Latest / Connecting the Dots / Connecting the Dots - 6-19-2025

Transcript

Connecting the Dots - 6-19-2025 · 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. Hello and welcome to Connecting the Dots,
• 0:17
  broadcasting from sunny Sacramento, California.
• 0:20
  I'm Alex. And I'm Morgan. It's great to be with
• 0:22
  you. We're here from Snarfl Solutions Group.
• 0:24
  ready to unpack the tech news that matters. That's
• 0:27
  right. Today is Thursday, June 19th, 2025, Juneteenth
• 0:31
  here in Sacramento, and it's looking like a beautiful
• 0:33
  sunny day, heading for a high of about 87 degrees
• 0:37
  Fahrenheit. Nice weather for the holiday here.
• 0:40
  Nationally though, bit of a different picture.
• 0:42
  Keep an eye out for strong, maybe even severe
• 0:44
  thunderstorms across parts of New England, the
• 0:47
  Mid -Atlantic, and down through the Carolinas
• 0:49
  today. Yeah, and the heat is really building,
• 0:51
  significant heat expanding across the West and
• 0:54
  Central Plains through Friday. And then it moves
• 0:56
  east. Looks like dangerous heat building into
• 0:58
  the eastern U .S. this weekend and likely sticking
• 1:00
  around for much of next week. So stay cool out
• 1:03
  there, folks. Definitely. And a quick look at
• 1:05
  the markets from yesterday's close Wednesday,
• 1:08
  June 18th. The Dow Jones Industrial Average finished
• 1:11
  at forty two thousand one hundred seventy one
• 1:13
  point six six and the S &P 500 closed up at five
• 1:16
  thousand nine hundred and eighty point eight
• 1:17
  seven. Holding strong. OK, so now it's time to
• 1:20
  connect the dots on today's top tech stories.
• 1:23
  We pull these directly from our internal Snarfle
• 1:26
  Tech article report. And we're not just summarizing.
• 1:28
  We want to dig into why these stories matter
• 1:31
  and how they might actually impact you, whether
• 1:33
  you're listening as a professional, running a
• 1:35
  business, or just managing everyday life. Exactly.
• 1:38
  Let's dive into our first major story. This one's
• 1:41
  really at the intersection of the whole AI data
• 1:44
  landscape and some pretty serious geopolitical
• 1:46
  currents. OK. The big headline, Scale AI. It's
• 1:51
  a six -year -old startup just locked down a massive
• 1:54
  $1 billion in new funding. That pushes its valuation
• 1:58
  to, get this, $14 billion. Wow. That's... That
• 2:02
  puts them in a very exclusive club, doesn't it,
• 2:04
  among those big generative AI companies? It absolutely
• 2:07
  does. But here's the twist. Despite this huge
• 2:10
  success, some of their rivals companies like
• 2:12
  Snorkel AI, Labelbox, Turing, they're actually
• 2:14
  seeing a surge in client interest. So interesting.
• 2:16
  So it's not about scale AI stumbling. Not at
• 2:19
  all. It seems to be driven by concerns about
• 2:21
  scale AI's independence, specifically after Meta
• 2:25
  made a significant investment in them. Ah, right.
• 2:28
  So if you're a company competing with Meta, you
• 2:30
  might start getting a little nervous about where
• 2:32
  your critical AI training data is being processed
• 2:35
  and sort of how neutral that process really is.
• 2:38
  Exactly. It really highlights the strategic sensitivity
• 2:41
  around this foundational AI stuff. Sure does.
• 2:44
  And scale AI's core business, just to remind
• 2:46
  everyone, is providing the human workers and
• 2:48
  and the software needed to label and test data.
