Latest / faafo with courtney / The Architecture of Liberation: AI, Identity, and Digital Ancestry

Transcript

The Architecture of Liberation: AI, Identity, and Digital Ancestry · faafo with courtney

• 0:00
  You know that feeling when you're like just moving
• 0:01
  through a workflow totally in the zone and your
• 0:04
  software suddenly throws up this pop -up window
• 0:06
  that just completely stops you in your tracks.
• 0:09
  Oh, yeah. The absolute worst. Right. It sort
• 0:10
  of refuses to let you click publish until you
• 0:13
  answer some mandatory question. Well, for Courtney
• 0:16
  Crossland, she's an American expat living down
• 0:19
  in Salvador Bahia, Brazil, and he goes by Cece
• 0:22
  in Brazil online. That sudden roadblock came
• 0:25
  from her podcast hosting platform. the service
• 0:28
  pushing out her RSS feed to Apple and Spotify.
• 0:32
  Exactly. And the question it asked her point
• 0:35
  blank was just, is this AI? Wow. I mean, it's
• 0:38
  the ultimate digital litmus test of our current
• 0:40
  moment. Just three little words. But they act
• 0:43
  as this massive lightning rod for just an immense
• 0:45
  amount of social, economic, and, well, cultural
• 0:49
  anxiety right now. And the context here is everything
• 0:52
  because Cece was getting asked this while working
• 0:54
  on an incredibly personal project. She was creating
• 0:57
  this audio documentary and a work of art that
• 1:00
  mapped out her own family tree spanning over
• 1:03
  500 years. That's a massive undertaking. It really
• 1:05
  is. And she's recording, producing and researching
• 1:08
  all of this from Bahia, Brazil. So when the platform
• 1:11
  demanded to know if she used AI -like, specifically
• 1:15
  large language models to help organize all that
• 1:17
  historical data, or synthetic voices for narration,
• 1:21
  she just confidently checked yes. No hesitation.
• 1:25
  None. And here is what makes this so fascinating,
• 1:27
  I think. She didn't say yes. because she was
• 1:30
  backed into a corner or forced to use these tools,
• 1:33
  she checked the box because she intentionally
• 1:35
  wanted to use them. Which is bold, because checking
• 1:38
  yes to utilizing AI in any creative space right
• 1:41
  now is basically like... kicking a hornet's nest.
• 1:44
  I mean, it just invites this level of intense
• 1:46
  scrutiny and honestly, an absolute flood of internet
• 1:49
  backlash. OK, let's unpack this because today's
• 1:52
  deep dive into the source material is going to
• 1:54
  explore exactly why she made that choice and
• 1:57
  why she remains completely unapologetic about
• 1:59
  it. We're looking at a stack of sources that
• 2:01
  pull this topic in entirely different directions
• 2:03
  today. We've got a highly compelling blog post
• 2:05
  written by Cece herself addressing the backlash.
• 2:08
  We have a very heated Reddit debate dissecting
• 2:11
  the environmental footprint of AI. And we have
• 2:14
  a passionate YouTube panel discussion arguing
• 2:16
  over whether this tech is a genuine tool for
• 2:20
  black liberation or if it's just a brand new
• 2:24
  engine for a very old type of extraction. It's
• 2:26
  a lot to cover. It is. The mission today is to
• 2:29
  explore the messy, sometimes hypocritical, and
• 2:31
  entirely fascinating intersection of AI, race,
• 2:35
  historical dead, and digital freedom. Yeah, and
• 2:38
  to really grasp the gravity of Cece's stance
• 2:40
  here, we have to look at the receipts she brings
• 2:42
  to the table. In response to the people who suddenly
• 2:45
  had negative things to say about her use of large
• 2:47
  language models, she published a blog post, and
• 2:49
  it has this beautifully defiant title. I said
• 2:51
  yes. and I would do it again. I love that. And
• 2:54
  she points out in there that she's been creating
• 2:56
  in digital spaces since 2006. Right, which is
• 2:59
  huge. Yeah. Yeah. We're talking about the era
• 3:01
  when podcasting meant sitting at some bulky desktop
• 3:04
  computer, wrangling a weird external recorder,
• 3:07
  and manually coding your own XML files just to
• 3:10
  get an episode online. Oh, man, the Wild West
• 3:13
  days. Exactly. She was building in the digital
• 3:15
  space long before the tools were frictionless.
