome to any new listeners who just joined us from TIK.
• 0:40
  T.O.K, I thought. Since we have had some new
• 0:43
  listeners start following the. Pod that I should.
• 0:45
  Introduce mini tech scroll. Because it is a little bit.
• 0:49
  Different to mini scroll, but I like to think that they
• 0:52
  complement each other nicely so. In mini tech scroll.
• 0:55
  Episodes We cover social media platforms and everything kind of
• 0:59
  going on in the tech space and often, unfortunately.
• 1:03
  It is AI related but we do feel like.
• 1:07
  An episode like this. Weekly is really valuable.
• 1:10
  Because I feel like sometimes when we're talking about
• 1:12
  Internet culture and the creator economy, we don't give enough
• 1:15
  time slash give enough information about these kind of
• 1:20
  themes. Because at the end of the day,
• 1:23
  the platforms and the people in charge of these platforms are
• 1:26
  often playing a key role in what goes viral and why we care about
• 1:29
  what we do online. So we try and kind of fill that
• 1:33
  gap with mini tech scroll once a week.
• 1:35
  So that is my elevator pitch. Why you should?
• 1:38
  Listen, it is definitely a different vibe.
• 1:40
  To mini scroll but I. Hope it brings something else to
• 1:43
  the network also before we get into the.
• 1:46
  Stories today I feel. I have to acknowledge my hair.
• 1:48
  I feel like my hair is looking a.
• 1:50
  Little bit crazy today. My colleague girls will
• 1:53
  understand the plight of trying to put your hair half up, half
• 1:57
  down and just not getting it right and then the more you
• 2:00
  touch it. The frizzier.
• 2:01
  It gets. So that's kind of what happened
• 2:02
  to me this morning and then I just had to leave.
• 2:04
  So here we are, and I just felt the need to.
• 2:07
  Preface that but. Also remember to rate, review
• 2:10
  and subscribe and star rate US on Spotify.
• 2:13
  All the things. But yes, let's get into the
• 2:16
  stories for. This week.
• 2:17
  The first story I wanted to talk about.
• 2:18
  Today is pop culture. I.
• 2:20
  Will say that, but. I wanted to break down all the
• 2:23
  AI allegations levelled against the new movie Stop That Train.
• 2:26
  So. I will say this.
• 2:27
  Is a little bit of me holding the audience hostage to talk
• 2:30
  about what I want to talk about, but it is relevant, I think.
• 2:34
  So Stop That Train is a new film coming out of.
• 2:37
  The RuPaul's Drag Race. Cinematic universe.
• 2:40
  We really cover reality TV beyond a mention on infinite
• 2:44
  scroll or mini scroll, but I do. Watch Drag Race.
• 2:47
  So I have been seeing this saga kind of bubble across my social
• 2:51
  media and I thought I may as well just include it in this
• 2:53
  week's episode to mix things up. But.
• 2:55
  For those of you. Who haven't heard of Stop That
• 2:57
  Train? It is an action comedy disaster.
• 3:00
  Film that stars RuPaul. Alongside some of the most
• 3:03
  famous Drag Race Queens, including Queens like Ginger
• 3:06
  Minge, Jujubee Simone Monet, Exchange and more.
• 3:10
  I know if you don't know RuPaul's.
• 3:11
  Drag Race you probably. Have no idea who I'm talking
• 3:13
  about, but believe me when I say they are pretty iconic in the
• 3:18
  space. The film has also been described
• 3:20
  as a parody of kind of classic disaster films like Aeroplane.
• 3:24
  And I should add here that I actually haven't seen the film
• 3:26
  at the time of recording, but I am just reporting on kind of the
• 3:29
  discourse around it. Many Drag Race fans, since it is
• 3:33
  a very culty fandom, were obviously excited that the
• 3:36
  franchise was expanding and moving into film.
• 3:41
  But that sentiment has kind of shifted with many viewers
• 3:44
  questioning the use of AI in the movie.
