Latest / Connecting the Dots / Trump’s AI Action Plan and Google’s Virtual Shopping Suite

Transcript

Trump’s AI Action Plan and Google’s Virtual Shopping Suite · Connecting the Dots

• 0:00
  Today's tech news moves fast, but understanding
• 0:03
  it shouldn't be hard. That's why Snarfl Solutions
• 0:06
  Group brings you Connecting the Dots, where we
• 0:09
  break down what's happening in tech and why it
• 0:11
  matters to you. Now, let's dive in with your
• 0:14
  hosts. Welcome, curious minds. It's Thursday,
• 0:18
  July 24, 2025, and once again, you've given us
• 0:22
  quite a stack of interesting stuff, articles,
• 0:24
  notes, research. Our job, like always, is to
• 0:28
  unpack it all, pull out those key bits of knowledge,
• 0:31
  and help you get really well -informed without
• 0:33
  getting totally swamped. That's right. And today,
• 0:36
  we're connecting some really significant dots,
• 0:38
  I think. We'll be looking at how artificial intelligence
• 0:41
  isn't just this high -level policy thing, you
• 0:43
  know, government and international relations,
• 0:45
  but also how it's actually changing things for
• 0:47
  you every day, like big national plans on one
• 0:49
  hand and how you shop online on the other. It's
• 0:51
  quite the spectrum. Exactly. So Settling, we've
• 0:54
  got a fair bit to explore connecting those really
• 0:56
  ambitious national goals with the practical tools
• 0:58
  shown up right on your phone. OK, so let's dig
• 1:01
  into this first set of sources you gave us. They
• 1:03
  really focus on the Trump administration's recently
• 1:05
  released AI action plan. And this isn't just
• 1:09
  some brief memo, right? It's a full 23 -page
• 1:11
  document. Sets out a pretty bold vision, global
• 1:14
  AI dominance for America. And the core idea seems
• 1:17
  pretty clear. Build, baby, build. It really signals
• 1:20
  they want AI companies moving fast. as fast as
• 1:23
  possible. Yeah, and what's really striking, I
• 1:25
  think, is how much the plan seems to, well, defer
• 1:29
  to the private sector, empower them, really.
• 1:31
  Your sources point out these deep ties between
• 1:33
  the administration and the big tech players,
• 1:36
  OpenAI, NVIDIA, Meta, Google, Microsoft. And
• 1:39
  there's even specific mention of this private
• 1:41
  thing called the Stargate Project. That's a huge
• 1:44
  team -up. OpenAI, Oracle, SoftBank. And they're
• 1:47
  not just talking about a little infrastructure.
• 1:49
  We're talking hundreds of billions. Oracle and
• 1:51
  OpenAI specifically are looking at building like
• 1:55
  4 .5 gigawatts of data center capacity. I mean,
• 1:57
  just to give you a sense of that scale, it's
• 1:59
  enough power for maybe a few million homes. So
• 2:01
  it shows the massive industrial scale they're
• 2:03
  thinking about for AI. It really does feel like
• 2:05
  Silicon Valley handed Washington their wish list.
• 2:08
  And well, here it is. Wow, 4 .5 gigawatts. That's
• 2:11
  huge. And this whole plan is built around three
• 2:13
  main ideas or pillars, as they call them. boosting
• 2:16
  AI innovation, building out more AI infrastructure,
• 2:19
  and promoting American AI around the world. And
• 2:22
  it seems like deregulation is really a big part
• 2:24
  of how they want to achieve that. Oh, absolutely.
• 2:26
  The White House proposals for cutting back on
• 2:29
  what they call onerous regulations and bureaucratic
• 2:32
  red tape, they're pretty sweeping. And if you
• 2:35
  look at the details you shared, these aren't
• 2:36
  just symbolic things. They have teeth. Exactly.
