Latest / Murder Files Unsealed / Good Cop/Bad Cop: Solving the Murder of Heather Bogle

Transcript

Good Cop/Bad Cop: Solving the Murder of Heather Bogle · Murder Files Unsealed

• 0:00
  Speaker 1: So every time you copy and paste a simple video link,
• 0:04
  Speaker 1: you're unwittingly deploying military grade encryption and you know, exploiting
• 0:08
  Speaker 1: Belgian geopolitical real estate and launching a microscopic surveillance drone.
• 0:13
  Speaker 1: Today we're dissecting exactly one line of text. It's just
• 0:16
  Speaker 1: forty three characters, and that is actually our entire source
• 0:19
  Speaker 1: material for this deep dive. Our listener handed us a
• 0:23
  Speaker 1: single YouTube url, and we are going to read it
• 0:26
  Speaker 1: not as a path to a video, but as a
• 0:28
  Speaker 1: digital fossil.
• 0:29
  Speaker 2: Yeah, it really is the ultimate minimalist source. We aren't
• 0:32
  Speaker 2: going to watch the video at points too, We're completely
• 0:34
  Speaker 2: ignoring the destination to focus entirely on the vehicle, because
• 0:37
  Speaker 2: hidden inside this seemingly mundane string of text is like
• 0:40
  Speaker 2: the entire architectural history of the modern Internet.
• 0:43
  Speaker 1: Okay, let's unpack this. I'm gonna read the exact string
• 0:45
  Speaker 1: just so we're looking at the same artifact.
• 0:46
  Speaker 2: Here.
• 0:47
  Speaker 1: It is https colon slash slash, YouTube, dot, b slash
• 0:52
  Speaker 1: b zero h seven t g dw dcqy question mark
• 0:56
  Speaker 1: c rg joe six q five ou dash six p.
• 0:59
  Speaker 2: Right forty three keystrokes. But if you know how to
• 1:01
  Speaker 2: read them. They break them into these very distinct geological
• 1:04
  Speaker 2: layers of technology.
• 1:05
  Speaker 1: Yeah, and let's start with the outermost layer of defense. Yeah,
• 1:08
  Speaker 1: the very first five letters HTTPS. I think most people
• 1:12
  Speaker 1: know the S stands for secure, and they know the
• 1:15
  Speaker 1: old Internet use regular HTTP because we used to just
• 1:19
  Speaker 1: send our web request in plain text, which was essentially
• 1:23
  Speaker 1: mailing postcards that anyone on the same coffee shop WiFi
• 1:26
  Speaker 1: could read.
• 1:27
  Speaker 2: Right, exactly. Yeah, in the early days, a simple packet
• 1:30
  Speaker 2: sniffing program could just watch your passwords and your emails
• 1:32
  Speaker 2: just floating through the airwaves. The transition to HTTPS really
• 1:37
  Speaker 2: changed the fundamental physics of how we communicate online.
• 1:40
  Speaker 1: But I want to talk about the mechanism here because
• 1:42
  Speaker 1: people always use the analogy of an armored car for HTTPS.
• 1:45
  Speaker 1: You know, like your data goes into a safe, the
• 1:48
  Speaker 1: safe is into the car and it drives down the highway, right,
• 1:51
  Speaker 1: But that doesn't actually explain the cryptography. How does my
• 1:54
  Speaker 1: web browser lock the safe if the server on the
• 1:56
  Speaker 1: other side of the country is the one holding the key.
• 1:59
  Speaker 2: Well, what's fascinating here is that that is the genius
• 2:01
  Speaker 2: of asymmetric encryption. It relies on a mathematical concept called
• 2:06
  Speaker 2: a public private key pair. So before you ever send
• 2:09
  Speaker 2: your web request, the server you're connecting to sends your
• 2:12
  Speaker 2: browser a public.
• 2:13
  Speaker 1: Key, right, so teuse a mechanical analogy. The server mails
• 2:17
  Speaker 1: an open padlock to my browser. Yeah, and millions of
• 2:20
  Speaker 1: these open padlocks are distributed to anyone who asks for one.