• 3:18
  And certainly long before the crowds who joined
• 3:20
  Twitter in like 2017 started walking around calling
• 3:24
  themselves Internet pioneers. And that timeline
• 3:27
  really matters because it gives her the perspective
• 3:29
  to see these cycles of tech outrage. Her core
• 3:32
  argument in the post is that the current fury
• 3:34
  over AI is completely saturated in hypocrisy.
• 3:37
  Yeah, it really is. She highlights this massive
• 3:40
  contradiction, right? People are furious right
• 3:42
  now about AI companies scraping their data without
• 3:44
  permission to train neural networks. Yet those
• 3:46
  exact same people will happily take a vacation,
• 3:49
  stroll through a massive institution like the
• 3:51
  British Museum, look at thousands of pillaged,
• 3:54
  stolen artifacts and simple We call it culture.
• 3:57
  Wow. Yeah, that specific comparison in her post
• 3:59
  stopped me cold. And she takes it even further
• 4:02
  by pointing out the profound irony of the ancestry
• 4:05
  industry. Oh, the ancestry .com point is so sharp.
• 4:09
  Right. Think about paying private corporations
• 4:11
  just to access historical records about your
• 4:14
  own family. Yeah. Logically, public census and
• 4:17
  birth records should be free. Obviously. But
• 4:20
  this hits on a totally different level for black
• 4:22
  people who were systematically legally denied.
• 4:26
  documentation of their very existence for centuries.
• 4:29
  And now the descendants of those people have
• 4:32
  to pull out a credit card and pay a private corporation.
• 4:35
  just to access the scattered scraps of their
• 4:37
  own lineage. Which perfectly illustrates her
• 4:40
  refusal to be lectured. Cece is essentially asking,
• 4:43
  why should I accept a lecture on the ethics of
• 4:45
  data usage from a society that has spent centuries
• 4:48
  commodifying my data, my history, and my family?
• 4:50
  Exactly. It makes the internet's selective outrage
• 4:52
  look incredibly performative. People are screaming
• 4:55
  about a tech company reading their public tweets
• 4:57
  while entirely ignoring the systemic hoarding
• 5:00
  of actual human history. It makes me wonder if
• 5:02
  the intense backlash we're seeing is actually
• 5:05
  about the mechanics of the technology itself,
• 5:08
  or if AI is just a convenient scapegoat. Like,
• 5:11
  maybe people are just relieved that all this
• 5:13
  existential dread finally has a brand name they
• 5:16
  can point a finger at. Well, what's fascinating
• 5:18
  here is that idea that AI is basically absorbing
• 5:21
  the blame for much larger systemic anxieties
• 5:24
  is actually playing out verbatim in another one
• 5:27
  of our sources. The Reddit thread? Yeah, if we
• 5:29
  look at the r slash awars community on Reddit,
• 5:32
  there's this massive thread titled the thirsty
• 5:34
  hypocrisy. And it tackles this exact dynamic
• 5:38
  of selective outrage, but purely from an environmental
• 5:41
  standpoint. Wait, are we talking about the water
• 5:42
  consumption statistics? Yeah. Because I've seen
• 5:44
  those numbers floating around, and they sound
• 5:46
  terrifying on the surface. Yes. Critics of artificial
• 5:49
  intelligence have been heavily circulating the
• 5:51
  statistic that a platform like ChatGPT requires
• 5:54
  500 milliliters of water, so basically a standard
• 5:57
  plastic water bottle just to cool its servers
• 5:59
  for every single prompt a user enters. OK, wait.
• 6:02
  I need to pause here because that statistic gets
• 6:04
  thrown around without any explanation of the
• 6:07
  mechanics. Why is a chatbot drinking water? Right,
• 6:11
  it sounds crazy. It comes down to the thermodynamics
• 6:14
  of how these models actually operate. When you
• 6:17
  type a prompt into an AI, it isn't just pulling
• 6:19
  up a pre -written file. Massive arrays of graphics
• 6:23
  processing units, or GPUs, are performing billions
• 6:26
  of calculations in real time. To predict the
• 6:28
  next word. Exactly. And those processors run
• 6:31
  incredibly hot. To prevent the silicon inside
• 6:33
  the data centers from literally melting, these
• 6:36
  facilities rely on enormous evaporative cooling
• 6:39
  towers. So water is continuously being evaporated
• 6:42
  into the atmosphere to pull the heat away from
• 6:45
  the hardware. Okay, so the argument is that generating
• 6:48
  a digital painting or summarizing an email simply
• 6:51
  does not justify the physical evaporation of
• 6:53
  our freshwater supply, which I mean objectively
• 6:55
  sounds like a pretty valid concern. It does.
