Latest / Connecting the Dots / Supply Chain Attacks, Billion-Dollar Wearables, and Space Risks

Transcript

Supply Chain Attacks, Billion-Dollar Wearables, and Space Risks · 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. Welcome to Connecting the Dots, brought
• 0:17
  to you by Snarful Solutions Group. I am Alex.
• 0:20
  And I'm Morgan. And today is Tuesday, March 31st.
• 0:24
  2026. We've got a really packed show for you
• 0:27
  today exploring our latest Morningstar tech report.
• 0:31
  Yeah, we are connecting a lot of dots today from
• 0:33
  software all the way up to space. Exactly. But
• 0:36
  before we get into the heavy tech news, let's
• 0:38
  just set the stage for the morning. If you're
• 0:40
  traveling today, you know, maybe double check
• 0:42
  your flights. There are currently about 306 active
• 0:47
  weather alerts across the U .S. Right, which
• 0:49
  is quite a bit of activity for a Tuesday morning.
• 0:51
  Yeah, though honestly, here in Sacramento, we
• 0:53
  can't complain. It's sitting at a really mild
• 0:55
  65 degrees with just some light rain. Not bad
• 0:58
  at all. And well, the financial markets are certainly
• 1:01
  looking bright today, too. Oh, absolutely. The
• 1:03
  Dow closed up over 266 points yesterday, and
• 1:07
  the S &P is climbing right alongside it. Plus,
• 1:09
  if you're holding crypto, Bitcoin is sitting
• 1:11
  really comfortably over $67 ,000 right now. Nice.
• 1:15
  Love to see those green arrows. And just quickly
• 1:17
  on the sports front, the Giants managed to squeeze
• 1:19
  out a 3 -2 victory over the Padres last night.
• 1:22
  Yeah, it was a close one. It really was. Sadly,
• 1:25
  the A's took a pretty tough 4 -0 loss to Atlanta.
• 1:28
  But you know, you win some, you lose some. That's
• 1:30
  baseball for you. Right. So before we jump into
• 1:33
  our first story, we want to thank our sponsor,
• 1:35
  Strong Coffee. It is the premium instant coffee
• 1:39
  that tastes like a traditional brew, but is packed
• 1:41
  with 15 grams of protein. MCTs and adaptogens
• 1:45
  like Oshawa Gonda to give you clean, sustained
• 1:48
  energy without the jitters. Which we definitely
• 1:51
  need for parsing through all this data today.
• 1:53
  Oh, for sure. And promo code SNARFL gets you
• 1:55
  20 % off your entire purchase at strongcoffeecompany
• 1:58
  .com. How do you recommend it? So let's jump
• 2:01
  straight into the Morningstar report. Because
• 2:03
  while the financial markets are doing great,
• 2:05
  the global software supply chain is, well, it's
• 2:09
  basically flashing bright red right now. Yeah,
• 2:10
  bright red is an understatement. We are looking
• 2:12
  at a massive compromise involving a project called
• 2:15
  Axios. Right, Axios. Now, for anyone who doesn't
• 2:18
  write JavaScript for a living, and you probably
• 2:20
  haven't heard of it, but I guarantee your digital
• 2:22
  life relies on it entirely. Absolutely. I always
• 2:24
  think of it like a universal translator between
• 2:26
  your phone and every single database on Earth.
• 2:29
  Like, when your weather app needs data from a
• 2:31
  server, Axios is the courier making that request.
• 2:34
  That is a really great way to conceptualize it.
• 2:36
  Yeah. I mean, the sheer volume is just staggering.
• 2:39
  Axios sees roughly 100 million weekly downloads
• 2:43
  on the node package manager registry. Wow, 100
• 2:46
  million. A week? Every single week, yeah. It
• 2:50
  is embedded in corporate dashboards, mobile apps,
• 2:52
  banking software. Pretty much everything modern
• 2:55
  relies on this library. Which obviously makes
• 2:57
  it the ultimate target for an attacker. Oh, exactly.