• 3:47
  So this really picked up when a filmmaker named Gloria Book
• 3:50
  posted a review on Letterboxed speculating that the film had.
• 3:54
  Used AI as her review reads quote.
• 3:56
  If the film wasn't bad enough on its own, it's one of the most
• 3:59
  conspicuous. Uses of AI I've.
• 4:01
  Seen in a film with a lot of. Visual effects.
• 4:03
  Looking like Gen. AI and doubt about how much of
• 4:06
  the. Obvious stock.
• 4:07
  Footage might also be confirmed in the end credits that Acme.
• 4:11
  AI was the top build visual effects.
• 4:13
  Company on the film so from here the speculation really
• 4:17
  snowballed and to begin with, this kind of claim wasn't
• 4:21
  actually that surprising given that in the recent seasons of.
• 4:25
  Drag races, actually. Seemed to be an.
• 4:27
  Uptick of AI. Usage, especially in the most
• 4:31
  recent season where it seemed that the production included AI
• 4:35
  generated portraits of the Queens.
• 4:37
  All that to say. After this letterbox.
• 4:39
  Review Gloria did go on to create TikTok videos backing up
• 4:43
  her claims, as well as posting various tweets and threads on X
• 4:46
  documenting. Exactly what she believed.
• 4:49
  To be AI in the movie and in one of her posts she points to.
• 4:54
  Various exterior shots. Of the train saying that it
• 4:57
  just. Felt off.
• 4:58
  To kind of back up her instinct, she notes that according to
• 5:02
  IMDb, no computer. Graphics artists are credited.
• 5:06
  While compositors are. This she says quote.
• 5:10
  Means no one is credited for the creation of the assets, only
• 5:13
  combining them. Of course, the list on IMDb
• 5:17
  might necessarily. Be the full credits list.
• 5:19
  But it's obviously something worth considering.
• 5:22
  Another thing Gloria notes is the inconsistencies from the
• 5:25
  outside of the train in different shots.
• 5:27
  So she shows how there seems to be a different number of windows
• 5:31
  in these shots and some of them there are like 8 windows and
• 5:34
  others. There are three or. 6 And with
• 5:37
  this she notes that when using computer graphic models, there
• 5:40
  would be no reason for these differences.
• 5:43
  My name is Gloria. Cook I'm a film maker and
• 5:45
  visual. Effects artist and last.
• 5:46
  Week, I tuned the New York premiere of RuPaul's new film
• 5:49
  Stop That Train and wrote a review, which grew up on social
• 5:51
  media for speculating that the film used AI based on what I saw
• 5:54
  on screen and the film's credits, including Acme AI and
• 5:57
  FX, the company known for using generative AI to cut costs.
• 6:01
  If Adam Shankman wants to make me look like a huge idiot, all
• 6:03
  he has to do are post wireframes of these scenes to show off the
• 6:06
  3D models. It's that easy.
• 6:08
  But I don't think he will, because I don't think they
• 6:10
  exist. And I'm going to make a big deal
• 6:11
  about this because this is going to be a big deal with Hollywood
• 6:14
  moving forward. Film makers are going to lie
• 6:17
  about not using AI, and when they get caught, they're going
• 6:19
  to claim it wasn't generative AI.
• 6:21
  And that'll look good when news sources are able to publish the
• 6:23
  headlines saying that like it's a fact.
• 6:25
  As if the film makers aren't buying ads on those websites to
• 6:28
  promote the film. And it will all be swept under
• 6:30
  the rug as they try to obfuscate the process by assuming the
• 6:33
  public doesn't understand visual effects.
• 6:35
  I do though. I'm a visual effects artist and
• 6:39
  when they claim this technology doesn't take away jobs from
• 6:41
  human artists, they're lying. I'm not saying human visual
• 6:44
  effects artist didn't work on this film.
• 6:46
  They did. But if they did use AI as all
• 6:48
  evidence seems to point to, and they got around meeting 3D
• 6:51
  modelers or animators for a cheaper workflow, a cheapness of
• 6:54
  the film makers are proud of when they talk about the film
• 6:56
  and how quickly and cheaply they were able to make it.