• 2:23
  Speaker 2: Precisely. Yeah, your browser puts your data. So your request
• 2:27
  Speaker 2: for this specific video inside a box, and it snaps
• 2:31
  Speaker 2: that public padlock shut. Oh okay, but here is the
• 2:33
  Speaker 2: critical part. The padlock is designed so that once it
• 2:37
  Speaker 2: clicks shut, it cannot be reopened with the same key
• 2:39
  Speaker 2: used to close it. The only thing in the universe
• 2:42
  Speaker 2: that can unlock it is the private key, which the
• 2:44
  Speaker 2: server never ever shares. It just stays locked inside their
• 2:48
  Speaker 2: data center.
• 2:48
  Speaker 1: Wow. Okay, So even if an attacker intercepts the box
• 2:52
  Speaker 1: in transit, and even if they have a copy of
• 2:54
  Speaker 1: the same open padlock from the server, it doesn't help
• 2:57
  Speaker 1: them at all because a math only works in one.
• 2:59
  Speaker 2: Direction, right exactly. The math is actually based on prime factorization.
• 3:04
  Speaker 2: It is computationally trivial for a machine to multiply two
• 3:07
  Speaker 2: massive prime numbers together to create a padlock, but it
• 3:11
  Speaker 2: is practically impossible I mean even for supercomputers to look
• 3:14
  Speaker 2: at that massive resulting number and reverse engineer which two
• 3:17
  Speaker 2: primes created it? Right, right, that one way mathematical function
• 3:21
  Speaker 2: is what allows you to securely ask for a video
• 3:24
  Speaker 2: over a compromised public Wi Fi.
• 3:27
  Speaker 1: Network, And all of that invisible, incredibly complex prime number
• 3:32
  Speaker 1: math is represented by just one letter, just the SS
• 3:36
  Speaker 1: at the front of our forty three character strength.
• 3:38
  Speaker 2: Yeah.
• 3:39
  Speaker 1: But right after that protocol we hit the domain, which
• 3:41
  Speaker 1: is Utwo dot b, which is this deliberate, highly engineered
• 3:45
  Speaker 1: deviation from the standard dot com format.
• 3:48
  Speaker 2: It's such a fascinating artifact of how human behavior forces
• 3:52
  Speaker 2: infrastructure to adapt. If you go back to the origins
• 3:54
  Speaker 2: of the Domain name system DNS, the whole point was
• 3:57
  Speaker 2: to create human readable descriptive addresses, right because the Internet
• 4:00
  Speaker 2: routes traffic using IP addresses, which are just strings of numbers.
• 4:04
  Speaker 2: So DNS was invented so you could type a logical
• 4:06
  Speaker 2: descriptive name instead of remembering a server's exact numerical coordinate.
• 4:11
  Speaker 1: But then the Internet went mobile and we entered this
• 4:13
  Speaker 1: era of strict character limits, like when Twitter kept you
• 4:18
  Speaker 1: at one hundred and forty characters. A standard descriptive URL
• 4:21
  Speaker 1: became a massive liability. If your link is seventy five
• 4:25
  Speaker 1: characters long, you have barely any room left to type
• 4:28
  Speaker 1: of thought.
• 4:29
  Speaker 2: Right, that constraint birthed the entire URL shortener industry. You
• 4:33
  Speaker 2: had services like bitly emerge acting as middlemen. You gave
• 4:37
  Speaker 2: them your really long URL and they gave you a
• 4:39
  Speaker 2: highly compressed, randomized short link that just redirected to your
• 4:43
  Speaker 2: original destination.
• 4:45
  Speaker 1: But the major platforms quickly recognized a dual threat in
• 4:48
  Speaker 1: relying on those third party shorteners. Right first, if the
• 4:51
  Speaker 1: shortener company goes bankrupt or their servers crash, millions of
• 4:54
  Speaker 1: links across the Internet suddenly just lead nowhere. Yeah. Total,
• 4:57
  Speaker 1: it's an unacceptable single point of failure.
• 4:59
  Speaker 2: It is. And the second threat was brand dilution. A
• 5:02
  Speaker 2: generic bitly link gives the user no context about where
• 5:05
  Speaker 2: they are about to go. So these massive platforms decided
• 5:09
  Speaker 2: to build their own in house URL shorteners. They wanted
• 5:12
  Speaker 2: the brevity, but they also really wanted to maintain their
• 5:15
  Speaker 2: brand identity.
• 5:16
  Speaker 1: Which brings us to the bizarre geopolitical reality of our
• 5:19
  Speaker 1: source texts today. Because to compress their brand name, this
• 5:23
  Speaker 1: specific platform sliced it in half and used dot B
• 5:27
  Speaker 1: as their top level domain. But dot B isn't a
• 5:29
  Speaker 1: generic tech domain, it's literally the sovereign country code for Belgium.