• 3:00
  If you compromise the translator, you don't even
• 3:02
  need to hack a bank's heavily guarded database.
• 3:05
  You just tell the translator to hand over the
• 3:07
  keys. Right, but the tricky part is, attacking
• 3:09
  a project that massive directly on a platform
• 3:12
  like GitHub, well, that would set off every alarm
• 3:15
  bell in the cybersecurity industry immediately.
• 3:18
  So how did they actually pull this off without
• 3:21
  triggering those alarms? I mean, dropping poison
• 3:24
  directly into the internet's water supply seems
• 3:27
  like it should be noticeable. You'd think so,
• 3:29
  but they bypassed the code entirely and went
• 3:32
  after the human element. Security researchers
• 3:35
  found that the attackers hijacked the account
• 3:37
  of the project's primary maintainer. Wait, they
• 3:40
  just went after the developer directly? Yeah,
• 3:43
  a developer named Jason Saman. They somehow managed
• 3:46
  to swap his registered email address to an anonymous
• 3:49
  proton mail account. Oh wow, so they completely
• 3:52
  locked him out of his own project. Completely
• 3:54
  locked him out. But OK, even if I steal someone's
• 3:57
  digital identity, massive enterprise open source
• 4:00
  projects have automated security guards, right?
• 4:03
  Like you can't just push bad code without a system
• 4:05
  checking it. Well, you're talking about the CI
• 4:07
  CD pipeline, the continuous integration and continuous
• 4:10
  deployment systems. Right, exactly. Don't those
• 4:12
  automated tests usually catch this stuff before
• 4:14
  it goes live? They usually do, yeah. Developers
• 4:17
  rely on them heavily. But because the attackers
• 4:20
  controlled the publisher's actual credentials,
• 4:22
  they just sidestepped that automated GitHub pipeline
• 4:25
  completely. Unbelievable. Yeah, they use a command
• 4:28
  line interface to manually force two compromised
• 4:30
  versions directly into the registry. Specifically,
• 4:34
  versions 1 .14fame .1 and .30 .4. And what really
• 4:39
  stands out to me in the report is the speed of
• 4:41
  this whole thing. They hit two different release
• 4:44
  branches within 39 minutes of each other. Which
• 4:46
  is incredibly fast. Right. So that means any
• 4:49
  corporate system set to, you know, automatically
• 4:52
  update its minor versions, it just pulled the
• 4:55
  poison till immediately. No human ever even reviewed
• 4:58
  it. Precisely. And the payload itself, I mean,
• 5:00
  it was a master class in deception. They didn't
• 5:02
  alter the actual Axios code at all, which definitely
• 5:05
  would have been noticed. So what did they do?
• 5:07
  They just added a single seemingly boring dependency
• 5:10
  to the package manifest. It was a library called
• 5:13
  plaincrypto .js version 4 .2 .1. Ah, okay. So
• 5:17
  that's classic type of squatting right there.
• 5:19
  It sounds exactly like a perfectly legitimate
• 5:21
  cryptography library that you would expect to
• 5:24
  see in a project like this. Yeah, it completely
• 5:27
  exploits the cognitive load on developers. If
• 5:29
  you are scanning a list of 50 different dependencies,
• 5:33
  plain crypto JS looks completely normal. Right,
• 5:35
  your brain just skips right over it. Exactly.
• 5:38
  Researchers noted it was staged about 18 hours
• 5:40
  in advance by an attacker using the email nrys
• 5:43
  at proton .me. But it's the obfuscation inside
• 5:47
  this fake library that is just truly remarkable.