• 2:39
  Like the plan talks about reviewing investigations
• 2:41
  from the Biden era FTC. Anything seen as, quote,
• 2:45
  Unduly burdening AI innovation, but it goes further
• 2:49
  right suggesting maybe cutting federal AI funding
• 2:52
  to states if their regulations are seen as unfriendly
• 2:55
  to AI So if a state wants a tough third -party
• 2:59
  audits for AI tech they might lose funding. That
• 3:02
  seems to be the implication. It puts a lot of
• 3:04
  pressure on states to, you know, create a friendlier
• 3:06
  climate for these AI companies. And this deregulation
• 3:08
  push, it even touches areas you might not expect,
• 3:10
  like environmental rules. Yeah, that's a really
• 3:12
  interesting and maybe controversial part. AI
• 3:15
  data centers, as we know, use enormous amounts
• 3:17
  of water and electricity. It's a big sustainability
• 3:20
  question. But this AI action plan... It kind
• 3:23
  of just waves away a lot of those environmental
• 3:25
  worries. It suggests streamlining, maybe even
• 3:28
  cutting back on environmental rules, possibly
• 3:30
  including things like Clean Air and Clean Water
• 3:32
  Act standards, just to build faster. And critics,
• 3:35
  as your sources note, are really jumping on this,
• 3:38
  calling it a fossil fuel subsidy in disguise,
• 3:40
  especially when you see predictions that data
• 3:43
  center electricity use might triple by 2028.
• 3:47
  So it really forces you to ask, what are the
• 3:49
  long -term costs here? What are we trading for
• 3:51
  this speed? It's a real tension. That's a huge
• 3:53
  point, the trade -offs. And it kind of leads
• 3:56
  to what seems like the core philosophy here.
• 3:59
  This idea of a dynamic, try -first culture for
• 4:01
  AI. What does that mean exactly? Build it first,
• 4:05
  test it out. Pretty much. Build it, deploy it,
• 4:07
  let it iterate quickly, and then figure out the
• 4:09
  usefulness or the risks down the line. And they
• 4:11
  want it used everywhere. Science, healthcare,
• 4:14
  farming, government services. And the military
• 4:16
  too, you mentioned. Very notably, yes. The Pentagon's
• 4:18
  already announced these big contracts, up to
• 4:21
  $200 million each, with OpenAI, Google, Anthropic,
• 4:25
  XAI. So... Yeah, so it's not just theoretical.
• 4:29
  It's happening now. It's happening. The implementation
• 4:31
  is definitely underway. And you mentioned the
• 4:33
  industry influence earlier. How clear is that
• 4:36
  link between what companies wanted and what's
• 4:38
  in the plan? Oh, it's incredibly clear. Recommendations
• 4:41
  from companies like OpenAI, Empropic, Google,
• 4:44
  Microsoft. They really seem to line up almost
• 4:47
  exactly with what's in the final document things
• 4:49
  like you know Categorical exclusions from environmental
• 4:52
  reviews for AI projects right limits on what
• 4:55
  states can regulate Creating these sandboxes
• 4:58
  for startups to just test things freely without
• 5:00
  much oversight initially It's pretty stark example
• 5:03
  of industry recommendations becoming policy.
• 5:05
  Okay, and then there's the whole intellectual
• 5:07
  property angle That's a massive fight right now
• 5:10
  huge president Trump's position according to
• 5:12
  the sources is that training AI on copyrighted
• 5:16
  stuff books articles art music should basically
• 5:19
  be fair use yeah his argument is it's like how
• 5:22
  humans learn right by taking in lots of information
• 5:26
  which is exactly what companies like open ai
• 5:28
  and microsoft are arguing for especially with
• 5:30
  what over 40 lawsuits filed against ai firms
• 5:33
  since 2022 about this Precisely. This could totally
• 5:37
  reshape copyright law as we know it, if that
• 5:40
  view prevails. OK, speed and innovation sound
• 5:42
  good, but what are the potential downsides experts
• 5:45
  are pointing out? Well, the big concern is safety
• 5:47
  and trust. Critics are saying this whole approach
• 5:49
  is sidestepping safety precautions that sustain
• 5:52
  public trust, rushing ahead without enough guardrails.