• 5:34
  Speaker 2: Yes, it's a practice known as domain hacking. When the
• 5:37
  Speaker 2: Internet Corporation for Assigned Names and Numbers i CAN was
• 5:41
  Speaker 2: standardizing the Web, they assigned two letter top level domains
• 5:45
  Speaker 2: to every recognized country based on their geography, so dot
• 5:49
  Speaker 2: US for the United States, dot UK for the United Kingdom.
• 5:52
  Speaker 1: But they inadvertently created this massive global digital real estate
• 5:56
  Speaker 1: market because tech companies realized they could string together their
• 5:59
  Speaker 1: brand name using the digital infrastructure of foreign nations exactly.
• 6:03
  Speaker 2: So Belgium happens to be dot D. Libya was assigned
• 6:06
  Speaker 2: dots LEE, which give us you know, hundreds of startups
• 6:08
  Speaker 2: ending in.
• 6:08
  Speaker 1: L y oh, right, like Bentley.
• 6:10
  Speaker 2: Yeah. And Tuvalu, which is this tiny island nation in
• 6:13
  Speaker 2: the Pacific, was assigned dot TV and they now generate
• 6:16
  Speaker 2: a significant portion of their national GDP just by leasing
• 6:19
  Speaker 2: that domain to media companies.
• 6:21
  Speaker 1: It's incredible. I mean, one of the wealthiest technology companies
• 6:24
  Speaker 1: in California is actively routing global web traffic through the
• 6:29
  Speaker 1: digital jurisdiction of the Kingdom of Belgium purely to save
• 6:33
  Speaker 1: four keystrokes in a textbox. Yeah, it really highlights how
• 6:37
  Speaker 1: the physical political map of the world has been contorted
• 6:41
  Speaker 1: to basically serve the linguistic conveniences of software engineering.
• 6:44
  Speaker 2: It really does, and that Belgian shortcut operates solely to
• 6:47
  Speaker 2: deliver us to the payload. You know, the real core
• 6:50
  Speaker 2: of our forty three character string. After the slash, we
• 6:53
  Speaker 2: have the unique video identifier which is B zero h
• 6:56
  Speaker 2: seven t g wzqduy a.
• 6:58
  Speaker 1: Live in characters. And so what does this all mean.
• 7:02
  Speaker 1: This is the exact coordinate of the file in the database. Now,
• 7:05
  Speaker 1: anyone looking at this knows it's a Base sixty four
• 7:07
  Speaker 1: encoding system because it's using uppercase letters, lower case letters,
• 7:11
  Speaker 1: numbers zero through nine, and a couple of symbols like
• 7:14
  Speaker 1: dashes or underscores to maximize data density.
• 7:17
  Speaker 2: Right, because if you only use numbers like Base ten,
• 7:19
  Speaker 2: eleven characters only gives you one hundred billion combinations for
• 7:22
  Speaker 2: a platform ingesting thousands of hours of video every single
• 7:26
  Speaker 2: minute globally, one hundred billion addresses would have been exhausted
• 7:30
  Speaker 2: years ago. So Base sixty four bumps the math to
• 7:33
  Speaker 2: sixty four to the eleventh power.
• 7:35
  Speaker 1: Which comes out to roughly seventy three quintillion possible unique identifiers.
• 7:41
  Speaker 1: But this is where I want to push back on
• 7:42
  Speaker 1: the architecture a little bit. Okay, I understand that seventy
• 7:45
  Speaker 1: three quintillion is a mathematically permanent ceiling. We could upload
• 7:49
  Speaker 1: a video every second for centuries and never hit the hardcap.
• 7:53
  Speaker 1: But they aren't assigning these ideas in a sequential order, right,
• 7:57
  Speaker 1: They are generating them randomly? Yeah? Is this string just
• 8:01
  Speaker 1: random alphabet soup? Aren't they eventually going to run out
• 8:03
  Speaker 1: of eleven character combinations if they just guess randomly.
• 8:06
  Speaker 2: Well correct Sequential ideas are a massive security risk and
• 8:10
  Speaker 2: they make database partitioning a total nightmare. So they are
• 8:13
  Speaker 2: generated pseudorandomly.
• 8:15
  Speaker 1: So if they are random, doesn't the platform run into
• 8:17
  Speaker 1: the birthday paradox long before they reach seventy three quintillion?