• 2:51
  This is absolutely critical for training AI models
• 2:53
  well. Critical and booming. Apparently 90 % of
• 2:57
  the revenue is now driven by spending on generative
• 2:59
  AI. Their annual recurring revenue tripled last
• 3:02
  year in 2023. Tripled! Yeah, and they're forecasting
• 3:05
  hitting $1 .4 billion by the end of this year,
• 3:08
  2024, aiming for profitability. It's just explosive
• 3:11
  growth. It really underscores the incredible
• 3:14
  demand for high -quality specialized data, doesn't
• 3:17
  it? Especially for these generative models that
• 3:19
  need so much input to get good. Right, and the
• 3:21
  CEO, Alexander Wang, he started this company
• 3:23
  at 19, dropped out of MIT basically to solve
• 3:27
  the data problem in AI. Quite a vision for a
• 3:30
  19 -year -old. They initially worked with autonomous
• 3:32
  driving companies, right? Cruze, Waymo? He did,
• 3:35
  and they even helped OpenAI train GPT -2 way
• 3:38
  back when. They've been foundational for a while.
• 3:40
  Okay, so what about this shift in how they do
• 3:43
  the work? I heard something about that. Yes,
• 3:45
  that's really fascinating. Initially, they relied
• 3:48
  heavily on tens of thousands of gig workers through
• 3:51
  a subsidiary called Remo Tasks. That division
• 3:55
  did face some criticism for low pay, by the way.
• 3:57
  Okay. But more recently, they've made a big pivot.
• 4:00
  They're hiring highly specialized workers, think
• 4:03
  PhDs, academics, lawyers, even poets, and paying
• 4:06
  them much better, like $30 to $60 an hour through
• 4:09
  different subsidiary calls outlier. So that's
• 4:12
  more than just labeling data, isn't it? It sounds
• 4:14
  like they need deep expertise now. Wang talks
• 4:16
  about giving these experts society -level impact,
• 4:19
  helping with scientific discovery. Exactly. It's
• 4:21
  a fundamental shift. As the AI models get smarter,
• 4:24
  They need more nuanced, context -aware, maybe
• 4:27
  even ethically reviewed input. So it transforms
• 4:29
  that labeling work from, you know, maybe commoditized
• 4:32
  clicks into real intellectual contribution, creating
• 4:35
  a whole new kind of specialized AI workforce.
• 4:38
  A new class of AI whispers, you could say. And
• 4:41
  it's not just commercial work. They're also expanding
• 4:43
  into government and military contracts, like
• 4:45
  with the U .S. Department of Defense. Right.
• 4:47
  And focusing on AI safety and evaluation, too.
• 4:50
  Didn't they run the platform for that DEF CON
• 4:52
  AI security challenge? the White House supported?
• 4:55
  They did. Which leads us neatly into the geopolitical
• 4:57
  side of things. Alexander Wang seems very focused
• 5:00
  on this, partly due to his background growing
• 5:02
  up in Los Alamos. Okay, the birthplace of the
• 5:04
  atomic bomb. That context matters. It seems to.
• 5:08
  He's been meeting with the White House, with
• 5:10
  Congress, warning about the threat from China's
• 5:13
  advancements in AI. He's specifically pointed
• 5:15
  to a Chinese startup called DeepSeek. DeepSeek.
• 5:19
  This is the one Mark and Dressen called AI's
• 5:22
  Sputnik moment, right? Because they claimed that
• 5:24
  they trained an advanced model really cheaply,
• 5:27
  less than six million dollars, and supposedly
• 5:29
  without Nvidia's top chips due to the US export
• 5:32
  controls. That's the one. It set off alarms,
• 5:35
  suggesting maybe China could catch up or even
• 5:37
  leapfrog the US without access to the absolute
• 5:39
  best hardware. But Wang's take is a bit different,
• 5:42
  isn't it? He's saying DeepSeek's model is actually
• 5:45
  roughly on par with the best American models.
• 5:47
  Right. And he suspects they do have advanced
• 5:50
  Nvidia chips, maybe tens of thousands of them,
• 5:53
  obtained somehow despite the export controls,
• 5:55
  and they just can't admit it. And U .S. officials
• 5:58
  are reportedly looking into whether DeepSeek
• 6:00
  did manage to circumvent those restrictions.
• 6:02
  So this isn't just about algorithms. It's about
• 6:04
  hardware access, sanctions, potentially even
• 6:07
  industrial espionage. It's a real strategic battle.
