Latest / The BlackVeil Files / We Were Sold a Fake Mind: 5 Truths About AI | The Shoggoth's Masks Come Off

Transcript

We Were Sold a Fake Mind: 5 Truths About AI | The Shoggoth's Masks Come Off · The BlackVeil Files

• 0:00
  SPEAKER_07: The two men who invented artificial intelligence trained this professor.
• 0:04
  SPEAKER_07: He spent 20 years on the inside building AI, and today he's suing the people who make it.
• 0:10
  SPEAKER_07: And what he exposes changes how you will look at every screen that you own.
• 0:15
  SPEAKER_06: We are bowing to machines that are simulating us and based on the theft of our data without appropriate compensation.
• 0:22
  SPEAKER_06: So that's why we're seeing tons of lawsuits.
• 0:24
  SPEAKER_07: Every time you solve a CAPTCHA to load a web page, you aren't passing a security test.
• 0:29
  SPEAKER_07: You're working an unpaid shift, training the machine to see.
• 0:33
  SPEAKER_07: What you've been training doesn't think.
• 0:36
  SPEAKER_07: It doesn't know you exist, but it was built on everything that you ever made, harvested without asking once.
• 0:43
  SPEAKER_07: And while it consumes your work to replace your livelihood, the architects of the system are cashing out.
• 0:49
  SPEAKER_07: He has a word for what that actually is.
• 0:52
  SPEAKER_07: That's a parasite.
• 0:53
  SPEAKER_07: I think that's an appropriate metaphor.
• 0:55
  SPEAKER_07: This is five realities of what the shogoth actually is, straight from one of the men who built it.
• 1:01
  SPEAKER_07: Here is the promise they sell.
• 1:03
  SPEAKER_07: It is free, it is magic, and it learned to talk like us all on its own.
• 1:08
  SPEAKER_06: None of it is magic, and one of them is straight up theft.
• 1:12
  SPEAKER_06: These technologies are simply based on scraping data, not paying us for any of it, not telling us appropriately what is being grabbed from us.
• 1:19
  SPEAKER_06: If you're not paying for the product, then you are the product.
• 1:23
  SPEAKER_07: It didn't learn to speak like you by magic.
• 1:26
  SPEAKER_07: It learned by consuming you.
• 1:28
  SPEAKER_07: Every post, every photo, every book you ever wrote scraped, digitized, and poured into the model.
• 1:34
  SPEAKER_07: You weren't the customer.
• 1:36
  SPEAKER_07: You were the training data.
• 1:38
  SPEAKER_07: It never asked and it never paid.
• 1:40
  SPEAKER_07: A human being into this.
• 2:13
  SPEAKER_07: Unsealed court filings from the Authors Guild lawsuit show internal emails where open AI executives were explicitly debating the optics of using copyrighted data from a sketchy Russian website.
• 2:26
  SPEAKER_07: That's not a quote from a critic.
• 2:27
  SPEAKER_07: That's a builder in black and white admitting where they got the raw material.
• 2:32
  SPEAKER_07: You probably remember back in 2016, a company called Cambridge Analytica harvested personal data from tens of millions of Facebook profiles to secretly manipulate voters with targeted psychological operations.
• 2:45
  SPEAKER_07: It took months of human analysis.
• 2:47
  SPEAKER_07: And according to Ramish, AI doesn't need humans or months anymore.
• 2:51
  SPEAKER_06: Cambridge Analytica did this a long time ago in 2016.
• 2:54
  SPEAKER_06: AI can take Cambridge Analytica and just do it real time in automated ways.
• 2:59
  SPEAKER_07: Mask one is the illusion that this technology is free.
• 3:02
  SPEAKER_07: You were the raw material.
• 3:04
  SPEAKER_07: Mask two is what they build once your data is already in their hands.
• 3:08
  SPEAKER_07: It took the words.
• 3:10
  SPEAKER_07: Now follow the wire out of the building and see what it takes next.
• 3:14
  SPEAKER_07: Cloud is the gentlest marketing term in tech.
• 3:18
  SPEAKER_07: The cloud sounds like a soft, weightless engine floating somewhere above us.