• 5:56
  And there's also this potential irony or paradox,
• 5:58
  right? The administration has cut some research
• 6:00
  funding, and there are still visa issues for
• 6:03
  foreign scientists. And these foreign scientists
• 6:05
  make up like 70 % of AI grad students in the
• 6:08
  US, a huge number. 70%, wow. So the worry is,
• 6:12
  by restricting talent and maybe underfunding
• 6:14
  basic research, we could actually be hurting
• 6:16
  our own competitiveness in the long run. Some
• 6:19
  critics are even calling it self -inflicted ignorance,
• 6:22
  basically shooting ourselves in the foot while
• 6:24
  trying to race ahead. And there's another layer
• 6:25
  to it, specific directives about idealizing Yes,
• 6:29
  that's in there too. The plan recommends federal
• 6:32
  buying guidelines should make sure the government
• 6:34
  only uses AI models that are, quote, free from
• 6:37
  top -down ideological bias. And it mentions a
• 6:40
  potential executive order to get rid of woke
• 6:42
  Marxist lunacy from AI and also to remove terms
• 6:46
  like misinformation, diversity, equity and inclusion
• 6:49
  and climate change from the government's AI risk
• 6:52
  management frameworks. So it introduces a very
• 6:55
  specific political viewpoint into how government
• 6:57
  AI should be assessed and used. So putting it
• 7:00
  all together, what's the big picture takeaway
• 7:02
  for someone listening? Well, I think it paints
• 7:04
  a really clear picture of an aggressive push
• 7:07
  for AI dominance, a really aggressive one, with
• 7:09
  a huge emphasis on deregulation, letting the
• 7:12
  private sector lead and speed above almost all
• 7:15
  else. And it really does highlight this tension,
• 7:17
  doesn't it, between that drive for innovation,
• 7:20
  the economic promise, and then the worries about
• 7:22
  safety, ethics, the environment, and even making
• 7:25
  sure we have the talent we need. Your sources
• 7:27
  sum it up by saying it puts the nation's AI future
• 7:31
  in Silicon Valley's hands. feels pretty accurate
• 7:34
  based on the plan. That's a really helpful summary
• 7:36
  of the big picture strategy. Yeah. Okay, let's
• 7:39
  shift gears completely now. From that huge national
• 7:41
  policy stage, let's zoom way, way down, down
• 7:44
  to something much more personal, your online
• 7:46
  shopping. Because our sources also talk about
• 7:48
  Google launching some pretty interesting new
• 7:50
  AI shopping features. And the headline grabber
• 7:53
  seems to be this virtual try -on tool for clothes.
• 7:56
  It's apparently available across the US now.
• 7:58
  Yeah, this is pretty neat actually. What's really
• 8:00
  interesting is how personal it gets. It's not
• 8:02
  just seeing clothes on standard models anymore.
• 8:05
  You can actually upload a photo of yourself,
• 8:07
  a full -length one. And the AI then shows you
• 8:10
  how that specific item might look on your body
• 8:12
  shape. Google says their AI understands. You
• 8:16
  know, the human body, how clothes drape, how
• 8:18
  different fabrics fold or stretch on different
• 8:20
  people. Think about that. It could seriously
• 8:23
  change how you decide what to buy online. Makes
• 8:26
  it feel much more tangible, maybe more confident.
• 8:28
  That sounds like a game changer. I mean, potentially
• 8:31
  fewer returns because something just doesn't
• 8:33
  fit, right? That's a big hassle for everyone.