• 5:50
  I want to dig into that mechanics of it, actually,
• 5:53
  because simply scrambling or minifying code usually
• 5:56
  still looks really suspicious to a good static
• 5:59
  analysis tool, right? It does, yeah. So how did
• 6:01
  this slip by entirely undetected by the enterprise
• 6:04
  security scanners? Well, standard security tools
• 6:06
  scan for known malware signatures. They look
• 6:09
  for suspicious IP addresses or recognizable malicious
• 6:12
  commands. To beat that, the attackers used a
• 6:15
  two -layer custom encoding scheme inside a script
• 6:17
  called setup .js. Okay, so they hid the actual
• 6:20
  instructions. How? They used a reversed base
• 6:23
  -64 encoding combined with an XOR cipher. Okay,
• 6:26
  for the non -engineers listening, what does an
• 6:28
  XOR cipher actually do in this context? Think
• 6:32
  of the XOR cipher as a very complex mathematical
• 6:35
  lock. They did every single character of their
• 6:38
  malicious code and applied a mathematical operation
• 6:41
  against it using a specific key, the word order
• 6:45
  7077 and a constant number. 333. So it's essentially
• 6:50
  like sending a locked briefcase through the mail,
• 6:53
  and the instructions on how to build the key
• 6:54
  are literally taped to the bottom of the briefcase.
• 6:57
  Exactly. And the automated post office scanners,
• 6:59
  they're only checking for explosives, right?
• 7:01
  They aren't looking for abstract math problems.
• 7:03
  Right. To the scanner, it doesn't look like an
• 7:05
  attack at all. It just looks like benign mathematics.
• 7:08
  It completely evades static analysis because
• 7:10
  the malicious commands literally do not exist
• 7:12
  in the file until the package is installed. Wow.
• 7:15
  Yeah, the post -install script just decodes itself
• 7:18
  directly in the machine's memory. And once it's
• 7:20
  decoded, it checks your operating system and
• 7:22
  delivers a highly tailored payload. And that's
• 7:25
  where the operational sophistication really gets
• 7:27
  scary. According to the Morningstar report, on
• 7:30
  a Mac... It uses an Apple script to disguise
• 7:33
  itself as a normal system daemon. Yeah, they
• 7:36
  deliberately named the process com .apple .act
• 7:39
  .mond. So if an engineer opens up their activity
• 7:42
  monitor because, say, their laptop fan is running
• 7:44
  loud, it just looks like another completely boring
• 7:47
  Apple background task. You never think to touch
• 7:49
  it. And on Windows, it's just as clever. It locates
• 7:53
  the built -in PowerShell tool and actually copies
• 7:55
  it to a totally new file named wt .exe. Which
• 7:59
  disguises it as the official Windows Terminal
• 8:02
  executable. Yes, which is a very well -known
• 8:04
  technique to evade endpoint detection software.
• 8:08
  And on Linux, it simply uses the curl command
• 8:10
  to drop a hidden Tython script. Okay, but here's
• 8:13
  the thing. Once this remote access Trojan is
• 8:15
  actually running on your machine, it has to phone
• 8:17
  home, right? It does. It has to connect to the
• 8:19
  attacker's server to receive commands. That outbound
• 8:22
  network traffic should be a massive red flag.
• 8:24
  You would think so. Our source has noted it connects
• 8:26
  to a command and control server at the domain
• 8:28
  sfrclack .com on port 8000. How on earth does
• 8:34
  a corporate network admin not see that weird
• 8:36
  domain and block it instantly? Because the malware
• 8:39
  heavily disguises its outgoing HTTP requests.
• 8:43
  It makes them look like routine developer traffic.
• 8:46
  When it sends data out, it formats the request
• 8:48
  to look like it is just pinging the normal software
• 8:51
  registry. So it's wearing a mask, basically.
• 8:53
  Exactly. On Mac, it labels the outgoing data
• 8:56
  as connecting to packages .npm .org. But wait,
• 8:59
  here is the brilliant, almost darkly funny part
• 9:01
  of this whole attack that I loved in the report.
• 9:03
  The domain npm .org does not actually belong
• 9:06
  to the node package manager. No, it doesn't.