• 6:58
  No one has yet reached out to me, either from the production
• 7:01
  or from the many new sources that are discussing my review
• 7:03
  and Twitter thread exposing this.
• 7:05
  I'm right here. I have nothing to hide.
• 7:07
  I'm a trans woman in a country that wants to get rid of me, an
• 7:10
  unemployed visual effects artist in an industry that sees me as a
• 7:13
  corner to cut. And I have nothing to lose and
• 7:15
  all the time to keep talking about this.
• 7:17
  After all of this started. Gaining traction, the film's
• 7:20
  director Adam Shankman took to Instagram to try and bring some
• 7:23
  clarity to the situation. He issued a statement that read
• 7:27
  quote every shot in Stop That Train was made by human hands.
• 7:31
  It's come to my attention that there is some online speculation
• 7:34
  that Stop That Train is. Full of fully.
• 7:36
  Generated AI shots and I'm here to tell you.
• 7:39
  That that's. Patently not true.
• 7:40
  There are some total of 0 shots conceived by AI in the movie.
• 7:45
  We employed hundreds. Of visual effects sex artists.
• 7:48
  Who all killed themselves getting this out for release and
• 7:50
  not one job was taken out of human hands.
• 7:53
  The movie is a fully human made piece of joy and laughter.
• 7:57
  The response to this statement has been mixed.
• 7:59
  Obviously some social media. Users are happy that it was
• 8:01
  addressed, but others had. Taken issue with the quote UN
• 8:04
  quote conceived by AI wording, questioning whether it was a way
• 8:09
  of getting around admitting that some AI was used in the film, as
• 8:13
  one user commented under the statement quote.
• 8:17
  No shots. Were 100% done by AI.
• 8:19
  But what about shots where generative?
• 8:21
  AI did 80. Percent of the job 50 percent 25
• 8:25
  The only acceptable amount of Gen.
• 8:26
  AI. Is 0.
• 8:27
  Other social media. Users felt that he.
• 8:29
  Could have done more to quell the rumors around Acme.
• 8:33
  AI, which was the. Leading visual effects company
• 8:36
  in the production for the movie and obviously while he didn't
• 8:39
  talk. About what they did.
• 8:40
  Specifically, according to. Variety quote a source familiar
• 8:43
  with the. Production told Variety that
• 8:45
  Acme AI was contracted. Exclusively for its visual.
• 8:49
  Effects work and any AI use was limited to background workflow
• 8:53
  processes and not anything that appeared on screen so.
• 8:57
  In response to that. Social media users asked him to.
• 9:00
  Clarify with these visual effects artists whether they.
• 9:03
  Used AI maybe without his knowledge, so with.
• 9:06
  All that it's clear that those associated with the.
• 9:09
  Film are denying the claims of AI.
• 9:12
  Usage, but viewers. Including Gloria did remain
• 9:15
  skeptical after the statement and last week it seems that they
• 9:18
  have been vindicated because in an interview with Adam Shankman
• 9:22
  that was published by Extra magazine, which is a.
• 9:24
  Digital LGBTQIA. Plus magazine slash like
• 9:27
  platform. He did admit that some AI was
• 9:30
  used in the movie, as his response reads quote some of the
• 9:35
  windows. Were AI there was.
• 9:36
  Only one AI shot. Or two AI.
• 9:39
  Shots of the train the. Rest are computer graphics.
• 9:42
  Most of the computer graphic is the train, then out the windows
• 9:46
  is mostly AI. But the Florida.
• 9:48
  Station all that that CG and different licensed imagery, the
• 9:52
  film's. Distributor Bleecker St.
• 9:54
  Also issued a statement to the publication offering more detail
• 9:57
  and specifying the. Use of AI so I'll also read.
• 10:00
  That statement now quote AI. Tools were used by the visual
• 10:04
  effects. Team to assist in three moments
• 10:06
  to adjust the angle of the live action footage outside the
• 10:09
  windows to match the angle of the camera in one shot and to
• 10:13
  format the CGI crafted train onto 2 frames of live action
• 10:17
  footage exteriors traditional. CGI was used.