• 8:36
  Exactly. Less waste, less frustration. And speaking
• 8:38
  of making better choices, they've also apparently
• 8:41
  rolled out smarter price alerts. Right. So instead
• 8:44
  of just getting an alert when any price drops,
• 8:46
  you can set it for very specific things like,
• 8:48
  only tell me if this shirt in blue, size medium,
• 8:51
  drops below $30. Much more control. Okay, that
• 8:54
  is smarter. It finds the deals you actually care
• 8:56
  about. Precisely. Tailoring the deals to your
• 8:59
  actual needs, not just blasting generic sales
• 9:01
  at you. So these tools... They're part of a bigger
• 9:04
  push from Google. Yeah, it seems like it's all
• 9:06
  part of Google expanding its AI mode. The goal
• 9:09
  seems to be making web browsing, especially shopping,
• 9:12
  easier, faster, and helping you save money or
• 9:15
  find exactly the right thing. Your sources even
• 9:18
  mentioned something coming this fall, AI -powered
• 9:20
  style inspiration. Oh, like what? Well, not just
• 9:24
  for clothes, but offering ideas for whole outfits,
• 9:27
  or even room designs using their vision technology.
• 9:30
  So you could maybe upload a picture of, say,
• 9:33
  a couch you like, and the AI suggests coordinating
• 9:36
  decor, that kind of thing. That's pretty futuristic.
• 9:38
  And then there's this other app they launched,
• 9:40
  Doffle. Yeah, that one's interesting. It's described
• 9:42
  as experimental, a standalone app. And it takes
• 9:46
  that virtual try -on idea even further. You can
• 9:48
  upload pictures of outfits from almost anywhere.
• 9:51
  Like a photo from a thrift store. or screenshot
• 9:53
  from Instagram exactly and visualize it on your
• 9:56
  digital self and apparently it can even turn
• 9:58
  those static images into little AI generated
• 10:01
  videos of you wearing the outfit so it really
• 10:04
  starts to blur the lines between just seeing
• 10:06
  something cool imagining it on yourself and maybe
• 10:08
  deciding to buy it or something like it So for
• 10:10
  the person shopping online, what's the bottom
• 10:12
  line here? What does this all mean? Well, on
• 10:14
  one level, these tools are definitely designed
• 10:16
  to make your shopping easier, more informed,
• 10:19
  hopefully less wasteful, fewer returns. But maybe
• 10:22
  think about the other side, too, for only businesses.
• 10:25
  This means having really high quality product
• 10:27
  photos is going to be even more crucial. Pricing
• 10:30
  needs to be sharp, dynamic, maybe, and being
• 10:33
  optimized for visual people searching with images,
• 10:36
  not just words, that's going to be huge. It really
• 10:39
  pushes retailers to up their game in this AI
• 10:41
  enhanced world. Maybe even forces physical stores
• 10:44
  to think about how they compete with this kind
• 10:47
  of personalization online. So quite a journey
• 10:50
  today. We've gone from these huge, sweeping national
• 10:53
  ambitions about winning a global AI race, involving
• 10:57
  massive policy changes, big controversies, all
• 11:00
  the way down to these very immediate practical
• 11:03
  tools that are changing how you might pick out
• 11:05
  a shirt online. It really does show how interconnected
• 11:08
  it all is, doesn't it? AI influencing everything
• 11:10
  from, yeah, national security and economics down
• 11:14
  to your shopping cart. And those big decisions
• 11:16
  being made at the government level about infrastructure,
• 11:19
  about regulation, they're going to shape the
• 11:21
  very technology that powers these consumer tools
• 11:24
  we're just starting to see. It just highlights
• 11:26
  how deeply technology is weaving itself into,
• 11:29
  well, pretty much every part of life now. Absolutely.
• 11:31
  And as AI keeps developing so quickly, integrating
• 11:35
  into government, into our daily lives. It really
• 11:38
  leaves us with a big question, I think. What's
• 11:40
  the right balance? How do you think we should
• 11:42
  weigh that relentless drive for innovation against
• 11:45
  the need for public safety, for ethical guidelines,
• 11:48
  and also protecting our individual privacy in
• 11:50
  this new landscape? It's something we'll all
• 11:53
  be grappling with more and more. Thank you for
• 11:55
  joining us as we explored your sources today.
• 11:56
  We really hope you found these insights valuable.
• 11:59
  Maybe give you something to think about. Till
• 12:00
  next time, keep exploring, keep asking questions,
• 12:03
  and keep learning.