• 9:08
  The real software registry is npmjs .org. The
• 9:12
  domain npm .org has actually belonged to the
• 9:14
  National Association of Pastoral Musicians since
• 9:16
  1997. Yeah, it's a psychological trick just as
• 9:19
  much as a technical one. Oh, totally. A security
• 9:21
  system or a human administrator sees a quick
• 9:23
  log entry with npm .org in the URL. They just
• 9:26
  assume it is a developer's machine fetching a
• 9:28
  package and they ignore it completely. They see
• 9:30
  the letters NPM and their brain just fills in
• 9:32
  the rest. Exactly. The malware even spoofs its
• 9:35
  own identity, pretending to be Internet Explorer
• 9:37
  8 running on Windows XP, just to blend into legacy
• 9:40
  traffic noise on corporate networks. It is just
• 9:43
  incredible. The attacker gets full arbitrary
• 9:45
  shell execution, system fingerprinting, the ability
• 9:48
  to inject more binaries, basically total control
• 9:51
  of the network, and then the cherry on top, it
• 9:54
  performs the ultimate anti -forensics. It cleans
• 9:57
  up after itself. It really does leave zero forensic
• 9:59
  trace. After executing, it deletes that setup
• 10:03
  .js script we talked about. It deletes the malicious
• 10:06
  package .json file that contained the hook. Just
• 10:09
  wiping the fingerprints. Exactly. And finally,
• 10:11
  it takes a clean, completely harmless file named
• 10:14
  package .md and renames it to package .json.
• 10:17
  So if a developer hears about this Axios vulnerability
• 10:20
  on the news and goes looking in their own system
• 10:22
  folders, everything looks perfectly normal. The
• 10:25
  evidence is just vaporized. It's gone. Security
• 10:28
  experts are saying that if you ran those affected
• 10:30
  versions, you have to assume total compromise.
• 10:33
  You're rotating every API key, changing every
• 10:36
  single password, and literally rebuilding your
• 10:39
  machines from scratch. Which is a nightmare.
• 10:42
  And because modern AI development tools rely
• 10:45
  so heavily on automated package fetching, the
• 10:47
  report noted that AI packages like at Shidenon
• 10:50
  OpenClaw caught the infection almost instantly.
• 10:54
  Yeah, it spreads like wildfire through automated
• 10:56
  systems. That cascading failure in the software
• 10:59
  pipeline is deeply unsettling. But you know,
• 11:01
  that same mathematical inevitability of a hidden
• 11:04
  defect, it applies to hardware too. It absolutely
• 11:07
  does. Especially when you scale it up to the
• 11:08
  size of SpaceX's current ambitions. So let's
• 11:12
  shift our focus from the invisible code on our
• 11:14
  servers to the massive physical ecosystem of
• 11:17
  low Earth orbit. We're moving from millions of
• 11:19
  lines of invisible code to thousands of physical
• 11:22
  objects flying at 17 ,000 miles per hour above
• 11:24
  our heads. Right. So according to the Morningstar
• 11:27
  report, SpaceX recently lost contact with Starlink
• 11:30
  satellite 34343. This happened on March 29th
• 11:34
  while the satellite was orbiting roughly 348
• 11:37
  miles or about 560 kilometers above Earth. Now,
• 11:42
  there are over 24 ,000 tracked objects in low
• 11:45
  Earth orbit right now. It's crowded up there.
• 11:47
  Very crowded. So my immediate assumption when
• 11:49
  a satellite goes offline is a collision. Did
• 11:52
  it hit a piece of space junk? Well, that would
• 11:55
  be the logical guess. But the data from Leolabs,
• 11:58
  they're an American space radar tracking company,
• 12:00
  tells a totally different story. OK, what did
• 12:02
  they see? They tracked tens of new fragments
• 12:04
  in orbit and explicitly attributed the event
• 12:07
  to an internal energetic source. Wow. So it didn't
• 12:10
  crash into anything. It wasn't hit by a solar
• 12:12
  flare or some geomagnetic storm. It literally
• 12:15
  blew up from the inside out. Yes, a spontaneous
• 12:18
  internal fragmentation event. And what makes
• 12:20
  this really significant is the pattern here.
