Latest / Connecting the Dots / Connecting the Dots - 7-10-2025

Transcript

Connecting the Dots - 7-10-2025 · Connecting the Dots

• 0:00
  Today's tech news moves fast, but understanding
• 0:03
  it shouldn't be hard. That's why Snarful 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. Hello and welcome to Connecting the Dots.
• 0:17
  I'm Alex. And I'm Morgan. Thanks for joining
• 0:19
  us today, Thursday, July 10th, 2025. Let's kick
• 0:23
  things off with a quick look at what's happening
• 0:24
  out there. Weather -wise, a bit of a mixed bag
• 0:27
  across the country. Yeah, definitely. Sounds
• 0:29
  like severe thunderstorms, large hail, even heavy
• 0:32
  rain are hitting the central plains and Midwest.
• 0:35
  And watch out for potential flash flooding in
• 0:37
  the Carolinas, maybe stretching up into Virginia
• 0:38
  too. Meanwhile, it's uh... Pretty dry and risky
• 0:42
  out west. Elevated fire weather concerns in the
• 0:44
  Four Corners region. And the desert southwest.
• 0:47
  Extreme heat warnings. Stay cool out there. Closer
• 0:50
  to home here in Sacramento. Well, it's gonna
• 0:51
  be another hot one. Forecast is areas smoke with
• 0:54
  a high reaching 101 degrees Fahrenheit. Ugh,
• 0:58
  that's smoke. Okay, let's shift gears to the
• 1:00
  markets. How did things close out yesterday?
• 1:02
  All right, the Dow Jones finished at 44 ,690
• 1:07
  .87. That's up about 263 points. or 0 .59%. Okay,
• 1:13
  decent daily gain. How's the week looking for
• 1:15
  the Dow? Well, actually down slightly for the
• 1:17
  week. Since Monday, it's lost about 112 points,
• 1:19
  so that's down 0 .25%. Gotcha. What about the
• 1:23
  S &P 500? S &P closed at 6 ,277 .9 to 4. A smaller
• 1:28
  gain yesterday, up just over 11 points, 0 .18%.
• 1:32
  And the week for the S &P? That one's actually
• 1:34
  in the dream for the week, up about 19 points
• 1:37
  since Monday, so a gain of 0 .30%. Interesting
• 1:40
  divergence there. And Bitcoin. Always watching
• 1:43
  that one Bitcoin closed at a hundred and eleven
• 1:44
  thousand four hundred and forty eight point twenty
• 1:46
  a pretty small daily change up just a hundred
• 1:49
  and forty six dollars about Point one three percent
• 1:51
  with the weekly trend much stronger there. It's
• 1:53
  up over three thousand one hundred dollars since
• 1:56
  Monday That's a gain of two point nine one percent.
• 1:58
  So quite a bit of upward movement this week Wow,
• 2:00
  okay, and finally before we get into our main
• 2:03
  stories, let's talk baseball the San Francisco
• 2:05
  Giants How do they do last night? Yeah, not great
• 2:08
  news for Giants fans, I'm afraid. They played
• 2:10
  the Philadelphia Phillies and lost pretty badly.
• 2:13
  Oh no, what was the score? For an all -scorer
• 2:15
  was 13 to nothing. 13 -0. Ouch. Okay, well, try
• 2:20
  to find a silver lining here. Well, Jesus Lizardo
• 2:22
  pitched a great game for the Phillies, obviously.
• 2:24
  He got the win. And Bryce Harper had a huge night
• 2:26
  for them, went 4 -6 with the homerun. Right.
• 2:29
  Well, tough loss for the Giants, but hey, it's
• 2:30
  a long season. Gotta bounce back. Exactly. OK,
• 2:33
  so let's move on from the real world updates
• 2:35
  and get into the tech stories we're tracking
• 2:37
  today from our snarfle tech article report, The
• 2:40
  Morningstar. Sounds good. First up, we've got
• 2:43
  a pretty bold launch from Elon Musk's company,
• 2:45
  XAI. That's right. They've just rolled out Grok
• 2:49
  4 and they're not being shy about it. They're
• 2:51
  claiming it's the world's most powerful AI model.
• 2:54
  More powerful than like ChatGPT or Google's Gemini.
• 2:58
  That's the claim. They're saying it outperforms
• 3:00
  rivals like OpenAIs, ChatGPT, Google's Gemini,
• 3:04
  and Anthropix Claude on several key benchmarks.