• 6:58
  But the pushback in that reddit thread from the
• 7:00
  pro -AI users is intense, isn't it? Oh, it is
• 7:03
  fierce. The users defending the technology start
• 7:06
  pulling up comparative macroeconomic data to
• 7:09
  put that 500 milliliters into perspective. Like
• 7:11
  what kind of data? Well, one user breaks down
• 7:14
  the numbers to show that all of these AI data
• 7:16
  centers combined actually use less water annually
• 7:19
  than a standard commercial pecan farm. A pecan
• 7:22
  farm? Yeah. And another user brings up the meat
• 7:25
  industry, highlighting that producing a single
• 7:27
  beef burger requires over 3 ,000 times the amount
• 7:31
  of energy and water as a single AI. See, that
• 7:34
  exposes the absurdity of the outrage. I haven't
• 7:36
  seen anyone staging massive protests outside
• 7:39
  of a local pecan pie bakery lately. Right. The
• 7:42
  pro -AI argument isn't necessarily that water
• 7:44
  conservation doesn't matter. It's that these
• 7:47
  critics didn't wake up yesterday with a sudden
• 7:49
  burning passion for global hydrology. They were
• 7:52
  just weaponizing environmentalism to attack a
• 7:54
  technology they already decided they hate. Exactly
• 7:56
  the point Cece was making. AI is just the newest
• 7:59
  punching bag for very old problems. And that
• 8:02
  brings us to the most visceral fear people have,
• 8:04
  right? Job displacement. Oh, for sure. Cece writes
• 8:07
  about this with incredible clarity. Before she
• 8:10
  was building digital media, she worked the third
• 8:12
  shift as a telephone operator. Typing numbers
• 8:15
  into computers for lonely elderly callers, right?
• 8:19
  And we all know how that story ends. That entire
• 8:21
  profession just vanished. Yeah, displaced by
• 8:24
  tech. But her point is how silently it happened.
• 8:27
  Nobody held a press conference to mourn the telephone
• 8:29
  operators. Society just absorbed the efficiency
• 8:32
  and moved on. As she writes in her post, AI did
• 8:36
  not invent displacement. It just finally gave
• 8:38
  it a name that people can argue about at dinner.
• 8:41
  Precisely. So her response isn't to boycott the
• 8:43
  technology out of fear. Instead, she is aggressively
• 8:46
  utilizing it. Cece uses large language models
• 8:49
  to buy herself the most scarce and valuable asset
• 8:52
  in the modern world, which is time. She's saving
• 8:55
  hours on administrative tasks. Exactly. So she
• 8:57
  has time to teach, time to create, and time to
• 9:00
  demonstrate to others, specifically black people
• 9:02
  and women, that financial freedom and mental
• 9:04
  liberation are actually the exact same project.
• 9:07
  She actually compares her use of these digital
• 9:09
  tools to an evolution of the Underground Railroad.
• 9:13
  The underlying philosophy is that true liberation
• 9:15
  always requires utilizing whatever infrastructure
• 9:19
  currently exists in your era. Right, you have
• 9:22
  to use the master's tools to build your way out.
• 9:24
  And right now, in this decade, that infrastructure
• 9:27
  includes artificial intelligence. It is a profound
• 9:30
  reframing. She acknowledges that the system is
• 9:33
  already broken, and the extraction of our data
• 9:35
  is already happening. Since you can't stop the
• 9:38
  machine, you must use the machine's capabilities
• 9:41
  to engineer your own freedom. And look. Because
• 9:44
  we are diving into incredibly deep, heavily debated
• 9:47
  waters today, touching on everything from global
• 9:50
  environmental policy to the history of American
• 9:52
  systemic inequality, I need to say this. It is
• 9:54
  important to remember our goal here. Yes, absolutely.