• 12:22
  We actually saw a very similar internal anomaly
• 12:25
  cause an explosion just a few months ago, back
• 12:27
  on December 17, 2025. Right. I remember that.
• 12:30
  Now, SpaceX released a statement saying that
• 12:32
  because of the relatively low altitude, these
• 12:35
  fragments will eventually just burn up in the
• 12:36
  atmosphere. I think they will. And they pose
• 12:38
  no immediate risk to the International Space
• 12:40
  Station, the upcoming Artemis's second launch
• 12:42
  or their own transporter 16 mission. But frankly,
• 12:46
  the current debris isn't what worries me. It
• 12:49
  is the trajectory of their scale. Because scale
• 12:52
  fundamentally changes the physics of risk. Right
• 12:55
  now, SpaceX has roughly 10 ,000 starling satellites
• 12:58
  in orbit. but as they prepare for this massive
• 13:01
  record -shattering initial public offering, their
• 13:04
  expansion plans are honestly almost difficult
• 13:07
  to comprehend. Yeah, the Morningstar report noted
• 13:09
  that in January, they submitted a request to
• 13:12
  the FCC for approval to launch up to one million
• 13:15
  satellites. One million. One million. And these
• 13:17
  aren't just simple internet relays anymore. They
• 13:19
  are AI -driven satellites designed to act as
• 13:22
  orbital data centers. Which completely shifts
• 13:24
  the risk model of the entire aerospace industry.
• 13:27
  Historically, if you manufacture 10 bespoke satellites,
• 13:31
  you have Human engineers inspect every single
• 13:34
  weld, every single circuit board, and every battery
• 13:36
  cell. Right, it's artisanal almost. Exactly.
• 13:40
  But when you are mass manufacturing satellites
• 13:42
  on an assembly line to reach a constellation
• 13:44
  of one million, you are applying the software
• 13:48
  industry's rapid scaling mentality directly to
• 13:51
  aerospace hardware. And that brings us right
• 13:53
  back to the mathematics of failure we just talked
• 13:55
  about with the Axio supply chain attack. Yep.
• 13:58
  If you have a manufacturing defect rate of one
• 14:00
  in a million and you launch a few hundred satellites,
• 14:04
  you're statistically pretty safe. But if you
• 14:06
  launch one million satellites, that hidden internal
• 14:09
  vulnerability goes from a statistical improbability
• 14:12
  to a guaranteed physical event. Multiple guaranteed
• 14:15
  events, in fact. SpaceX president Gwynne Schmottwell
• 14:18
  actually noted recently that she was surprised
• 14:20
  the 1 million satellite request didn't make bigger
• 14:23
  waves in the press. Well, the aerospace community
• 14:25
  is certainly paying attention now. Oh, without
• 14:27
  a doubt. An internal energetic explosion inside
• 14:29
  a constellation of a million objects drastically
• 14:32
  increases the risk of a chain reaction. The Kessler
• 14:35
  syndrome. Yes, exactly. Where one explosion creates
• 14:38
  shrapnel that hits another satellite, creating
• 14:41
  more shrapnel, until low Earth orbit basically
• 14:44
  becomes an unusable blender. flying debris. Is
• 14:47
  the physical atmospheric equivalent of that malicious
• 14:50
  code cascading through an automated registry?