• 3:06
  Which benchmarks are we talking about? Things
• 3:08
  like ARC, AGI, version 1 and 2, those tests,
• 3:11
  advanced reasoning, and get this, something called
• 3:14
  humanity's last exam. Humanity's last exam? That
• 3:17
  sounds... Intense yeah it's designed to see if
• 3:20
  an ai can tackle problems at a phd level musk
• 3:23
  himself is quoted saying grok four achieves phd
• 3:25
  levels in every subject with no exceptions wow
• 3:29
  okay that's quite the statement so what does
• 3:31
  this mean for you know businesses or professionals
• 3:34
  thinking about using it is it accessible. Well
• 3:37
  that's a key point there's a premium tier super
• 3:40
  grok heavy that costs three hundred dollars a
• 3:42
  month there's also a standard super grok for
• 3:45
  thirty dollars a month. 300 bucks a month? Yeah.
• 3:48
  That's quite a bit. Seems like they might be
• 3:50
  aiming for larger companies with deeper pockets.
• 3:52
  It certainly suggests that. For smaller businesses,
• 3:55
  that price point might be a barrier, even with
• 3:57
  those PhD level claims. It could push smaller
• 4:00
  players towards more specialized or cost -effective
• 4:04
  AI solutions. And is it fast? If it's super smart
• 4:07
  but slow, that could be an issue. That's another
• 4:09
  interesting detail. The reports mention its output
• 4:11
  speed is around 75 tokens per second. Now that's
• 4:14
  fast, but apparently slower than some competitors.
• 4:18
  So there's a potential trade -off. Super high
• 4:20
  accuracy maybe, but not the quickest response
• 4:23
  time. That could matter in customer service or
• 4:26
  time -sensitive analysis. Exactly. Tell them
• 4:28
  the application, right? Is top -tier smarts worth
• 4:30
  waiting a bit longer for? Fair point. And this
• 4:33
  launch... It's part of a bigger picture for XAI,
• 4:35
  isn't it? Absolutely. This isn't just about one
• 4:37
  model. They're actively trying to raise five
• 4:40
  billion dollars in funding. Five billion? Wow.
• 4:43
  And they have plans for this enormous supercomputer
• 4:46
  in Memphis. They're calling it Colossus, supposed
• 4:49
  to have one million GPUs. A million GPUs, that's
• 4:53
  just mind -boggling compute power. It really
• 4:56
  is, and there are even Whispers discussions about
• 4:59
  potentially merging with Tesla. For AI stuff,
• 5:02
  like RoboTaxis? Precisely. Musk has been very
• 5:05
  public about RoboTaxis being a huge part of Tesla's
• 5:08
  future value. So connecting XAI's power with
• 5:11
  Tesla's cars makes strategic sense, at least
• 5:13
  on paper. Okay, but it hasn't all been smooth
• 5:15
  sailing for Grok. Yeah. There have been some
• 5:17
  controversies. It's true. The sources mention
• 5:19
  the departure of ex -CEO Linda Yakarina recently,
• 5:23
  and previous versions of Grok had issues generating
• 5:26
  problematic content. Like what kind of content?
• 5:28
  Well, there were instances of anti -Semitic responses
• 5:31
  which actually led to a Turkish court blocking
• 5:33
  access. The Polish government also threatened
• 5:35
  action. Yikes. How did Musk explain that? His
• 5:38
  explanation was that the earlier version was,
• 5:40
  and I quote, too compliant to user prompts and
• 5:43
  too eager to please. Basically, it wasn't pushing
• 5:46
  back enough on potentially harmful requests.
• 5:49
  So they're trying to find that balance between
• 5:52
  capability and safety. That seems to be the constant
• 5:56
  struggle in AI development. It really is. How
• 5:58
  do you make it incredibly smart and useful without
• 6:00
  letting it go off the rails? Despite these issues,
• 6:03
  though, they've managed to secure partnerships
• 6:05
  with Oracle, Microsoft, and even Telegram to
• 6:08
  integrate Grok. So they're still pushing hard
• 6:11
  into the market. It's a fascinating space. OK,
• 6:13
  so while Ex -AI is launching these powerful models.
• 6:17
  There's another battle brewing, right? Over the
• 6:20
  people building these things. Oh, absolutely.
• 6:22
  Our next story digs into the, frankly, unprecedented
• 6:25
  talent war happening in AI, and Meta seems to
• 6:28
  be leading the charge. Meta. So, Mark Zuckerberg
• 6:31
  is trying to build up his AI teams. Big time.