• 8:20
  Speaker 2: Oh?
• 8:20
  Speaker 1: Interesting? Like, if I have a room of just twenty
• 8:24
  Speaker 1: three people, there's a fifty percent chance two of them
• 8:27
  Speaker 1: share a birthday, even though there are three hundred and
• 8:30
  Speaker 1: sixty five days in the year. In database architecture, if
• 8:33
  Speaker 1: you are randomly generating strings, the chance of a collision,
• 8:37
  Speaker 1: like accidentally generating the exact same eleven character ID for
• 8:41
  Speaker 1: two completely different videos skyrockets way before the name space
• 8:44
  Speaker 1: is actually full.
• 8:45
  Speaker 2: That is an excellent deduction, and it is exactly why
• 8:48
  Speaker 2: the name space has to be so astronomically huge. If
• 8:52
  Speaker 2: we connect this to the bigger picture, you are entirely
• 8:54
  Speaker 2: correct about the mathematics of collisions. If their ceiling was
• 8:58
  Speaker 2: a trillion and they were generating random to my ds,
• 9:01
  Speaker 2: they would start seeing collisions where the system tries to
• 9:03
  Speaker 2: assign an existing ID to a new upload far too often.
• 9:07
  Speaker 2: The database would constantly have to check, reject and regenerate keys,
• 9:11
  Speaker 2: which creates massive latency, and at.
• 9:13
  Speaker 1: Their scale, a few milliseconds of latency on every single
• 9:17
  Speaker 1: upload destroys the serverload.
• 9:19
  Speaker 2: Exactly by expanding the available name space to seventy three quintillion,
• 9:23
  Speaker 2: they pushed the probability of a birthday paradox collision so
• 9:26
  Speaker 2: far down into the decimal points that it effectively becomes zero.
• 9:30
  Speaker 2: The sheer size of the eleven character based sixty four
• 9:33
  Speaker 2: string isn't just about having enough room to store videos.
• 9:37
  Speaker 2: It is an engineering buffer. It ensures that the random
• 9:40
  Speaker 2: number generators can fire blindly at incredible speeds without ever
• 9:44
  Speaker 2: having to worry about hitting the same coordinate twice. If
• 9:47
  Speaker 2: every human on Earth uploaded a video every second for
• 9:50
  Speaker 2: their entire lives, we still wouldn't put a dent in
• 9:53
  Speaker 2: the available combinations. This eleven character string is mathematically permanent.
• 9:57
  Speaker 1: It's a brilliant solution. You saw the computational processing problem
• 10:01
  Speaker 1: simply by making the digital universe too vast for a
• 10:04
  Speaker 1: coincidence to ever occur. So the browser has securely connected VIAHDTPS.
• 10:08
  Speaker 1: It used the Belgian DNAs shortcut to find the server,
• 10:11
  Speaker 1: and it queried the database for this specific collision proof
• 10:14
  Speaker 1: eleven character id. The video is found, the core job
• 10:18
  Speaker 1: is done right well.
• 10:19
  Speaker 2: And yet our tech string isn't finished because the destination
• 10:22
  Speaker 2: isn't the end of the story. This specific URL provided
• 10:25
  Speaker 2: by the listener includes a trailing piece of information that
• 10:28
  Speaker 2: reveals exactly how it.
• 10:30
  Speaker 1: Travels right, because there is a question mark followed by
• 10:33
  Speaker 1: a final sixteen character string, which is c equals eqnsgr
• 10:39
  Speaker 1: six q five nou dashed six P. Here's where it
• 10:43
  Speaker 1: gets really interesting, because we're moving past just finding a video.
• 10:46
  Speaker 2: Yeah, this is where we pivot from database architecture into
• 10:49
  Speaker 2: behavioral telemetry. That question mark signifies the beginning of a
• 10:53
  Speaker 2: query string, and to understand why it's there, we have
• 10:56
  Speaker 2: to look at how the web fundamentally operates. HDTP is
• 11:00
  Speaker 2: a stateless protocol, meaning.
• 11:01
  Speaker 1: The server has no memory. It has the object permanence
• 11:04
  Speaker 1: of a goldfi.
