Latest / Connecting the Dots / Connecting the Dots - 7-17-2025

Transcript

Connecting the Dots - 7-17-2025 · 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. Hey, everyone, and welcome back to Connecting
• 0:17
  the Dots. I'm Alex. And I'm Morgan. We're coming
• 0:20
  to you from Snarful Solutions Group right here
• 0:22
  in Sacramento. It is Thursday, July 17th, 2025,
• 0:26
  and we're ready to connect some dots for you
• 0:28
  on today's biggest tech stories. That's right.
• 0:31
  But first, let's get a quick pulse check on what's
• 0:33
  happening outside the tech bubble. Weather first.
• 0:36
  Sounds good. Nationally, things are pretty active.
• 0:39
  We've got heavy rain, potential flash floods
• 0:41
  along the central Gulf Coast. That's sticking
• 0:43
  around into Saturday. OK, so folks down there
• 0:45
  need to stay alert. Definitely. And then widespread
• 0:49
  showers and storms, pretty much from the central
• 0:52
  plains moving east. through the mid Mississippi
• 0:55
  Valley, the Ohio Valley, right into the central
• 0:58
  Appalachians. Sounds like a soggy middle of the
• 1:01
  country. Yeah, a bit. And meanwhile, it's the
• 1:03
  opposite problem further south and east. Dangerous
• 1:06
  heat is expected from the lower Mississippi Valley
• 1:08
  all the way to the mid Atlantic. So stay cool.
• 1:12
  Stay hydrated if you're out there. Good advice.
• 1:14
  And what about here at home in Sacramento? Any
• 1:16
  relief? Actually, yeah. Looks like a beautiful
• 1:19
  day shaping up for us. Sunny skies high around
• 1:21
  88 degrees Fahrenheit. Pretty nice July day.
• 1:24
  Okay, I'll take that. Much better than the heat
• 1:26
  dome further east. Alright, quick market check.
• 1:28
  Let's do it. So looking back at Wednesday's close,
• 1:30
  July 16th, the Dow Jones finished at 44 ,254
• 1:35
  .78. Okay, and how did it do for the day? It
• 1:38
  was up slightly, gained about 102 points. That's
• 1:41
  a 0 .23 % increase. And the week so far? since
• 1:44
  monday looking at the week since monday the 14th
• 1:46
  it's actually down a little bit off by about
• 1:48
  91 points or 0 .21 percent so kind of shredding
• 1:52
  water gotcha how about the s &p 500 s &p also
• 1:55
  closed slightly up on wednesday finished at 6
• 1:58
  263 .70 that was a gain of just over nine points
• 2:02
  of 0 .15 percent for the day in its weekly picture
• 2:04
  a little better than the dow since monday the
• 2:06
  s &p is actually up about eight and a half points
• 2:08
  or 0 .14 percent so slight gain there interesting
• 2:11
  and the crypto world Bitcoin Bitcoin always moving
• 2:14
  as of this morning Thursday the 17th It's trading
• 2:16
  around a hundred and seventy thousand seven hundred
• 2:18
  and seventy point six two and the recent trend
• 2:20
  down slightly today off about nine hundred and
• 2:22
  fifty bucks or point eight zero percent and looking
• 2:25
  back since Monday It's down about two thousand
• 2:27
  eighty three dollars, which is a dip of one point
• 2:30
  seven four percent So a bit of a pullback this
• 2:32
  week. Okay, always volatile that space and finally
• 2:36
  any action from her San Francisco Giants last
• 2:38
  night Nope, no game for the Giants on Wednesday,
• 2:40
  July 16th. They had the night off Yeah, OK. A
• 2:43
  little breather for them, then. Yep. All right,
• 2:45
  let's shift gears. Time to dive into the big
• 2:46
  tech stories making waves today, pulled straight
• 2:49
  from our Morningstar report. Let's get into it.
• 2:51
  And this first one, it definitely raised some
• 2:53
  eyebrows. We're talking about Scale AI. Right.