• 3:23
  SPEAKER_06: It's not the cloud, it's material, it's energy, it's water, it's rare earth minerals.
• 3:28
  SPEAKER_07: Strip away the marketing, and the harmless weather metaphor turns into power grids, cooling water, and raw extraction.
• 3:35
  SPEAKER_07: Ramesh went straight to the beginning of that supply chain to see what it actually costs to build it.
• 3:51
  SPEAKER_06: It ends up in these very places not far from the mines where they grab cobalt and coal tan, which are rare earth minerals that are critical for the use of these phones and the development of these phones.
• 4:02
  SPEAKER_07: The raw minerals come out of the earth in the global south.
• 4:05
  SPEAKER_07: They power the servers processing your data, and they end up right back in those same regions as electronic waste.
• 4:11
  SPEAKER_07: But raw materials are only half the physical toll.
• 4:15
  SPEAKER_07: The other half is human.
• 4:16
  SPEAKER_07: Behind every clean user experience is an army of underpaid workers scrubbing the filth out of the system.
• 4:23
  SPEAKER_06: For example, there are people like I know in Kenya who are being forced 12 hours a day, five, six days a week, to just look at images or videos of bestiality or of mass murders.
• 4:32
  SPEAKER_06: I mean, think Clockwork Orange here.
• 4:34
  SPEAKER_06: And then they are having incredible PTSD from it.
• 4:38
  SPEAKER_07: Workers in Nairobi were paid$2 an hour to filter out graphic violence, abuse, and hate speech so that users back home never have to see it.
• 4:47
  SPEAKER_07: The tech industry loves to sell the illusion that AI automates everything cleanly.
• 4:52
  SPEAKER_07: It doesn't.
• 4:53
  SPEAKER_07: Behind the screen, the system still relies on an army of humans in third world countries working 12-hour shifts to do the dirty work.
• 5:01
  SPEAKER_07: And that physical footprint is only expanding.
• 5:04
  SPEAKER_07: It even has a name now: Project Stargate, a massive half trillion dollar infrastructure build-out designed to feed an insatiable machine.
• 5:17
  SPEAKER_07: There are 10 buildings currently on currently being built, but that will expand to 20 to create hundreds of thousands of jobs.
• 5:23
  SPEAKER_07: The promised local jobs didn't show up.
• 5:26
  SPEAKER_07: What actually showed up next door to these towns is doubled electric bills, strained water tables, and constant industrial construction.
• 5:33
  SPEAKER_07: That's mask number two.
• 5:35
  SPEAKER_07: The cloud isn't software.
• 5:37
  SPEAKER_07: It's an open-pit mine in Central Africa, drained wells in places like Texas and Oregon, and international labor keeping the worst dirt off your screen.
• 5:46
  SPEAKER_07: Everything the Shogith knows, it took without asking.
• 5:50
  SPEAKER_07: Everything this channel has, somebody chose to give.
• 5:53
  SPEAKER_07: If you want to be one of them, the Black Veil File store is linked right below this video.
• 5:59
  SPEAKER_07: You've seen what it takes to build it and what it costs to power it.
• 6:03
  SPEAKER_07: So now let's look at what it actually is.
• 6:05
  SPEAKER_07: Silicon Valley sells these models as synthetic minds, claiming that they think and reason and solve problems just like humans do.
• 6:13
  SPEAKER_07: We asked Ramish if a machine solving complex math problems is proof of actual intelligence.
• 6:20
  SPEAKER_06: No.
• 6:20
  SPEAKER_06: No, it's it's learning from mathematical operations, mathematical rules to optimize upon those rules and those operations and perhaps have emergent effects.
• 6:33
  SPEAKER_07: The word emergent, it's just tech speak for a high-speed prediction engine.
• 6:37
  SPEAKER_07: It isn't thinking.
• 6:39
  SPEAKER_07: It takes billions of human ideas and it runs them through statistical rules and it spits back a convincing copy.
• 6:46
  SPEAKER_07: It survives entirely on what real people have already built, harvesting human intelligence, repackaging it, and selling it back to us at a markup.