• 10:20
  Including for the train and the Stormaganza and hundreds of
• 10:23
  visual. Effects artists were.
• 10:25
  Hired and paid for their work on the film, Gloria has since made
• 10:28
  a TikTok in reaction to this article, and I will.
• 10:31
  Play part of. It for you.
• 10:33
  Now there's a. Specificity here that might make
• 10:35
  people think it must be the full truth.
• 10:38
  I'm skeptical, though it feels like perhaps an attempt to
• 10:40
  concede on some small points because because they know that a
• 10:43
  lot of people have seen the visual evidence in my last video
• 10:46
  and know about Acme AI being in the credits, and so it's not
• 10:49
  worth it to keep pretending there's no generative AI and
• 10:52
  admitting but downplaying can give a plausible way to deflect.
• 10:55
  They're conceding to my bigger point, but they're not actually
• 10:57
  addressing the evidence or giving explanations that would
• 11:00
  justify any of that. But despite how little they're
• 11:02
  admitting to, this update is still big news to me.
• 11:05
  I feel finally, officially vindicated for continuing down
• 11:09
  this rabbit hole, even when the official sources said I was
• 11:12
  wrong. But mostly, I just hope news
• 11:14
  outlets make as big a deal of Adam Shankman admitting to using
• 11:17
  AI after previously denying it, as they made of Shankman's
• 11:20
  original denial. Maybe I'm just inflating my own
• 11:22
  ego, but I don't think Shankman would have said this if I hadn't
• 11:25
  posted my last video. And that's kind of scary, isn't
• 11:28
  it? He lied to you.
• 11:30
  He tried to get. People who wouldn't.
• 11:31
  Want to support AI art to give him money by lying.
• 11:34
  Creators are going to lie and in a lot of cases they will get
• 11:37
  away with it. I'm not really sure what we can
• 11:39
  do about that so I'm excited to update.
• 11:42
  Y'all. I was right, but I'm mostly just
• 11:45
  worried about the future. So obviously.
• 11:47
  This is a very important story in times of.
• 11:49
  How AI will potentially. Be used in filmmaking and how.
• 11:53
  It will be talked about to the public, but the reason?
• 11:57
  Why I wanted? To cover this story is because.
• 12:01
  There have been continuous calls to boycott the movie.
• 12:04
  For its use of. AI online like across X and
• 12:07
  Reddit people have been pointing out that this is kind of a.
• 12:10
  Double edged although which I thought was.
• 12:12
  Quite an interesting point of discussion because choosing to
• 12:16
  boycott this film because. It's AI usage is also.
• 12:19
  Kind of choosing to undermine a queer film at a time when the
• 12:23
  world has become a lot less hospitable and welcoming to this
• 12:27
  community. And I do think what makes.
• 12:29
  All of this difficult to. Kind of reckon with is because
• 12:32
  there just isn't enough kind of clear industry standards and
• 12:36
  disclosure norms. About AI use.
• 12:38
  In films, because obviously Adam Shankman's initial denial was
• 12:42
  like definitely vague, saying that zero shots were conceived
• 12:46
  by AI, quote UN quote, which is obviously very different from
• 12:48
  saying AI wasn't used at all. That distinction between like
• 12:52
  fully generative AI and AI. AI assisted visual effects.
• 12:56
  Is something that the industry just like doesn't have clear
• 12:59
  language around yet? But as Gloria.
• 13:02
  Did put it in her video, she questioned whether.
• 13:04
  This would have had. Even come to light if she hadn't
• 13:06
  even said anything in this kind of started picking up traction
• 13:09
  online, which is even more concerning because this isn't
• 13:12
  going to happen with every single film release.
• 13:15
  But yeah, I feel. Like that was a bit of a ramble
• 13:17
  but I am curious to. Hear your thoughts about.