• 2:56
  So Scale AI, big name in the AI data labeling
• 2:59
  and model training space. They just announced
• 3:02
  layoffs about 200 full -time employees, which
• 3:05
  is what 14 % of their staff Yeah around 14 %
• 3:08
  of the 1 ,400 people and on top of that they
• 3:10
  let go of 500 contractors That's that's a significant
• 3:14
  cut. It really is and what makes this so I guess
• 3:17
  Surprising is the timing this comes just one
• 3:20
  month after Meta poured a huge amount of money
• 3:22
  into the company Huge is right. We're talking
• 3:24
  fourteen point eight billion dollars and it wasn't
• 3:27
  just cash, right? Metta took a 49 % stake. Exactly.
• 3:31
  A nearly half stake in the company. Yeah. And
• 3:33
  they also hired Scales founder and former CEO
• 3:37
  Alexander Wang. Yep. Wang's now heading up Metta's
• 3:41
  new super intelligence group. So you have this
• 3:43
  massive investment, this strategic hiring of
• 3:45
  the founder, and then boom, significant layoffs.
• 3:48
  It feels counterintuitive. Totally. It's like
• 3:51
  buying a championship sports team and immediately
• 3:53
  benching your star players. So what's the official
• 3:55
  reason? Well, the interim CEO, Jason Droge, sent
• 3:58
  out an email. He basically said it's a restructuring
• 4:00
  of their generative AI organization. Restructuring?
• 4:03
  Okay, meaning what exactly? He cited, let me
• 4:05
  get the quote right, having ramped up their generative
• 4:08
  AI capabilities too quickly, which apparently
• 4:11
  led to, you know... the usual suspects, inefficiencies
• 4:13
  and redundancies, and even excessive bureaucracy.
• 4:16
  Ah, the classic growing pains explanation after
• 4:18
  hypergrowth. So they're trying to slim down the
• 4:21
  Gen. AI part of the business. Yeah, seems like
• 4:23
  it. George mentioned that they had something
• 4:25
  like 16 internal pods or project teams within
• 4:28
  Gen. AI. The plan is to cut that down to just
• 4:31
  five. Five core areas. What are they? The email
• 4:35
  listed them as code. languages, experts, experimental,
• 4:39
  and audio. So focusing on what they see as the
• 4:41
  most impactful areas. Okay, so maybe it is more
• 4:44
  of a strategic shift than, say, a sign of financial
• 4:48
  distress. They did just get billions from Metta.
• 4:51
  That's certainly the company line. They're emphasizing
• 4:53
  they're still well -funded and actually plan
• 4:56
  to hire hundreds more employees, but in other
• 4:59
  areas like enterprise and government AI. But
• 5:02
  it still seems pretty rough for the employees
• 5:04
  affected. Reports are saying they were just locked
• 5:06
  out of systems, Slack deactivated, kind of overnight.
• 5:09
  Yeah, that part sounds abrupt. Definitely not
• 5:11
  ideal. The Severance package details are out
• 5:13
  there. They're paid through September 15th, plus
• 5:15
  at least four more weeks if they sign the agreement.
• 5:17
  Okay. Still, you got to wonder about the bigger
• 5:19
  picture here. The founder leaves. A major investor
• 5:23
  takes a huge stake. Some big clients like Google
• 5:26
  reportedly paused projects after the meta deal.
• 5:28
  Right, and there have been some questions swirling
• 5:30
  about their security practices around training
• 5:32
  data too. It's a lot happening all at once. It
• 5:35
  really makes you think about the pressures of
• 5:37
  scaling so fast in this AI gold rush. Even with
• 5:41
  massive funding, maybe hitting the brakes or
• 5:44
  restructuring becomes necessary if things get
• 5:46
  unwieldy. Remember, these guys work on huge models
• 5:49
  like XAI Scrock. Google's Gemini. Exactly. It's
• 5:53
  not just an internal shuffle. It potentially
• 5:54
  impacts the wider AI ecosystem, too. So, yeah,
• 5:58
  a story worth watching. It's a reminder that
• 6:00
  huge investment doesn't automatically equal smooth
• 6:02
  sailing, especially in a field moving this quickly.
• 6:05
  Absolutely. OK, let's move from AI growing pains
• 6:08
  to, well, digital growing pains of a different
• 6:10
  sort. Crypto theft. Yeah. And the numbers for
• 6:13
  the first half of this year, 2025, are. Honestly,
• 6:17
  pretty alarming. How alarming are we talking?