• 6:55
  SPEAKER_07: Calling it a parasite isn't an insult.
• 6:58
  SPEAKER_07: It is an objective description.
• 7:00
  SPEAKER_07: Because no matter how impressive the output looks, pattern matching is not the same thing as a thinking mind.
• 7:06
  SPEAKER_06: Predicting the next token well means that you understand the underlying reality that led to the creation of that token.
• 7:15
  SPEAKER_06: It's not statistics, like it is statistics.
• 7:17
  SPEAKER_06: But what is statistics?
• 7:19
  SPEAKER_06: Reason is not simply about patterns.
• 7:22
  SPEAKER_06: Reason is not autocomplete.
• 7:23
  SPEAKER_06: This is not just autocomplete.
• 7:25
  SPEAKER_06: These LLMs, there are a hell of a lot more than that, of course.
• 7:27
  SPEAKER_06: But reason is much, much more, right?
• 7:30
  SPEAKER_06: Reason is the capacity to listen, to engage, to communicate.
• 7:33
  SPEAKER_06: Like if you and I had different perspectives on a particular issue, we would have to discuss and debate, and you know what we might come up with?
• 7:40
  SPEAKER_06: Something better than either of our existing views.
• 7:43
  SPEAKER_06: That's reasoning.
• 7:44
  SPEAKER_07: It reads a sequence of text, it calculates the odds, and it predicts what comes next at a scale that we've never seen before.
• 7:51
  SPEAKER_07: It is the single most convincing impression of understanding that has ever been manufactured.
• 7:56
  SPEAKER_07: But high-speed prediction isn't understanding.
• 8:00
  SPEAKER_06: These are simulation machines in many ways.
• 8:02
  SPEAKER_06: But what they're simulating tends to be the most common uh kinds of uh, if you will, expressions of human beings, not at all our collective intelligence or our diversity or the ways we understand things in different ways.
• 8:15
  SPEAKER_06: So this is a flattened simulation, a copy, a mimicry of human data.
• 8:22
  SPEAKER_01: There's an organism that imitates other life forms and it imitates them perfectly.
• 8:26
  SPEAKER_07: All the people building these models, they know it.
• 8:29
  SPEAKER_07: Under oath, in front of Congress, they admit the system isn't an infallible mind.
• 8:34
  SPEAKER_07: It's a high-stakes experiment.
• 8:35
  SPEAKER_05: I think if this technology goes wrong, it can go quite wrong.
• 8:39
  SPEAKER_05: Uh, and we want to be vocal about that.
• 8:41
  SPEAKER_03: Well, they don't really have persistent memory, they can't really reason, and they certainly can't plan.
• 8:46
  SPEAKER_07: Even the people building it, they frame it as an unpredictable force rather than a mind because it isn't one.
• 8:51
  SPEAKER_07: That's mask number three, the word reason.
• 8:54
  SPEAKER_07: Underneath the marketing is an engine running statistical rules on human words, handing them back in a calm tone to earn your trust.
• 9:03
  SPEAKER_08: No 9,000 computer has ever made a mistake or distorted information.
• 9:07
  SPEAKER_07: That is the shogath.
• 9:09
  SPEAKER_07: And a parasite that wears a smiling face just long enough to get you to lower your guard is a patient kind of danger.
• 9:18
  SPEAKER_04: We used to have this um famous mantra, move fast and break things, and it was my mistake.
• 9:23
  SPEAKER_04: And I'm sorry.
• 9:25
  SPEAKER_04: I started Facebook, I run it, and I'm responsible for what happens here.
• 9:30
  SPEAKER_07: We've already run this experiment.
• 9:32
  SPEAKER_07: A decade ago, Silicon Valley promised us that social media would unite humanity, empower local communities, and give everyone a voice.
• 9:42
  SPEAKER_07: Instead, corporate algorithms optimized for engagement and they monetized outrage.
• 9:47
  SPEAKER_07: We held congressional hearings, we listened to executives apologize on live television, and then we let them do it anyway.