• 13:19
  All of this and like the boycotting of a film, is this
• 13:22
  deserving of nuance? But I guess nothing online these
• 13:26
  days can be nuanced. My next story is a dramatic
• 13:30
  change of pace. But we.
• 13:31
  Need to talk? About what is probably.
• 13:33
  The biggest story in tech this week, and that is the latest
• 13:35
  back and forth between the US government and Anthropic.
• 13:38
  OK, so if any of you use Claude, you may have noticed that.
• 13:42
  There's like a pop up above the chat.
• 13:44
  Box that currently reads. Quote Claude Fable. 5 is
• 13:47
  currently unavailable. Claude Fable 5 was rolled out to
• 13:51
  the public. Last week and was widely billed
• 13:53
  as Anthropic's most powerful model yet, with advanced
• 13:57
  capabilities when it comes to like reasoning, coding,
• 14:00
  cybersecurity, and just about everything else.
• 14:02
  So Fable is essentially another version of Mythos.
• 14:06
  If you run in any tech cycles online you have definitely heard
• 14:09
  of Mythos. Although ironically, I don't
• 14:12
  think I've actually covered it on the pod.
• 14:14
  Before, but it has. Been making the rounds.
• 14:16
  For a while so. Let me explain.
• 14:18
  Basically, Anthropic has been saying for literal months that
• 14:21
  its latest model mythos. Was too powerful to be released
• 14:25
  to the public. Part of this was.
• 14:27
  Because mythos. Was so highly skilled at tasks.
• 14:30
  Relating to cybersecurity and hacking, which obviously in the
• 14:34
  wrong hands could be extremely dangerous to kind of combat
• 14:37
  this, Anthropic ended up launching this initiative called
• 14:40
  Project Glass Wing, which allowed select partners to use
• 14:44
  Mythos. Preview I will.
• 14:46
  Say back in April when I think this.
• 14:48
  All started there was. Quite a bit of panic around
• 14:51
  Mythos, and I did consider covering that on the part of the
• 14:53
  time, but many social media users, rightfully or not.
• 14:57
  Believed that this. Panic wasn't actually entirely
• 15:00
  warranted and was just another marketing ploy by Anthropic to
• 15:04
  drum up publicity. And it did work if that was the.
• 15:07
  Case anyway Fable 5. Was basically Anthropic's answer
• 15:11
  to the threats that Mythos posed.
• 15:13
  It's being positioned as a version of Mythos that was safe
• 15:17
  enough to use. By the public.
• 15:19
  And alongside the launch of Fable last week, Anthropic also
• 15:22
  launched Clawed Mythos 5 to a quote small.
• 15:26
  Group of cyber. Defenders and infrastructure
• 15:28
  providers as an upgrade of Claude Mithos preview and so it
• 15:32
  is the same underlying model. As Fable 5, but it has.
• 15:36
  Safeguards lifted in some areas. This launch, while very exciting
• 15:41
  for some, did not grow smoothly because three days later,
• 15:44
  anthropic. Was forced to.
• 15:45
  Pull these new models from the platform after an export control
• 15:49
  directive from the US government.
• 15:51
  So the purpose behind these kinds of directives is to
• 15:54
  essentially to protect national security interests and measures.
• 15:57
  Anthropic has issued a statement on the situation in a blog post
• 16:01
  and I will be pulling a lot of my reporting from.
• 16:04
  That, but here's basically. What went down?
• 16:06
  So the US government ordered Anthropic to suspend.
• 16:09
  Access to Fable 5. And mythos. 5 by quote.
• 16:13
  Any foreign national, whether inside or outside the United
• 16:16
  States, including foreign national anthropic employees.
• 16:19
  The last section being very harmful to anthropic.
• 16:22
  Of course, because they. Wouldn't only have U.S. citizens
• 16:25
  as employees and on top. Of this this directive.
• 16:28
  Is obviously something that's very difficult to control.
• 16:32
  Slash monitor like this information is not something
• 16:35
  that anthropic. Would have about its.
• 16:36
  Users so in response to that. Anthropic decided to disable.