• 6:19
  We're looking at somewhere between $2 .17 billion
• 6:21
  and $2 .47 billion worth of crypto stolen just
• 6:25
  in the first six months. Wait, say that again?
• 6:27
  Over two billion? Yep. And get this, that's already
• 6:29
  more than the total amount lost in all of 2024.
• 6:32
  Wow. Okay, that is staggering. What on earth
• 6:36
  is driving such a massive spike? Is it just more
• 6:40
  attacks or are they bigger? It's mostly the latter.
• 6:43
  The Surge was really driven by two enormous incidents.
• 6:46
  You had the Bybit breach. I remember hearing
• 6:48
  about that one. How much was that again? Estimated
• 6:50
  between 1 .4 and 1 .5 billion dollars. Just that
• 6:54
  one incident. It happened back in February at
• 6:57
  the Dubai -based exchange. Unbelievable. A billion
• 7:00
  and a half. And who's suspected? The usual suspects,
• 7:03
  unfortunately. The Ladris Group, that North Korean
• 7:06
  state -sponsored hacking collective, is believed
• 7:09
  to be behind it. If those numbers hold, it's
• 7:11
  the single largest crypto theft ever. And what
• 7:16
  was the other big one? That was the Cetus Protocol
• 7:18
  incident. About 225 million dollars stolen there.
• 7:21
  That happened more recently in May on the Sui
• 7:23
  blockchain. Okay, so those two alone account
• 7:25
  for a massive chunk of that two billion dollars
• 7:27
  plus total. Exactly. Without those two mega hacks,
• 7:30
  the overall losses would look quite different.
• 7:32
  But it does show the extreme vulnerability when
• 7:34
  these large platforms get hit. So even if the
• 7:37
  number of smaller hacks isn't necessarily exploding,
• 7:40
  the potential damage from a single successful
• 7:42
  attack on a big target is just enormous. Precisely.
• 7:46
  And another interesting data point, the average
• 7:49
  loss per incident in the first half of this year
• 7:51
  was $7 .18 million. Okay. How does that compare
• 7:55
  to last year? It's more than double. Last year
• 7:57
  the average loss per incident was around $3 .1
• 8:00
  million. So the hacks that are happening are
• 8:03
  becoming much more costly on average. And what's
• 8:05
  the main way thieves are getting in? Is it sophisticated
• 8:07
  platform hacking or something else? The report
• 8:11
  highlights wallet compromise as the single costliest
• 8:14
  attack vector, meaning hackers getting hold of
• 8:17
  individual users private keys or those secret
• 8:19
  recovery phrases. Once they have those, they
• 8:22
  can just drain the wallet directly. It's less
• 8:24
  about cracking the exchanges main vault and more
• 8:27
  about targeting individual users credentials.
• 8:29
  So it really underscores the personal responsibility
• 8:32
  aspect of crypto security. It's not just about
• 8:34
  trusting the platform. You have to guard your
• 8:36
  own keys like. Well, like the keys to billions
• 8:39
  of dollars sometimes. Exactly. It's the digital
• 8:42
  equivalent of leaving your bank account password
• 8:44
  written on a sticky note. There's a bit of a
• 8:47
  slowdown in the volume of hacks in the second
• 8:48
  quarter compared to the first. Q2 saw about $465
• 8:52
  million lost versus Q1's $1 .77 billion. Okay,
• 8:57
  so a drop off. Is that good news? Better security
• 9:00
  maybe? It's hard to say definitively. Could be
• 9:03
  better defenses, could be what some call attacker
• 9:06
  fatigue, or maybe just luck. The bottom line
• 9:08
  is the risk is still incredibly high. So for
• 9:12
  platforms and users, What's the takeaway? What
• 9:15
  needs to happen? For platforms, it's about a
• 9:17
  multi -layered approach. You know, constant code
• 9:19
  audits, real -time monitoring for suspicious
• 9:21
  activity, having solid incident response plans,
• 9:24
  and crucially, training employees not to fall
• 9:26
  for phishing scans. Right, the human element
• 9:28
  again. Always. And for individual users, it's
• 9:31
  hammering home the absolute necessity of securing
• 9:33
  those private keys and recovery phrases. Use
• 9:36
  hardware wallets, be skeptical of everything,