• 6:09
  A battle for AI. supremacy, playing out on multiple
• 6:12
  fronts. Now, contrast that sort of high -stakes,
• 6:16
  potentially covert race with what Meta just did.
• 6:19
  They released Llama 3 .1. It's a powerful new
• 6:22
  AI model, rivals OpenAI and Google's best. But
• 6:26
  they made it free and open source. Ah, the open
• 6:29
  source play. That's meta strategy to shake things
• 6:31
  up, isn't it? By giving away powerful tools,
• 6:33
  they potentially undermine their competitors'
• 6:36
  paid models. Exactly. And they make advanced
• 6:38
  AI much more accessible for startups, researchers,
• 6:42
  anyone, really. They're trying to build an ecosystem,
• 6:44
  maybe something like Android did for mobile phones.
• 6:47
  The argument being that open source leads to
• 6:49
  faster innovation, more eyes on the code for
• 6:51
  security and bias checks, and it prevents just
• 6:54
  a few giant companies from controlling everything.
• 6:57
  That's the upside. But there's always a flip
• 6:59
  side with open access, right? Always. Critics
• 7:02
  are definitely raising concerns about misuse.
• 7:05
  If anyone can download and modify these powerful
• 7:07
  models, well, that includes bad actors, fraudsters,
• 7:10
  state hackers, criminals. Yeah, there have been
• 7:12
  reports about open source AI tools being used
• 7:15
  for harmful purposes, like generating child sexual
• 7:18
  abuse material. This is a serious complex debate
• 7:20
  balancing innovation with safety. Absolutely.
• 7:24
  And meanwhile, Meta is weaving its AI into everything.
• 7:27
  WhatsApp, Instagram, Facebook, Messenger for
• 7:30
  image generation, answering questions, all that
• 7:32
  stuff. OK, so let's connect the dots for everyone
• 7:34
  listening. If you're running a business and thinking
• 7:36
  about AI. Well, you've got a real strategic choice
• 7:38
  now, don't you? Do you go with an established
• 7:40
  service like Scale AI, maybe pay more, navigate
• 7:43
  those independence concerns? Or do you explore
• 7:46
  the open source route with something like Llama
• 7:48
  3 .1? That could mean lower costs, more flexibility.
• 7:52
  But also maybe more risk and needing more in
• 7:55
  -house expertise It really impacts your whole
• 7:58
  innovation approach and for professionals the
• 8:00
  nature of AI work itself is clearly changing
• 8:03
  It's moving beyond basic data tasks to needing
• 8:06
  deep specialized expertise, right? So new opportunities
• 8:09
  are opening up, but they demand higher level
• 8:11
  skills critical thinking It's redefining what
• 8:14
  being an AI professional even means and finally
• 8:16
  for everyone listening this intense US China
• 8:19
  AI competition plus the whole open versus closed
• 8:22
  model debate. It has huge implications. It's
• 8:25
  about global security, economic power, who sets
• 8:29
  the standards for this incredibly powerful technology,
• 8:31
  and ultimately how it shapes the devices and
• 8:34
  services you use every single day. Well, it's
• 8:36
  a big deal. It really is. OK, let's shift gears
• 8:39
  now to health care. This is an area where A .I.'s
• 8:42
  impact could be incredibly positive and very
• 8:46
  personal. Definitely. The big news here is about
• 8:49
  a company called Cameer. They provide what's
• 8:51
  called ambient AI plus tools for revenue cycle
• 8:54
  management and workflow automation in health
• 8:56
  care. And they just raised $200 million from
• 8:59
  General Catalyst. That's a significant investment.
• 9:01
  It is. And it speaks to the potential here. Comeer's
• 9:04
  whole goal with their ambient AI platform is
• 9:06
  to streamline all that background stuff, data
• 9:08
  collection, documentation, coordinating care.
• 9:11
  Basically handling the admin burden that doctors
• 9:13
  and nurses always talk about. Exactly. The aim
• 9:16
  is to cut down on that administrative time so
• 9:18
  healthcare providers can spend more actual time
• 9:20
  focusing on the patient, fostering a more human
• 9:22
  connection. Imagine the AI just quietly handling
• 9:25
  the notes and the background. That sounds like
• 9:28
  a dream for so many clinicians dealing with burnout.