• 16:41
  These models for all. Users to be sure that.
• 16:43
  They complied with the directive.
• 16:45
  The reason why the US government.
• 16:47
  Issued this directive was. Because they had.
• 16:50
  Discovered a way of. Jailbreaking Fable 5 And
• 16:53
  actually the concerns came from Amazon, which is and interesting
• 16:57
  in and of itself and for context in general, jailbreaking is the
• 17:00
  process of bypassing guardrails around a system and making
• 17:04
  modifications to the system that one intended I remember back in
• 17:08
  the day when I was. Like in primary.
• 17:10
  School, I think people were like jailbreaking iPod touches and
• 17:14
  like iPhones to customize them, and that was like my
• 17:17
  introduction to what jailbreaking is.
• 17:20
  But. Bringing it back to anthropic,
• 17:21
  we don't know the exact nature of the jailbreak, and this term
• 17:24
  is. Obviously used quite broadly, so
• 17:27
  it is really. Difficult to know the extent
• 17:29
  this is what Anthropic said about it in the statement.
• 17:31
  Quote To date the government has only given us verbal evidence of
• 17:35
  a potential narrow non universal jailbreak, which essentially
• 17:39
  consists of asking the model to. Read a specific code.
• 17:41
  Base and fix any software flaws. If we were to assume what?
• 17:45
  A jailbreak might look like in this context.
• 17:47
  It could potentially be something about bypassing the.
• 17:50
  Guardrails that anthropic. Is put in place.
• 17:52
  To limit cybersecurity attacks. And things of that nature.
• 17:56
  In saying that, Anthropic does contend that the model is still
• 17:59
  safe. To use.
• 18:00
  Throughout the statement, Anthropic notes a couple of key
• 18:02
  points. Firstly, they claim that.
• 18:05
  Perfect jailbreak. Resistance is just not possible
• 18:09
  for any model. Provider at this stage.
• 18:11
  They also note that they have yet to identify any quote UN
• 18:15
  quote universal jailbreak for the model, which is essentially
• 18:19
  A jailbreak method that can broadly bypass the model
• 18:22
  safeguards. On top of that, and perhaps most
• 18:24
  controversially, Anthropic adds. That the level of capability
• 18:28
  that led to this directive is actually available.
• 18:31
  By other models as well and they mentioned open AI.
• 18:34
  S GB T. 5.5 As an example, as the statement says quote we are
• 18:38
  complying with the governments legal directive and are removing
• 18:42
  access. To fable. 5 and Mythos 5.
• 18:44
  For all users, however, we. Disagree that the finding of a
• 18:48
  narrow potential. Jailbreak should be the cause
• 18:50
  for. Recalling a commercial model
• 18:52
  deployed to hundreds of millions of people if this standard was
• 18:56
  applied across the industry. We believe.
• 18:58
  It would essentially hold all new model deployments for all
• 19:01
  frontier model. Providers.
• 19:03
  So that's essentially the breakdown and there have been a
• 19:05
  couple interesting points of discussion that have come up
• 19:09
  online and from experts about. All this.
• 19:12
  So the response that's first prompted questioning about
• 19:15
  whether Anthropic has been unfairly.
• 19:17
  Targeted by the US. Government.
• 19:20
  This of course isn't entirely out of the question, considering
• 19:22
  the back and forth between them earlier.
• 19:24
  This year. I did cover this on the podcast,
• 19:26
  but as a quick reminder, Anthropic was essentially
• 19:29
  pushing back against the US government when it came to how
• 19:32
  its models. Could be used.
• 19:34
  By the Pentagon slash Like federal agencies, Anthropic
• 19:38
  really doubled down on the. Fact that Claude.
• 19:40
  Was not intended to be used. For.
• 19:42
  Domestic surveillance or deadly means after this anthropic was
• 19:46
  essentially blacklisted. Because it was designated as a
• 19:50
  supply. Chain risk, which required
• 19:52
  defense contractors to stratify that they don't use Claude when
• 19:56
  doing work for the Pentagon. It was like a whole thing I did
• 19:59
  cover on the pod like I mentioned and Anthropic has
• 20:02
  since gone on to. Sue the US.