• 14:53
  It really is. The underlying theme here seems
• 14:55
  to be our immense reliance on automated quality
• 14:58
  assurance. SpaceX says they are investigating
• 15:01
  the root cause and will rapidly implement corrective
• 15:04
  actions. Right, but a hidden vulnerability sealed
• 15:06
  inside a metal chassis in the vacuum of space
• 15:09
  is infinitely harder to patch. than a line of
• 15:12
  code on a server. Oh, absolutely. It requires
• 15:14
  us to place absolute trust in the mass production
• 15:16
  systems of a single private company. OK, so we
• 15:19
  have radar systems searching for anomalies in
• 15:22
  orbit, and we have security scanners searching
• 15:24
  for anomalies in our code. But what about the
• 15:26
  technology we use to monitor the internal anomalies
• 15:29
  of our own bodies? Let's bring this down from
• 15:32
  the atmosphere right down to your wrist. Yeah,
• 15:34
  we are shrinking the scale down to the ultimate
• 15:36
  frontier here, human biology. We are looking
• 15:39
  at reporting on the wearable tech company Whoop.
• 15:42
  They just raised a staggering $575 million in
• 15:47
  a Series G funding round led by Collaborative
• 15:49
  Fund. That is massive. It rocketed their valuation
• 15:52
  to an incredible 10 .1. billion dollars. And
• 15:56
  much like SpaceX, Whoop is aggressively checking
• 15:58
  off financial milestones as they prepare for
• 16:01
  their own IPO. Right. Now, for you listening,
• 16:03
  if you aren't familiar, Whoop makes those really
• 16:05
  popular screenless fitness bands. They track
• 16:08
  your physiology 24 hours a day, your sleep stages,
• 16:12
  your cardiovascular strain and your heart rate
• 16:15
  variability. They're essentially acting as the
• 16:17
  Leo Labs for the human body. That's a great way
• 16:19
  to put it. They're trying to detect internal
• 16:22
  energetic anomalies. like an impending sickness
• 16:25
  or severe overtraining before you physically
• 16:27
  crash. It's a perfect analogy and the financial
• 16:29
  metrics validating this model are pretty remarkable.
• 16:32
  They are projected to hit one billion dollars
• 16:35
  in annual recurring revenue by the end of 2025.
• 16:37
  A billion dollars purely from health subscription
• 16:40
  software. Yeah. But what I find most revealing
• 16:43
  in the Morningstar report is where this exponential
• 16:46
  growth is actually coming from. Right. The geography
• 16:48
  of it. Exactly. The U .S. market usually dominates
• 16:51
  early adoption for continuous biometric tracking.
• 16:55
  But 60 percent of Whoop's 2025 sales were international.
• 17:00
  Specifically, there is a massive push into the
• 17:03
  Middle East. If you look at this latest funding
• 17:06
  round, there is heavy participation from Middle
• 17:09
  Eastern sovereign wealth funds. Exactly. Specifically,
• 17:12
  the Mubidala Investment Company and the Qatar
• 17:14
  Investment Authority. Wow. That is serious institutional
• 17:18
  capital. You don't get sovereign wealth funds
• 17:19
  involved unless you are building foundational
• 17:22
  societal infrastructure. Exactly. Whoop's CEO,
• 17:26
  Will Ahmed, has positioned the Gulf Cooperation
• 17:28
  Council, or GCC region, as central to their long
• 17:31
  -term strategy. The region is investing heavily
• 17:34
  in forward -looking health, human performance,
• 17:36
  and longevity initiatives. Whoop is rapidly expanding
• 17:39
  their local teams and their retail footprint
• 17:41
  to capture that market. They have even brought
• 17:44
  Christiano Ronaldo on board, right? Yeah, he's
• 17:46
  an investor in this round and a global ambassador
• 17:49
  now. He publicly stated that the platform is
• 17:52
  one of his most important tools for long -term
• 17:54
  health and performance. Which is a huge endorsement.
• 17:57
  We are seeing a massive global cultural shift
• 18:01
  toward preventative medicine driven by continuous
• 18:04
  data monitoring. But, you know, this brings us
• 18:06
  right back to our core theme today of hidden
• 18:08
  vulnerabilities. We are treating the human body
• 18:11
  like a complex software system that can be debugged,
• 18:15
  optimized, and upgraded provided we have enough
• 18:17
  data. Right. millions of people are trusting
• 18:19
  these algorithms to tell them when they're stressed,
• 18:22
  when their immune system is compromised, and
• 18:24
  how to govern their daily energy. Which requires
• 18:27
  absolute unwavering trust in the integrity of
• 18:30
  the data pipeline. Exactly. If the software reading
• 18:33
  our biometrics has a hidden vulnerability like,
• 18:35
  say, a compromised dependency from an open source
• 18:38
  library, like we saw with Axios, the consequences
• 18:40
  aren't just a hacked database. No, it is corrupted
• 18:43
  medical and physiological guidance. Exactly.