• 9:57
  We are not endorsing a specific political view
• 9:59
  or taking sides on these issues. Our mission
• 10:02
  is strictly to examine the raw arguments as they're
• 10:04
  laid out in these sources and see what they reveal
• 10:07
  about the human reaction to this technology.
• 10:09
  Right. And exploring this concept of AI as a
• 10:12
  mechanism for black liberation leads us directly
• 10:15
  into a massive ideological collision in our next
• 10:18
  source. It's a fiery panel discussion from a
• 10:21
  YouTube show on the Dr. Boyce Watkins Black Excellence
• 10:25
  Channel. Yeah, this panel was convened to react
• 10:27
  to a highly controversial statement made by the
• 10:29
  political commentator, Van Jones. He recently
• 10:32
  threw absolute gasoline on this debate by publicly
• 10:36
  claiming AI is the closest thing the reparations
• 10:39
  that black people are going to get. Which is
• 10:40
  a heavy claim. His logic is based on the idea
• 10:43
  of radical democratization. Van Jones argues
• 10:46
  that AI acts as an unprecedented societal equalizer.
• 10:49
  Because it's accessible? Right. You may not have
• 10:51
  had access to an elite private high school and
• 10:53
  you may have been locked out of a fancy Ivy League
• 10:55
  college, but today you have a free tool sitting
• 10:57
  on your phone that allows you to write, ideate,
• 11:00
  and function at the exact same level as someone
• 11:02
  who did. But the YouTube panel which features
• 11:04
  Tiara Williams, Yolanda Spivey, Mary Dean, and
• 11:07
  Micah. They absolutely tore that claim apart.
• 11:10
  I mean, they overwhelmingly rejected the idea
• 11:12
  that a piece of software could ever be equated
• 11:15
  to reparations. And the rejection was really
• 11:17
  rooted in the fundamental definition of the word.
• 11:20
  Yolanda Spivey makes a very sharp distinction
• 11:23
  during the discussion. She points out that artificial
• 11:25
  intelligence is simply a ubiquitous technology.
• 11:29
  Everyone on the planet with an internet connection
• 11:30
  is going to use it. Whereas reparations are totally
• 11:33
  different. Exactly. Reparations, by definition,
• 11:36
  are an owed debt designed to repair specific
• 11:39
  historical harm, intentional disenfranchisement,
• 11:42
  and systemic exclusion. You can't settle a ledger
• 11:45
  of centuries of stolen wages, stolen land, and
• 11:48
  intentional exclusion from wealth building engines
• 11:51
  like the GI Bill by handing someone a free app
• 11:54
  they can download from the iOS store. No, you
• 11:56
  can't. By calling a chatbot Reparations. Are
• 11:59
  we accidentally letting the actual institutions
• 12:01
  and governments that owe these massive, heavily
• 12:04
  documented historical debts completely off the
• 12:06
  hook? That is the exact danger the panel highlights.
• 12:09
  And if we connect this to the bigger picture,
• 12:11
  the conversation dives into the actual, tangible
• 12:14
  history of reparations in America, which is a
• 12:16
  history a lot of people are completely unaware
• 12:18
  of. They specifically pull up the 1866 Indian
• 12:21
  treaties. Yes, this part of the source material
• 12:23
  was incredibly dense, but so vital. Following
• 12:27
  the American Civil War, the U .S. government
• 12:29
  required the five civilized tribes, who had actively
• 12:32
  participated in chattel slavery, to not only
• 12:35
  free their enslaved people, but to provide them
• 12:37
  with legally binding reparations. Yeah, under
• 12:40
  these treaties, black freedmen and black Native
• 12:43
  Americans were legally entitled to very specific
• 12:46
  material repairs. It was 160 acres of land, right?
• 12:49
  160 acres, $150 in cash, and tax -exempt status.
• 12:54
  This was not a metaphor. It was a legal mandate
• 12:56
  for wealth transfer. But the panel breaks down
• 12:59
  how that legal mandate was systematically dismantled
• 13:02
  by the bureaucracy. How exactly did the government
• 13:04
  step in and block people from getting the land
• 13:07
  they were legally owed? It was done through administrative
• 13:09
  weaponization. The panel discusses how, primarily
• 13:12
  between 1938 and 1941, the federal government
• 13:15
  and the Bureau of Indian Affairs utilized a mechanism
• 13:18
  called the quantum blood law. The quantum blood
• 13:20
  law. Yes. To receive the benefits outlined in
• 13:23
  the 1866 treaties, bureaucrats required black
• 13:26
  folks to prove a very specific fractional percentage
• 13:29
  of Native American blood. Which is an impossible
• 13:32
  administrative trap. You're demanding a pristine,
• 13:35
  unbroken paper trail of ancestry from a population
• 13:38
  of people who are legally considered property
• 13:40
  and specifically denied birth records and documentation.