• 9:31
  All those hours spent charting. It's a huge factor
• 9:34
  in burnout, so this technology tackles that head
• 9:36
  -on. And it's not just theoretical, Commure is
• 9:39
  deploying this with Tenant Physician Resources,
• 9:41
  which is a huge network. With thousands of providers,
• 9:44
  right? And they've ported with APCA Healthcare
• 9:46
  too. Yes, another major player. And Commure has
• 9:50
  been aggressive in building out their capabilities.
• 9:53
  They recently bought Augmetics, an AI note -taking
• 9:56
  company, for $139 million. Okay. And they also
• 10:01
  merged with a Thelus, which did something similar.
• 10:03
  So they're consolidating these AI -driven clinical
• 10:06
  documentation tools into their platform. Tens
• 10:09
  of thousands of physicians are already using
• 10:10
  their tech across dozens of health systems. So
• 10:13
  connecting these dots. For healthcare professionals
• 10:15
  listening, this could genuinely change your workday.
• 10:18
  Less screen time, less after hours charting,
• 10:21
  more time for actual patient care, maybe even
• 10:23
  better decision making with more complete records.
• 10:26
  That's the promise. Less burnout, potentially
• 10:28
  better care, and for patients and families. You
• 10:31
  might get a doctor who feels less rushed, who's
• 10:33
  more present and focused on you during the visit,
• 10:36
  not just typing into a computer. Which can lead
• 10:38
  to better communication, more trust, potentially
• 10:41
  better health outcomes. It brings the focus back
• 10:44
  to the human element of care. And for healthcare
• 10:46
  businesses, hospitals, clinics. AI like this
• 10:49
  is shifting from a nice -to -have to a critical
• 10:52
  investment. You see Tenet increasing their spending
• 10:55
  on AI. It's about efficiency, tackling workforce
• 10:58
  shortages, managing costs, optimizing the entire
• 11:01
  operation. But it's not all smooth sailing, right?
• 11:03
  There must be concerns. Oh, absolutely. Crucial
• 11:06
  things to watch. is paramount. You can't have
• 11:09
  errors in medical records. Bias in the AI models
• 11:12
  is a huge concern. Could it lead to unequal care?
• 11:15
  And cybersecurity. Introducing more connected
• 11:17
  tech always brings new vulnerabilities, especially
• 11:20
  with sensitive patient data. Precisely. And it's
• 11:23
  a competitive space, too. Microsoft, Amazon,
• 11:25
  Oracle, the tech giants are all investing heavily
• 11:29
  in AI for health care, particularly these AI
• 11:31
  scribe type tools. Comeer has its work cut out
• 11:35
  for it, but the potential impact is undeniable.
• 11:37
  Okay, one more major story to connect the dots
• 11:40
  on today. Shifting to the world of IT management.
• 11:43
  Open source is making waves here too. Right.
• 11:45
  The news is that Fleet, a company known for its
• 11:48
  open source device management tools, just raised
• 11:51
  a $27 million Series B round. Led by 10 -11 ventures.
• 11:56
  So what makes Fleet stand out? Their unified
• 11:59
  approach. They offer a single open source platform
• 12:02
  to manage basically all your company's devices.
• 12:05
  Apple Macs, Windows PCs, Linux machines, and
• 12:08
  now, crucially, iPhones and iPads too. Wow, okay.
• 12:12
  So instead of using, say, Jam for Apple, Microsoft,
• 12:16
  Intune for Windows, maybe something else for
• 12:17
  Linux, an IT department could potentially manage
• 12:20
  everything through Fleet. That's the idea. Simplify
• 12:23
  the whole thing. Reduce the number of tools you
• 12:25
  need to buy, learn, and maintain. cut down on
• 12:28
  complexity and costs. That sounds incredibly
• 12:30
  appealing for IT teams. Are big companies using
• 12:32
  it? They are. Companies like Snowflake, Uber,
• 12:35
  Atlassian, even government agencies are using
• 12:37
  Fleet, which suggests it's robust enough for
• 12:40
  demanding environments. And they have this interesting
• 12:42
  GitOps approach. What's that about? Yeah, it's
• 12:44
  clever. They treat device management like software
• 12:47
  development. You use version control like you
• 12:49
  would with code in GitHub to manage device configurations.