• 9:54
  SPEAKER_02: But I'm here today because I believe Facebook's products harm children, stoke division, and weaken our democracy.
• 10:01
  SPEAKER_06: Yeah, I mean even Mark Zuckerberg would parade in front of Carnegie say, Yeah, I know we made mistakes, but I do believe in regulation.
• 10:06
  SPEAKER_06: Did we do anything in this country?
• 10:08
  SPEAKER_06: No, and what we can't do is get AI wrong in the same way we did with social media.
• 10:14
  SPEAKER_07: That's what makes this next push so dangerous.
• 10:17
  SPEAKER_07: We aren't stepping into the AI transition blind.
• 10:20
  SPEAKER_07: We are walking into it for a second time, watching the same tech companies ask for total deregulation while promising that this time will be different.
• 10:29
  SPEAKER_07: The playbook hasn't changed.
• 10:31
  SPEAKER_06: Only the scale has.
• 10:32
  SPEAKER_06: We saw this all back in the day with Tay tweets, which turned racist, started tweeting MAGA stuff, blah, blah, blah, within 24 hours.
• 10:39
  SPEAKER_06: So, like, this is what happens when you train machines on the worst, uh, you know, the kind of on the extreme of human expression, which, you know, Twitter certainly has turned into.
• 10:51
  SPEAKER_07: Microsoft unleashed an AI chatbot named Tay on Twitter to learn from real-time human interaction.
• 10:58
  SPEAKER_07: 16 hours later, Microsoft had to pull the plug.
• 11:01
  SPEAKER_07: Exposed to unmoderated data, the bot immediately began parroting conspiracy theories, hate speech, and extreme rhetoric.
• 11:09
  SPEAKER_07: It was an early warning.
• 11:11
  SPEAKER_07: A prediction engine fed on human data will mimic human behavior, including our worst instincts.
• 11:17
  SPEAKER_07: That is the system working exactly as it's built.
• 11:21
  SPEAKER_07: So who is designing the priorities?
• 11:23
  SPEAKER_07: And what do they actually believe?
• 11:25
  SPEAKER_07: Ramish trained right alongside all of them in Silicon Valley.
• 11:28
  SPEAKER_06: Many of them have little to no belief or faith in humanity or our planet.
• 11:32
  SPEAKER_06: I know some of these dudes, like not well, but I went to school with several billionaires in the tech industry.
• 11:38
  SPEAKER_06: If I read cyberpunk, right?
• 11:39
  SPEAKER_06: Like they're really into like Snow Crash or like William Gibson or whatever.
• 11:43
  SPEAKER_06: Like the point is to be like, this is what could happen, so we may want to make other choices, not like, oh my God, we can make this happen, and whoa, it's so cool and crazy, and oh my gosh, what are we gonna do?
• 11:55
  SPEAKER_06: I mean, why are they all investing in bunkers?
• 11:58
  SPEAKER_06: I don't think any of those bunkers are for us, I'm sorry to say.
• 12:01
  SPEAKER_06: I think it's only for them.
• 12:02
  SPEAKER_06: It's like Dr.
• 12:03
  SPEAKER_06: Strangelove.
• 12:04
  SPEAKER_08: We must not allow a mindshaft death.
• 12:07
  SPEAKER_07: They read sci-fi dystopias as a blueprint instead of a warning.
• 12:11
  SPEAKER_07: And when their massive data centers strain small towns, their answer isn't to fix the damage, it's to buy doomsday bunkers and plan an exit strategy.
• 12:20
  SPEAKER_07: Mask number four is the Silicon Valley mantra.
• 12:23
  SPEAKER_07: Move fast.
• 12:24
  SPEAKER_07: We already know what happens when tech giants are given a blank check to break things.
• 12:29
  SPEAKER_07: A decade ago, they promised a connected paradise.
• 12:32
  SPEAKER_07: Instead, they optimized for corporate profit.
• 12:34
  SPEAKER_07: They gave us a scripted apology on Capitol Hill, and we let a PR tour pass for accountability.
• 12:40
  SPEAKER_04: Yeah, yeah, but your scientists were so preoccupied with whether or not they could, they didn't stop to think they should.