• 11:04
  Speaker 2: Exactly, you ask the server for video, it hands you
• 11:07
  Speaker 2: the video, and the transaction is completely closed. The server
• 11:09
  Speaker 2: immediately forgets you exist. The next time you click something,
• 11:12
  Speaker 2: you are a brand new stranger all over again. But
• 11:15
  Speaker 2: modern platforms can't survive on stateless interactions. They need to
• 11:19
  Speaker 2: maintain a continuous narrative of your behavior to serve you, ads,
• 11:23
  Speaker 2: recommend content, and map user interactions.
• 11:26
  Speaker 1: So how do you force a goldfish to remember something?
• 11:30
  Speaker 1: You write it down and staple it to the request.
• 11:33
  Speaker 1: That's what the query string is. It's kind of like
• 11:35
  Speaker 1: buying a book, but the bookstore glues a highly detailed
• 11:38
  Speaker 1: receipt to the back cover, so the publisher always knows
• 11:41
  Speaker 1: exactly who handed the book to you.
• 11:43
  Speaker 2: And when this raises an important question, right, if you
• 11:46
  Speaker 2: strip everything after the question mark off this URL, the
• 11:50
  Speaker 2: video still loads perfectly. The video ID did its.
• 11:53
  Speaker 1: Job, right. I do that all the time. Actually, I
• 11:55
  Speaker 1: just delete the end of the link before I send it, right, But.
• 11:57
  Speaker 2: If you don't, the platform gains a crew piece of intelligence.
• 12:01
  Speaker 2: In our string, we see SI equals that stands for
• 12:04
  Speaker 2: source identifier, and this is.
• 12:06
  Speaker 1: Essentially a tracking drone glued to the URL. When a
• 12:10
  Speaker 1: user clicks the share button on this platform, it dynamically
• 12:13
  Speaker 1: generates this unique side code and tacks it onto the end.
• 12:17
  Speaker 2: Let's break down how a graph database actually ingests that parameter.
• 12:21
  Speaker 2: In back end data structures, platforms map behavior using nodes
• 12:24
  Speaker 2: and edges. A node is a data point like user,
• 12:27
  Speaker 2: a user b or a specific video, and edge is
• 12:30
  Speaker 2: the relationship connecting them.
• 12:32
  Speaker 1: Okay, so I am blogged into my account on my phone.
• 12:35
  Speaker 1: I watched this video, I hit share, and I text
• 12:37
  Speaker 1: the link to you.
• 12:38
  Speaker 2: At the exact moment you hit share, the platform's database
• 12:41
  Speaker 2: creates an edge. It records that user a generated link
• 12:45
  Speaker 2: parameter eqs R six Q five nouh six P four
• 12:50
  Speaker 2: video X.
• 12:51
  Speaker 1: Okay.
• 12:52
  Speaker 2: Later I receive your text message, I click the link.
• 12:54
  Speaker 2: My browser sends that entire forty three character string to
• 12:58
  Speaker 2: the server. The server looks at the eleven character ID
• 13:00
  Speaker 2: and serves me the video, but the telemetry engine strips
• 13:03
  Speaker 2: off the query parameter and logs it.
• 13:05
  Speaker 1: So the database is the exact code I generated earlier exactly.
• 13:09
  Speaker 2: The system now creates a new edge, connecting me user
• 13:12
  Speaker 2: B to that same code. The graph database instantly resolves
• 13:15
  Speaker 2: the triangle. It now has mathematical proof that user A
• 13:18
  Speaker 2: and user B share a direct off platform communication channel.
• 13:22
  Speaker 1: Wow, it's mapping a massive, invisible social network without anyone
• 13:26
  Speaker 1: ever sending a friend request. As a user, I just
• 13:29
  Speaker 1: think I'm sending a funny video. The design of the
• 13:31
  Speaker 1: interface implies that the link is just a neutral pointer,
• 13:33
  Speaker 1: but by embedding that stateful parameter, the platform has turned
• 13:38
  Speaker 1: a simple recommendation into a trackable, quantifiable data point. There
• 13:43
  Speaker 1: is this inherent tension here right between my desire for
• 13:47
  Speaker 1: frictionless sharing and the platform's desire for systemic surveillance, and.
• 13:52
  Speaker 2: They feed that edge directly into the recommendation engine. If
• 13:55
  Speaker 2: you share a video with me and I watch the
• 13:57
  Speaker 2: entire thing, the algorithm deduces that our behavioral clusters overlap.
• 14:01
  Speaker 2: Right the next time I log in, it might subtly
• 14:04
  Speaker 2: start prioritizing content that you have previously engaged with because
• 14:08
  Speaker 2: our newly established digital edge suggests we have aligned preferences.