• 20:03
  Government which is currently ongoing and I know I've
• 20:06
  mentioned this here and there throughout different episodes,
• 20:08
  but we know that the Trump administration responds very.
• 20:11
  Well, to like. Praise ideological alignment and
• 20:14
  just like positive personal relationships, which is
• 20:17
  something that many of the big tech companies slash tech.
• 20:20
  Bros have. Really embraced anthropic hasn't
• 20:24
  in the same way and has instead pushed back.
• 20:27
  So with that background, people are questioning the legitimacy
• 20:31
  of the directive in the 1st place.
• 20:33
  Another interesting point of discourse that has come out of
• 20:36
  this and I am putting on my international relations hat for
• 20:38
  this one. We're going back to university,
• 20:40
  but many experts have noted. That this whole saga.
• 20:44
  Epitomizes the problem with relying on US based artificial
• 20:47
  intelligence and. That those outside the US.
• 20:51
  Probably need to develop their own sovereign capabilities.
• 20:55
  I do think this. Idea of like self.
• 20:56
  Reliance has been something we have been seeing across the
• 21:00
  board when it comes to international relations and
• 21:02
  foreign policy, especially since COVID.
• 21:05
  Nevertheless, this is probably one of the most obvious
• 21:08
  instances when it comes to. AI, as Robert Hart put it for a
• 21:12
  piece for The Verge quote. Trump may be restricting mythos
• 21:15
  and fable as a matter of national security, but the
• 21:17
  argument? Cuts both ways.
• 21:19
  And with Washington now asking if AI is too important for
• 21:22
  everyone to have access, other governments are asking whether
• 21:25
  they can afford for Washington to.
• 21:27
  Decide who does. This is obviously all an
• 21:30
  evolving story. At the time of recording, Fable
• 21:32
  still isn't accessible by the masses, but I am curious to hear
• 21:36
  if you had thoughts about all of this and I will.
• 21:38
  Keep you updated. If you guys are interested, OK,
• 21:41
  another change of pace, but the last thing I wanted.
• 21:43
  To talk about today is a pretty. Short story, but I wanted to
• 21:46
  discuss these dopamine sites that have started to make the
• 21:49
  rounds online. So these dopamine sites.
• 21:51
  I believe. Have originated in Korea and
• 21:54
  they are essentially mimicking online shopping, food delivery
• 21:59
  and that kind of stuff. So you can go to a website and
• 22:02
  fill in a. Cart of like food.
• 22:04
  Clothing, whatever. Exactly as you would regularly.
• 22:07
  But you don't actually order anything.
• 22:09
  But you are getting that like dopamine hit.
• 22:11
  Of as if you. Were ordering something so I
• 22:14
  first saw. This covered by a TikTok.
• 22:16
  Creator named Angel, who covers tech and AI and it's since been.
• 22:21
  Picked up by various. Outlets including Vice and
• 22:24
  Duxterto. One of the earliest English
• 22:27
  language pieces I found was from the Korea Times at the end of
• 22:31
  May though, so this has kind of been circulating.
• 22:34
  For a while, but. That piece describes the
• 22:37
  dopamine sites as quote online. Spaces that offer quick.
• 22:40
  Stimulation. Aside from fake delivery.
• 22:43
  Online shopping apps. There is also a.
• 22:45
  Version that mimics a smoke break, so you log on and you can
• 22:49
  see who else is there and leave little messages.
• 22:52
  For each other the. Career Time spoke to a 24.
• 22:55
  Year old college. Student named Lee who visited
• 22:57
  the smoking break site during exam periods and he said, quote,
• 23:02
  I'm not actually smoking, but it feels like I'm taking a break
• 23:04
  with someone. So it's strangely comforting.
• 23:06
  So I thought the concept of these websites were really
• 23:10
  interesting and worth covering because like, first and
• 23:12
  foremost, it's just another example of how for young people,
• 23:16
  the online and offline worlds are just completely intertwined.