• 12:52
  OK, so you can track changes, roll back if something
• 12:55
  goes wrong. have a clear audit trail. Exactly.
• 12:59
  Much greater visibility, control, and auditability.
• 13:02
  That's huge for compliance and security. And
• 13:04
  speaking of security, they offer an open API
• 13:07
  for every endpoint. How does that help? It's
• 13:09
  a big boost for Zero Trust security strategies.
• 13:13
  Zero Trust requires constant verification of
• 13:16
  users and devices. An open API lets IT teams
• 13:20
  build custom integrations and enforce security
• 13:23
  policies much more deeply and consistently across
• 13:25
  all those different devices. Makes sense. The
• 13:28
  CEO, Mike McNeil, has this vision about open
• 13:31
  source enabling future generations to trust and
• 13:33
  understand the code, right? Empowering organizations,
• 13:36
  avoiding vendor lock -in. It's a powerful vision.
• 13:38
  It's about giving control back to the organizations
• 13:41
  themselves. Okay, let's connect these dots. For
• 13:43
  IT professionals and businesses listening, especially
• 13:45
  if you're juggling Macs, PCs, Linux, maybe mobile
• 13:48
  devices. This could be a major simplification.
• 13:51
  Consolidate tools, potentially lower costs, reduce
• 13:54
  the need for platform -specific training. It
• 13:57
  offers more flexibility than many proprietary
• 13:59
  systems. And for security teams. The open source
• 14:02
  transparency and that GitOps approach are gold.
• 14:05
  Easier auditing, better control over configurations,
• 14:08
  which is exactly what you need for robust security
• 14:11
  and zero trust. So the bigger picture here is?
• 14:14
  It's about democratizing advanced IT management.
• 14:17
  making enterprise -grade control and security
• 14:19
  more accessible even for smaller organizations
• 14:22
  or those without huge budgets for proprietary
• 14:25
  software suites. It challenges that old model
• 14:28
  of just buying more and more software, offering
• 14:30
  a more flexible, community -driven alternative.
• 14:33
  Really interesting shift. That brings us towards
• 14:35
  the end. Let's quickly recap the main threads
• 14:38
  we pulled today. We saw the high -stakes world
• 14:40
  of AI data with scale AI's massive funding and
• 14:43
  their shift to specialized work, contrasted with
• 14:45
  the U .S.-China tech race involving DeepSeq and
• 14:48
  Meta's disruptive open -source Llama 3 .1 strategy.
• 14:51
  Then we looked at AI transforming health care
• 14:54
  through companies like Cameer, aiming to reduce
• 14:56
  clinician burnout and improve the patient experience
• 14:59
  while navigating important ethical and security
• 15:02
  considerations. And finally, how open source
• 15:04
  through platforms like Fleet is simplifying and
• 15:07
  improving IT device management, offering a unified,
• 15:11
  transparent alternative to proprietary systems.
• 15:14
  These aren't just separate stories. They show
• 15:16
  technology reshaping industries, creating new
• 15:19
  jobs, new challenges, new ways of working and
• 15:22
  living. So a final thought to leave you with.
• 15:24
  As these technologies develop, think about who
• 15:27
  controls the underlying data and the platforms.
• 15:29
  Is it open? Is it closed? What values are embedded
• 15:33
  in the code? Because that will increasingly shape
• 15:35
  transparency, trust, and maybe even power balances
• 15:38
  in the years ahead. Lost to think about. Thanks
• 15:40
  for joining us today. We're here to help it snarfle,
• 15:43
  so reach out on our website if you have any more
• 15:44
  questions. That's a wrap on today's top stories.
• 15:51
  Remember, tech doesn't have to be complicated,
• 15:53
  and we're here to help connect the dots. If you've
• 15:56
  got questions or want to learn more, visit us
• 15:59
  at snarful .com. Thanks for listening, and we'll
• 16:02
  catch you next time.