• 14:12
  Speaker 1: It takes the concept of word of mouth and digitizes
• 14:15
  Speaker 1: it into a highly structured, queriable matrix. It really is
• 14:18
  Speaker 1: a marvel.
• 14:18
  Speaker 2: It is. When you look at the entirety of what
• 14:20
  Speaker 2: we've discussed today, it's astonishing how much functional density is
• 14:24
  Speaker 2: packed into forty three characters.
• 14:26
  Speaker 1: Yeah, just summarizing our journey here, the listener provided exactly
• 14:29
  Speaker 1: forty three characters of text. From that, we unpack the
• 14:32
  Speaker 1: cryptographic public key exchange of HTTPS, protecting the data in transit.
• 14:37
  Speaker 2: To the DNS domain hacking that leverages Belgian national sovereignty
• 14:41
  Speaker 2: for platform branding.
• 14:43
  Speaker 1: Right, and the combinatorial mathematics of base fifty four encoding
• 14:46
  Speaker 1: that builds a seventy three quintillion slot database to avoid
• 14:50
  Speaker 1: those birthday paradox.
• 14:51
  Speaker 2: Collisions, right down to the stateless protocol workarounds that allow
• 14:55
  Speaker 2: query parameters to map human relationships across a graph database.
• 15:00
  Speaker 2: All of this executing perfectly millions of times a second,
• 15:03
  Speaker 2: every single day.
• 15:05
  Speaker 1: It really makes you look at a simple blue hyperlink
• 15:07
  Speaker 1: in a completely different way. Every single link you click, copy,
• 15:11
  Speaker 1: or share is a complex machine operating perfectly in milliseconds.
• 15:15
  Speaker 1: But you mentioned something at the very beginning of this
• 15:17
  Speaker 1: deep dive. You called this URL a digital fossil. Yeah,
• 15:21
  Speaker 1: and I keep thinking about the physical infrastructure required to
• 15:23
  Speaker 1: keep this forty three character string functional.
• 15:25
  Speaker 2: Right, It is entirely dependent on a living, breathing Internet.
• 15:29
  Speaker 2: The Belgian DNS registry has to remain active and governed.
• 15:33
  Speaker 2: The certificate authorities must continuously validate the cryptographic keys. The
• 15:37
  Speaker 2: database servers must remain powered and maintained to resolve that
• 15:40
  Speaker 2: a loving character ID.
• 15:42
  Speaker 1: So what happens when the machine stops? Centuries from now?
• 15:45
  Speaker 1: Platforms will inevitably collapse, you know, servers will be decommissioned,
• 15:48
  Speaker 1: and domain redistries will change hands or just expire.
• 15:51
  Speaker 2: The machine breaks. And when it does, this highly engineered,
• 15:54
  Speaker 2: incredibly dense string of texts, this vehicle we've spent the
• 15:58
  Speaker 2: last half hour dissecting, becomes completely inert.
• 16:01
  Speaker 1: It truly becomes a fossil without the massive hidden infrastructure
• 16:05
  Speaker 1: to interpret the cryptography, route the domain and resolve the
• 16:08
  Speaker 1: database query that exact text hgtps coal and slash slash
• 16:13
  Speaker 1: utuo dot b slash b zero a seven t g
• 16:16
  Speaker 1: w zqdy question mark side equals eq as g R
• 16:19
  Speaker 1: six two five nou dash six p reverse to just
• 16:22
  Speaker 1: being a meaningless arrangement of letters and numbers, but for.
• 16:25
  Speaker 2: A brief moment in time, it was a precise mathematical
• 16:28
  Speaker 2: key that summoned a very specific, fleeting piece of human
• 16:31
  Speaker 2: expression out of the ether.
• 16:33
  Speaker 1: It makes you wonder consider the ultimate life span of
• 16:36
  Speaker 1: this exact string. Every message we send, every link we
• 16:39
  Speaker 1: in bed, every digital fingerprint we leave behind. They are
• 16:42
  Speaker 1: all entirely dependent on a machine that requires constant electricity.
• 16:45
  Speaker 1: We are generating billions of these complex little artifacts every
• 16:49
  Speaker 1: single day, just waiting for the servers to power down,
• 16:52
  Speaker 1: leaving behind a digital archaeology that future generations may never
• 16:55
  Speaker 1: be able to unlock. What other digital fossils are we
• 16:58
  Speaker 1: unknowingly archiving today