• 18:46
  If I am generating a billion data points about
• 18:49
  my cardiovascular health, and that is being processed
• 18:52
  by a global entity, the security of that pipeline
• 18:55
  is just as critical as any enterprise software
• 18:58
  pipeline. When you scale biological tracking
• 19:00
  to a global population, a corrupted algorithm
• 19:03
  could lead to millions of people misinterpreting
• 19:06
  their own health data simultaneously. The hardware
• 19:09
  on your wrist is really only as reliable as the
• 19:12
  massive invisible cloud infrastructure processing
• 19:15
  its signals. It's wild to think about. It really
• 19:18
  is. Well, before we wrap up today's show, we
• 19:20
  want to mention our sponsor, Old Glory. They
• 19:23
  are an iconic music and sports fan merch store,
• 19:27
  which is Honestly great stuff for your techie
• 19:29
  friend or family member. They really do have
• 19:31
  an amazing selection They do and if listeners
• 19:33
  use promo code snuffle at checkout You can get
• 19:36
  15 % off your entire purchase the website is
• 19:39
  oldglory .com Definitely check them out So just
• 19:41
  to recap what we've covered today when you look
• 19:43
  at our morning star sources You can see a pretty
• 19:47
  terrifying pattern emerging across all these
• 19:48
  domains. Yeah, the through line is scale Whether
• 19:51
  you are a developer pushing code to the cloud,
• 19:53
  a citizen living under a sky rapidly filling
• 19:56
  with thousands of satellites, or an athlete trusting
• 19:59
  a wearable to tell you how your heart is doing,
• 20:01
  you are deeply embedded in these massively scaled
• 20:04
  networks. And the health of every single one
• 20:06
  of these networks relies entirely on the assumption
• 20:09
  that we can detect and neutralize anomalies before
• 20:12
  they detonate. Which, as we've seen, is getting
• 20:15
  harder and harder to guarantee. But here is a
• 20:18
  thought to leave you with today. We are rapidly
• 20:21
  moving toward a future where our most intimate
• 20:23
  biological data might be beamed up to one of
• 20:26
  SpaceX's million AI -driven orbital data centers,
• 20:30
  processed by software built on millions of lines
• 20:32
  of open source norm that no single human has
• 20:35
  entirely verified. It's a house of cards built
• 20:37
  on algorithms. It really is. And if we are completely
• 20:39
  outsourcing our intuition and our biological
• 20:41
  monitoring to AI, maybe the ultimate vulnerability
• 20:44
  isn't in the code or the hardware at all. Right,
• 20:46
  it might be our total willingness to surrender
• 20:48
  our own physical intuition to the dashboard.
• 20:52
  As these systems become so incredibly complex
• 20:55
  and intertwined that human oversight is physically
• 20:57
  impossible, well, the greatest risk might simply
• 21:01
  be our blind trust in the machine to tell us
• 21:03
  who we are and how we feel. Something to think
• 21:06
  about the next time you check your wrist to see
• 21:07
  how you slept last night. Definitely. We're here
• 21:10
  to help at Snarfful, so reach out on our website
• 21:12
  if you have any more questions. That's a wrap
• 21:18
  on today's top stories. Remember, tech doesn't
• 21:21
  have to be complicated, and we're here to help
• 21:23
  connect the dots. If you've got questions or
• 21:26
  want to learn more, visit us at snarful .com.
• 21:29
  Thanks for listening, and we'll catch you next
• 21:31
  time.