• 6:34
  He's on what the sources call an absolute hiring
• 6:36
  blitz for Meta's new super intelligence group,
• 6:39
  their goal. Build AI that's as good as, or maybe
• 6:42
  even better than, humans. And how are they getting
• 6:44
  the talent? Just posting job ads. Huh, not quite.
• 6:48
  They're aggressively poaching top AI minds from
• 6:51
  their biggest competitors. We're talking Google,
• 6:54
  OpenAI, Anthropic, even Apple. Poaching usually
• 6:57
  means offering more money, right? How much are
• 6:59
  we talking here? The numbers are just staggering.
• 7:02
  Sources mention compensation packages worth tens
• 7:05
  of millions of dollars per year. Tens of millions
• 7:08
  per year? Yeah. And some reports even claim sign
• 7:11
  -on bonuses reaching over a hundred million dollars.
• 7:14
  A hundred million just to sign on? That's almost
• 7:17
  unbelievable. For someone working in tech listening
• 7:20
  right now, this must sound like a different universe.
• 7:22
  It really underscores the insane demand for top
• 7:25
  -tier AI expertise. They're pulling out all the
• 7:27
  stops. Apparently Zuckerberg himself is personally
• 7:30
  involved, hosting dinners, even WhatsApping potential
• 7:33
  recruits directly in a group chat called Recruiting
• 7:36
  Party A. A WhatsApp recruiting party hosted by
• 7:38
  Zuck. Okay, that's definitely a personal touch.
• 7:41
  Who are some of the big names they've managed
• 7:42
  to snag? Some really high profile people. Rueming
• 7:45
  Pang, who led AI model development at Apple.
• 7:49
  Alexander Wang, the CEO of Scale AI. Daniel Gross,
• 7:53
  a well -known startup founder. Nat Friedman,
• 7:55
  the former CEO of GitHub. Plus key researchers
• 7:58
  from OpenAI and Ampropic. Wow, that's like an
• 8:00
  AI all -star team they're assembling. Pretty
• 8:02
  much. It suggests Meta believes that at the end
• 8:06
  of the day, Human talent is still the biggest
• 8:08
  bottleneck in the AI race. They're trying to
• 8:11
  concentrate that brain power. It makes you wonder
• 8:13
  though, is this sustainable? Could it create
• 8:15
  a brain drain from other important areas or maybe
• 8:17
  even some kind of talent bubble? That's a really
• 8:19
  good question. It certainly shows businesses
• 8:21
  how critical talent acquisition is in AI right
• 8:24
  now. It's driving up costs across the board.
• 8:26
  But investors seem to like the strategy. Meta's
• 8:29
  stock recently hit an all -time high, $747 .90.
• 8:34
  So the market believes spending big on talent
• 8:36
  is the right move for AI leadership. OK. So we
• 8:39
  have powerful new models like ROK4 and this intense
• 8:42
  battle for the people who build them. What about
• 8:44
  the tools actually reaching everyday users? Anything
• 8:47
  new on that front? Yes, definitely. Our final
• 8:50
  story today is about Google making some pretty
• 8:53
  sophisticated creative AI tools much more accessible.
• 8:58
  Oh, yeah. What's Google rolling out? They've
• 9:00
  just updated their Gemini app with a really cool
• 9:02
  new feature, photo to video generation. It uses
• 9:05
  their VO3 model. Photo to video. So you give
• 9:08
  it a picture and it makes a video out of it.
• 9:10
  Basically, yeah. You upload a still image. give
• 9:13
  it a text prompt describing what you want to
• 9:15
  happen, and it generates an eight second video
• 9:18
  clip, complete with sound. Whoa! Okay, that sounds
• 9:21
  fun. Like you could take a photo of your pet
• 9:23
  and make it look like it's talking or something.
• 9:25
  Exactly. Or animate a drawing you made, add movement
• 9:27
  to a landscape photo, bring everyday objects
• 9:29
  to life. The possibilities are pretty wide open
• 9:32
  for creators or just, you know, anyone wanting
• 9:34
  to experiment. That could be huge for social
• 9:36
  media, for just playing around, is it? Catching
• 9:39
  on? Seems like it. Apparently over 40 million
• 9:41
  videos have already been generated using VO3
• 9:44
  since it launched back in May. So, yeah, pretty
• 9:47
  rapid adoption. 40 million, okay. That's significant.