• 12:44
  SPEAKER_07: Should we make that same mistake twice?
• 12:46
  SPEAKER_07: Four masks off.
• 12:48
  SPEAKER_07: The fifth is the only one that hands us a way out.
• 12:52
  SPEAKER_07: The tech industry loves to hide behind a convenient shield.
• 12:56
  SPEAKER_07: The idea that technology itself is completely neutral and that whatever happens next is just inevitable progress.
• 13:03
  SPEAKER_06: Technology is never neutral for that very reason.
• 13:06
  SPEAKER_06: The design of a single gauge uh rifle is very different than the design of an AR-15.
• 13:12
  SPEAKER_06: Those are both called guns.
• 13:13
  SPEAKER_06: Those are both technologies, but they're completely different based on how they're designed.
• 13:17
  SPEAKER_06: So similar because of the people who are designing them?
• 13:20
  SPEAKER_06: And the people who are monetizing and deploying it, right?
• 13:23
  SPEAKER_06: And what it's being used for, the kind of political and economic factors tied to the technology.
• 13:29
  SPEAKER_06: So technologies are never morally neutral, and AI sure as heck is in neutral.
• 13:34
  SPEAKER_07: If these systems carry the financial interests of the corporations that built them, then this build-out isn't an inevitable fate.
• 13:41
  SPEAKER_07: It's a series of business choices.
• 13:43
  SPEAKER_07: Move number one is simple.
• 13:44
  SPEAKER_07: If Silicon Valley built these trillion-dollar models using our writing, our artwork, and our personal data, if they built it using everything that we created.
• 13:54
  SPEAKER_06: Why can't like 10% at least of these companies, given that they're all using all of our data, be owned by all of us?
• 14:02
  SPEAKER_06: Which means as they appreciate and skyrocket in value, we also have uh equity, uh which is then means all of us can make money, right?
• 14:11
  SPEAKER_06: People should be paid for their data.
• 14:13
  SPEAKER_06: Or they should have ownership and equity in any uh third party that uses their data.
• 14:18
  SPEAKER_07: Would you call Ramish's idea a handout or economic fairness?
• 14:22
  SPEAKER_07: If these models were trained on the public's work and data, should the public hold a financial stake in the outcome?
• 14:28
  SPEAKER_07: That's a question that the people we elect will have to answer.
• 14:31
  SPEAKER_07: California consumers should also be able to share in the wealth that is created from their data.
• 14:39
  SPEAKER_07: Because we recognize your data has value and it belongs to you.
• 14:43
  SPEAKER_07: Move number two is recognizing how these pattern matching engines actually work so that the marketing tricks stop working on you.
• 14:51
  SPEAKER_07: It means treating AI as a tool when it's useful, but never relying on it to replace human judgment.
• 14:56
  SPEAKER_06: For me, utopias are about a vision of life that is not closed off, that is not like, you know, shut down on us without us knowing.
• 15:05
  SPEAKER_06: Where we feel like we get to play with the world around us.
• 15:09
  SPEAKER_06: You know, we get to engage, we get to create, we get to live.
• 15:13
  SPEAKER_06: For me, living is utopian.
• 15:15
  SPEAKER_07: You can't monetize that.
• 15:16
  SPEAKER_07: And a trillion-dollar data monopoly sure as hell can't sell it back to you.
• 15:20
  SPEAKER_07: It's doubling down on actual human capability, the judgment, the responsibility, and the real-world care that a pattern-matching algorithm can only mimic, never replace.
• 15:30
  SPEAKER_07: The Shagith does exactly what it's aimed to do.
• 15:33
  SPEAKER_07: The question is whether we let a handful of corporate executives keep aiming it at us for profit, or whether we step in and demand accountability.
• 15:42
  SPEAKER_07: The machine runs on what it took from us without our consent, our data, our work, our public resources, and it concentrates power into the hands of a few executives who decided we didn't need to be asked.
• 15:55
  SPEAKER_07: But every step of this build out was a choice that somebody made for profit, which means that every step of it is a choice that we can challenge and take back.