• 23:19
  The activities and the feelings that we're kind of chasing for
• 23:23
  or experience in our real life can be replicated online and
• 23:26
  vice versa. And Angel actually framed the
• 23:29
  site as a form of. Digital Self.
• 23:32
  Of regulation like you can. Perhaps go fake?
• 23:35
  Online shopping to curb a shopping impulse.
• 23:37
  Or take a fake. Smoke break to combat
• 23:39
  loneliness, like that kind of stuff.
• 23:41
  I think that these apps are actually just people trying to
• 23:44
  regulate themselves through these apps.
• 23:47
  The whole phenomena can showcase like how a society is feeling
• 23:50
  and the fact that this is on the rise.
• 23:53
  I mean, people are lying for something and that they need
• 23:55
  these sites to fill it and that maybe doomscrolling on playing
• 23:58
  video games are just not it anymore and people want to feel
• 24:00
  that sense of familiarity and comfort.
• 24:03
  These websites gaining popularity in Korea, which.
• 24:06
  I feel like could. Eventually trickle across to
• 24:08
  other countries. Made me think about like what
• 24:11
  the Internet is actually for now, a few.
• 24:13
  Weeks ago we talked about how Google search.
• 24:16
  Is really entering its AI era and I think the dopamine site
• 24:20
  phenomenon kind of points to a similar.
• 24:23
  Shift and way that we use. The Internet because it's
• 24:26
  becoming less and less about finding information and more
• 24:29
  about actually doing things where the action.
• 24:32
  Itself like whether that's. Browsing or like adding to cart
• 24:35
  or checking out has become the whole point and I did.
• 24:38
  Discuss that a little. Bit in that episode.
• 24:41
  This is a little. Bit of a rogue comparison, but
• 24:43
  there's this other platform that I've been seeing go.
• 24:45
  Viral over the past. Week and it's also about taking
• 24:49
  action, but it's actually about paying someone to do something
• 24:52
  for you and this website. Is a little bit evil, but it's
• 24:55
  called. Pump Fun Go and it's a bounty
• 24:58
  marketplace that launched earlier this month with the
• 25:00
  tagline quote pay anyone to do anything and on the Internet
• 25:05
  that can be a bit concerning in a piece.
• 25:08
  For garbage day. Ryan Broderick picked up on this
• 25:10
  and he wrote quote. The majority of active bounties
• 25:13
  on the site are mostly just. Elaborate ways to take.
• 25:16
  Advantage of vulnerable people, bounties asking them to get meme
• 25:20
  coins tattooed on their foreheads and bounties asking
• 25:22
  them to hurt animals. It's about as dark.
• 25:24
  As the Internet gets and I feel like these websites and the
• 25:28
  pursuit of dopamine shows. That's where we are.
• 25:31
  At with the Internet right now, on the one end, you have like
• 25:34
  young people turning to. These quiet, cozy little.
• 25:37
  Corners of the Internet just to like feel something and have
• 25:39
  like some kind of community or carbon impulse.
• 25:42
  And on the other, you have platforms like pump fun Go,
• 25:45
  which is kind of a mouthful who are completely monetizing that
• 25:48
  same impulse and pointing it to the most vulnerable people
• 25:53
  online. It's the same kind of dopamine
• 25:55
  seeking tendencies, but I guess with like completely different
• 25:58
  intentions and ethics. Well, that brings us to the end
• 26:01
  of Mini Tech. Scroll for today.
• 26:03
  That was a crazy episode. Like those three stories will.
• 26:07
  Whiplash each time, but I hope you guys enjoyed.
• 26:10
  Lauren will be back. To speak to you.
• 26:13
  Tomorrow and thanks for listening.
• 26:15
  Bye. Centennial World Podcast Network
• 26:17
  operates on Gadigal land. We would like to acknowledge the
• 26:20
  Gadigal of the Eora Nation, the traditional custodians of this
• 26:23
  land, and pay our respects to elders both past and present.