• 13:43
  Right. They weaponized genealogy to shut them
• 13:45
  out of the well. It's just devastating. It is.
• 13:48
  Which brings us back to the stark reality of
• 13:50
  today. If a ubiquitous digital tool is not true
• 13:54
  reparations, the panel asks a terrifying next
• 13:57
  question. Is this new AI landscape actually leveling
• 14:01
  the playing field? Or are we just eagerly feeding
• 14:04
  our data into a brand new, highly sophisticated
• 14:07
  extractive machine? The economic data they reference
• 14:10
  to answer that question is staggering. The panel
• 14:13
  points to a McKinsey report that predicts artificial
• 14:15
  intelligence could generate up to $13 trillion
• 14:17
  in global economic activity by the year 2030.
• 14:20
  Trillion with a T. $13 trillion. But right alongside
• 14:24
  that unfathomable number, they highlight the
• 14:26
  infrastructure reality. Currently, 59 % of black
• 14:29
  adults report lacking access to high -speed internet.
• 14:31
  That's a massive gap. And furthermore, only 4
• 14:35
  % of venture capital funding for AI startups
• 14:38
  is currently going to black founders. So the
• 14:41
  implication is massive. You cannot have a democratized
• 14:45
  level playing field when the demographic that
• 14:48
  is supposed to benefit the most lacks the basic
• 14:50
  physical infrastructure to access the tool and
• 14:53
  is simultaneously locked out of the boardrooms
• 14:56
  where the technology is being funded and built.
• 14:58
  It feels like the old digital divide from the
• 15:00
  early Internet days is rapidly mutating into
• 15:03
  an intelligence divide. And if you're locked
• 15:05
  out of the boardroom with AI is being built,
• 15:07
  you aren't the builder. You're just the raw material.
• 15:10
  Exactly. Which brings us to the concept of mind
• 15:13
  cloning, which the panel discusses, and it sounds
• 15:16
  completely like science fiction. It does sound
• 15:18
  like a sci -fi movie, but Micah, the AI expert
• 15:21
  on the panel, breaks down how it is happening
• 15:22
  right now in the corporate world. High -end consultants
• 15:25
  are approaching burnt -out CEOs and offering
• 15:28
  to create a highly specialized AI clone of them
• 15:31
  to handle their workload. Okay, I need to know
• 15:33
  how this actually functions. How do you clone
• 15:35
  a CEO's mind into a machine? Well, it relies
• 15:38
  on tokenization and predictive modeling. The
• 15:41
  consultants take massive databases of the CEO's
• 15:44
  proprietary data years, of their outgoing emails,
• 15:47
  hours of their past video calls, internal memos,
• 15:49
  and their written decision -making frameworks.
• 15:51
  So everything they've ever output at work. Yeah.
• 15:54
  The AI ingests all of this, vectorizes the data,
• 15:58
  and learns to predict the exact syntax, tone,
• 16:01
  and logic pathways the CEO uses. So the AI isn't
• 16:05
  just mimicking their voice. It has internalize
• 16:07
  their business logic. Instead of the human CEO
• 16:10
  taking a mundane Tuesday morning meeting, the
• 16:13
  AI clone logs into the Zoom call, speaks with
• 16:16
  their voice, and makes strategic decisions exactly
• 16:19
  how the human would. based on historical data.
• 16:22
  That is exactly how it works. But Mary Dean,
• 16:24
  the lawyer on the panel, throws up a massive
• 16:26
  red flag about ownership here. IP rights. Yeah,
• 16:29
  who actually owns the neural network that holds
• 16:30
  all of that synthesized intelligence. She heavily
• 16:33
  warns the audience about the boilerplate terms
• 16:35
  and conditions that every single one of us clicks
• 16:37
  agree on without reading. Oh, guilty. Right.
• 16:41
  Users are unknowingly signing away their intellectual
• 16:44
  property and their commercial image rights just
• 16:46
  to gain access to these shiny new software platforms.