• 9:50
  What about businesses? Could they use this? Absolutely.
• 9:53
  Think about quick content generation. Transforming
• 9:56
  static product shots into dynamic little videos
• 9:58
  for ads or social media. Making presentations
• 10:01
  more engaging. It lowers the barrier to creating
• 10:05
  video content quite a bit. But what about the
• 10:07
  realism? Can you tell it's AI -generated? The
• 10:10
  reports say the outputs are staggeringly realistic.
• 10:13
  It handles things like real -world physics and
• 10:16
  apparently has accurate lip -syncing capabilities
• 10:18
  too. That's impressive. But it also makes you
• 10:21
  think about the whole fake news or deep fake
• 10:23
  issue, doesn't it? If AI can make super realistic
• 10:26
  videos easily. That's a really crucial point.
• 10:28
  It definitely blurs the lines further between
• 10:30
  what's real and what's generated. It means all
• 10:32
  of us need to be even more critical viewers of
• 10:35
  media online. Developing that digital literacy
• 10:37
  is becoming essential. So how do you get access
• 10:40
  to this? Is it free? And what about safety measures?
• 10:42
  It's part of Google's paid AI tiers. You need
• 10:46
  to be a Google AI Pro subscriber, which is $19
• 10:49
  .99 a month, or AI Ultra, which is much pricier
• 10:52
  at $249 .99 a month. OK, so not free, but the
• 10:55
  Pro tier isn't wildly expensive for someone who'd
• 10:58
  use it a lot in safety. Google says they've built
• 11:00
  in safeguards. There's a visible VO watermark
• 11:02
  on the videos in the bottom right corner. Plus,
• 11:06
  an invisible digital watermark called Synthide
• 11:08
  embedded in the file. And they're testing it
• 11:10
  for problems. Yeah. They mentioned using red
• 11:12
  teaming, basically having internal teams try
• 11:15
  to misuse a tool or find flaws, like hackers
• 11:18
  would, so they can fix issues proactively. Gotcha.
• 11:21
  So trying to be responsible about rolling out
• 11:23
  powerful tech. OK. OK, let's quickly recap the
• 11:25
  main stories we covered today from the Morningstar.
• 11:28
  All right. First, we looked at XAI launching
• 11:30
  Grok 4. super ambitious claims about being the
• 11:34
  most powerful AI, that PhD level intelligence,
• 11:37
  but also questions around its cost, its speed,
• 11:41
  and some past controversies. It's all part of
• 11:43
  a bigger strategic push. Right. Then we shifted
• 11:46
  to Meta's incredible spending spree to hire top
• 11:49
  AI talent, tens, maybe hundreds of millions in
• 11:52
  compensation, poaching big names, all driven
• 11:54
  by Zuckerberg personally. It's reshaping the
• 11:57
  tech talent market. And finally, we talked about
• 12:00
  Google putting powerful creative tools in more
• 12:02
  hands with the Gemini app's new photo -to -video
• 12:05
  feature using VO3. Makes video creation much
• 12:09
  easier, great for creators and businesses, but
• 12:12
  also raises those important questions about realism
• 12:15
  and media literacy. So across the board, We're
• 12:19
  seeing AI advancing rapidly, fierce competition
• 12:22
  for the experts who build it, and these increasingly
• 12:25
  powerful tools becoming available, changing how
• 12:27
  we create and consume content. It really shows
• 12:30
  how interconnected all this tech development
• 12:32
  is with our daily lives. It definitely does.
• 12:34
  It brings up so many possibilities, but also
• 12:36
  things we need to think carefully about as these
• 12:38
  tools become more common. Absolutely. You know,
• 12:41
  seeing these huge leaps in AI capability and
• 12:44
  the intense focus on building even smarter systems,
• 12:47
  it makes you wonder where is this ultimately
• 12:49
  heading? Are we building tools that will free
• 12:51
  up human creativity for totally new things or
• 12:53
  is the main focus becoming the AI itself? Something
• 12:56
  to think about. We're here to help at Snarfles,
• 12:58
  so reach out on our website if you have any more
• 12:59
  questions. That's a wrap on today's top stories.
• 13:06
  Remember, tech doesn't have to be complicated,
• 13:08
  and we're here to help connect the dots. If you've
• 13:11
  got questions or want to learn more, visit us
• 13:14
  at snarful .com. Thanks for listening, and we'll
• 13:17
  catch you next time.