• 9:38
  enable multi -factor authentication, basic digital
• 9:41
  hygiene, but super critical. Makes sense. Was
• 9:44
  there any good news in all this? Did any of the
• 9:46
  stolen funds get recovered? Actually, yes. In
• 9:48
  the CEDA's protocol case, the validators on the
• 9:51
  SUI blockchain apparently managed to freeze and
• 9:53
  eventually return about 162 million dollars of
• 9:57
  the 225 million dollars stolen. Oh wow. That's
• 10:01
  significant and pretty rare in these situations,
• 10:03
  right? Very rare, but definitely a positive development
• 10:05
  still you have other major losses lingering out
• 10:08
  there like that hundred million dollar rug pull
• 10:10
  at Libra right where the developers just vanished
• 10:12
  with the money. Yep, and 90 million dollars from
• 10:15
  Nobitex 70 million dollars from UPC X the list
• 10:19
  goes on it's still very much a Wild West environment
• 10:22
  in some corners of the crypto space Definitely
• 10:24
  sounds like it Okay, for our final story today,
• 10:26
  let's shift from the risks to, well, a surprising
• 10:29
  pocket of intense innovation. Yeah, we're talking
• 10:32
  about the generative AI scene, but focusing on
• 10:34
  a specific country that's really punching above
• 10:36
  its weight, Israel. Israel's startup nation reputation
• 10:39
  is well known, but what's happening specifically
• 10:41
  in GenEI? The numbers are kind of startling.
• 10:44
  As of 2025, there are reportedly 342 Israeli
• 10:48
  startups focused specifically on generative AI.
• 10:51
  342. Wow. How does that compare to previously?
• 10:54
  It's a huge jump. Reports from around May 2024
• 10:57
  put the number somewhere between 67 and 144.
• 11:01
  So in about a year, the number has potentially
• 11:03
  more than doubled, maybe even quintupled, depending
• 11:05
  on which baseline you use. That's explosive growth.
• 11:08
  Is there money following this boom? Oh, yeah.
• 11:11
  These startups collectively have raised over
• 11:13
  20 billion dollars to date. One older report
• 11:16
  mentioned they'd already pulled in 2 .3 billion
• 11:19
  dollars way back by September. 2021. So the momentum
• 11:23
  has been building for a while and it's accelerating.
• 11:25
  So what's driving this? Why Israel and why now
• 11:28
  for Gen. AI? Well, it builds on Israel's existing
• 11:32
  strength. You've got deep expertise in areas
• 11:34
  that are fundamental to AI, like deep learning,
• 11:37
  computer vision, natural language processing.
• 11:39
  Right. A lot of that comes from their strong
• 11:40
  tech sector and even the military, the IDF, right?
• 11:43
  That pipeline of technical talent. Exactly. There's
• 11:46
  a historical strength in cybersecurity, which
• 11:48
  has a lot of overlap with AI development and
• 11:50
  just a really dense ecosystem of engineers and
• 11:52
  entrepreneurs. Plus you have venture capital
• 11:54
  firms like Reimagine Ventures, as one mentioned,
• 11:57
  actively fueling this specific gen AI growth.
• 12:00
  It's fascinating how a relatively small country
• 12:03
  consistently generates so much tech innovation
• 12:06
  like that success breed, success idea and action.
• 12:09
  We saw it with navigation like Waze, now AI.
• 12:12
  It really is. And a lot of Israeli tech, especially
• 12:15
  in AI, tends to focus on B2B business -to -business
• 12:18
  and enterprise solutions rather than just consumer
• 12:21
  apps. They're building tools for data -heavy
• 12:23
  industries. Like what kinds of industries? Entertainment
• 12:26
  is a big one mentioned. Reimagine Ventures, for
• 12:29
  example, has backed companies like Site. They
• 12:32
  do image recognition for online retail, find
• 12:34
  similar products just from a photo. Okay, I've
• 12:36
  seen tech like that. What else? There's Hour
• 12:38
  One. They create synthetic video using Jan's
• 12:41
  generative adversarial networks. Basically, AI
• 12:43
  creating realistic video presenters or characters
• 12:46
  without needing cameras or real people. Whoa.