• 16:50
  So you want to use the cool new tool to summarize
• 16:52
  your emails. But the hidden price of admission
• 16:55
  is granting a tech giant a perpetual license
• 16:58
  to your own face and thoughts. Exactly. And Yolana
• 17:02
  Spivey takes that fear of extraction even further.
• 17:04
  She argues that AI companies are actively studying
• 17:07
  human behavior and specifically scrape the internet
• 17:10
  for black culture, slang, tone and flavor. To
• 17:14
  train the models. Right. They feed all of that
• 17:16
  cultural data into the bots to make them sound
• 17:18
  more authentic, relatable and human. It is a
• 17:21
  massive concern regarding cultural extraction
• 17:23
  without compensation. The machine requires human
• 17:26
  soul and culture to appear intelligent, but the
• 17:29
  human beings generating that culture are legally
• 17:31
  signing away their rights to it in the fine print.
• 17:33
  It's digital strip mining. It's like a corporation
• 17:36
  clear -cutting a beautiful ancient forest to
• 17:39
  build a massive subdivision and then naming all
• 17:42
  the new paid streets after the trees they just
• 17:44
  destroyed. They are extracting the flavor and
• 17:47
  syntax of black culture to train the neural network,
• 17:50
  but cutting the actual creators completely out
• 17:53
  of the equity and the profits. That is a perfect
• 17:56
  way to visualize it. It's the ultimate paradox
• 17:58
  of the AI era. So what does this all mean? We've
• 18:00
  looked at this technology from so many different
• 18:03
  conflicting angles today. Well, when you pull
• 18:05
  all these threads together, the fierce numbers
• 18:08
  -driven debates on Reddit about environmental
• 18:10
  hypocrisy, the intense historical and legal discussions
• 18:13
  on YouTube about reparations and mine cloning,
• 18:16
  and CC's deeply personal 500 -year family tree
• 18:19
  project, One undeniable truth emerges from the
• 18:22
  source material, which is artificial intelligence
• 18:25
  is not and never will be a neutral tool. It is
• 18:27
  just a mirror. Exactly. It acts as a massive
• 18:30
  computational mirror. It is reflecting our existing
• 18:34
  societal debts, our deeply ingrained biases,
• 18:37
  our history of extraction and our urgent desires
• 18:39
  for freedom. It simply magnifies whatever human
• 18:42
  nature is already there, which brings us right
• 18:44
  back to Cece in Brazil and her blog post. Her
• 18:48
  approach isn't naive. She isn't ignoring the
• 18:51
  dangers Mary Dean or Yolanda Spivey brought up.
• 18:53
  Not at all. Her approach is the ultimate pragmatic
• 18:56
  survival strategy. She knows the system is flawed.
• 19:00
  She knows the massive corporate databases have
• 19:02
  already scraped her decades of writing and podcasting
• 19:05
  without paying her. Right. But she flat out refuses
• 19:07
  to be left behind while everyone else argues
• 19:10
  about the ethics at the dinner table. She is
• 19:13
  confidently, unapologetically taking the exact
• 19:16
  tools that are available right now to build her
• 19:18
  own systems of freedom and document her own history.
• 19:21
  She represents individual agency in an era where
• 19:24
  it constantly feels like the machines and the
• 19:26
  massive corporations behind them are taking all
• 19:29
  of our agency away. If you want to read her brilliant
• 19:31
  unapologetic perspective in her own words, and
• 19:34
  I highly suggest that you do, you need to go
• 19:36
  read the full blog post. You can find it by going
• 19:38
  to faafo .appp. Wait, can you spell that one
• 19:42
  more time? Yeah, that is faafo .acp. Awesome.
• 19:46
  Understanding her perspective is really essential
• 19:48
  for grasping the human element behind these vast,
• 19:52
  intimidating technological shifts. Before we
• 19:54
  go, think about this. We spend so much time worrying
• 19:58
  about whether AI is going to clone our voices,
• 20:00
  steal our jobs, or drain our reservoirs. But
• 20:02
  if you were given the ability to mind -clone
• 20:04
  all of your own lived experiences, your family
• 20:07
  struggles, and your generational knowledge into
• 20:09
  a digital format right now, would you trust the
• 20:12
  machine to hold your history, or would you keep
• 20:14
  it to yourself?