• 12:49
  Okay, that brings up the whole deep fake concern,
• 12:51
  doesn't it? Yeah. How are companies like HoraOne
• 12:53
  handling the epics of that? That's a really important
• 12:56
  point. The report mentions Huron specifically
• 12:58
  committing to avoiding problematic uses, like
• 13:00
  using their tech for political campaign videos.
• 13:03
  So... There seems to be an awareness of the potential
• 13:05
  downsides from the start, at least for some players.
• 13:08
  That's somewhat reassuring, building in ethical
• 13:10
  guardrails early on. Any other examples? Yeah,
• 13:13
  another one in Remagin's portfolio is called
• 13:15
  Vault. They use deep learning to try and predict
• 13:17
  how well a film will do things like box office
• 13:20
  revenue, even Rotten Tomatoes scores, applying
• 13:23
  AI to creative decision making. Interesting.
• 13:25
  Using data to predict art's success. And beyond
• 13:29
  that specific VC firm, are there other big Israeli
• 13:33
  gen -AI names making waves? Absolutely. You've
• 13:37
  got companies like AI21 Labs, which is apparently
• 13:39
  valued at $1 .4 billion now. Lighttricks is another
• 13:43
  one focused on visual media generation tools.
• 13:46
  And taste -wise, using AI to analyze trends for
• 13:49
  the food industry, it's a really diverse landscape.
• 13:51
  It sounds incredibly dynamic. And that connection
• 13:54
  to cybersecurity strength, does that position
• 13:56
  them well for tackling things like detecting
• 13:58
  AI -generated fakes? You'd think so. The expertise
• 14:01
  is definitely there. The question might be, who
• 14:03
  takes on that task at scale? Is it the AI companies
• 14:07
  themselves, or does the burden fall more on the
• 14:10
  big social media platforms where this content
• 14:12
  spreads? Good question. Probably the platforms,
• 14:15
  ultimately, given the volume and the potential
• 14:17
  impact on things like advertising. If brands
• 14:20
  end up next to harmful deepfakes. Could be. But
• 14:22
  yeah, the bottom line is Israel has rapidly become
• 14:24
  a major global hub for generative AI development
• 14:27
  built on its unique mix of talent, investment,
• 14:30
  and existing tech strengths. OK, so quite a journey
• 14:33
  today. We started with those surprising layoffs
• 14:35
  at scale AI right after that huge meta investment.
• 14:39
  Then dove into the really concerning surge in
• 14:41
  crypto theft, topping $2 billion already this
• 14:44
  year, driven by massive breaches like Bybit.
• 14:47
  and wrapped up with the incredible boom in Israel's
• 14:50
  generative AI startup scene, just showcasing
• 14:52
  innovation happening at lightning speed. And
• 14:55
  these aren't just abstract tech stories, right?
• 14:57
  They touch on job security for people working
• 15:00
  in AI, the safety of your digital money if you're
• 15:03
  involved in crypto, and just the sheer pace of
• 15:06
  technological change globally that affects businesses
• 15:09
  and, well, all of us. Absolutely. It really makes
• 15:12
  you step back and think. Are we seeing necessary
• 15:15
  corrections and maturing in these fast -moving
• 15:17
  tech sectors, or is this maybe just a pause before
• 15:20
  things get even more disruptive? That's the big
• 15:22
  question hanging over all of this. How this all
• 15:25
  shakes out will really shape our digital future,
• 15:27
  impacting everything from innovation pipelines
• 15:29
  to how secure we feel online. Lots to keep watching,
• 15:32
  and remember, we're here to help it snarfle,
• 15:34
  so reach out on our website if you have any more
• 15:36
  questions. That's a wrap on today's top stories.
• 15:43
  Remember, tech doesn't have to be complicated,
• 15:45
  and we're here to help connect the dots. If you've
• 15:48
  got questions or want to learn more, visit us
• 15:50
  at snarful .com. Thanks for listening, and we'll
• 15:54
  catch you next time.