Health Ranger - Mike Adams - AI cognition keeps PLUMMETING in price while human costs keep rising

All Episodes

Previous Next

Nov. 5, 2025 - Health Ranger - Mike Adams

19:08

AI cognition keeps PLUMMETING in price while human costs keep rising

Large

Audio Only |

Scroll

Time	Text
	There's some very important trends in AI that you need to be aware of.
	And I know that not everybody is thrilled about this phenomenon.
	The fact that AI is rising so quickly and that it's replacing so many human jobs.
	It has begun.
	It's accelerating.
	I don't know if you saw, but Goldman Sachs put out a new investigative report that talks about the fact that about a quarter of all white-collar jobs can be replaced by AI right now.
	And those are concentrated in areas like customer service or sales consultations, but also in areas like legal and journalism and engineering and architecture, etc.
	I mean, it's a shocking across the board analysis of how quickly AI is going to replace human jobs.
	And again, I know that not everybody's thrilled about this.
	I know that it's scary for many people because you might be thinking, well, is AI going to replace my job?
	Well, my job, this is my job.
	My job is to help you not be replaceable.
	My job is to help you understand what AI is doing so that you can upgrade your own skills or your own qualities that AI can't touch.
	I want you to be successful.
	I want you to continue to earn income or pursue your passion or whatever it is that you do.
	And it's critical to understand what's happening in the world of AI in order to be able to navigate all of this.
	And I also think, by the way, as a side note, it's critical to use AI to enhance what you do.
	We're at the point in the job market, and I've said this before, if I were hiring, although I'm not, I'm not the HR manager of my own companies, but if I were, I would not hire anybody who wasn't using AI actively to do their job.
	If somebody comes in and says, oh, I'm the best middle manager.
	I know everything about this industry.
	I know how to do spreadsheets and inventory management, supply chain management, everything, but I don't use AI and I don't want to use AI.
	I will never use AI.
	I would say, well, then we can't hire you because there's no way that you're going to be able to keep up.
	It's impossible.
	You're going to have to learn AI to do your job competitively.
	You know, someone saying right now that, oh, I don't want to use AI.
	It would be like somebody in 1997 saying, well, I don't want to use the internet.
	The internet's a fad.
	It's not going to last.
	We don't need the internet.
	I just do everything on Microsoft Word or whatever.
	It's like, if you don't use the Internet, you're not going to be relevant to the economy that is emerging.
	And the same thing is true with AI today.
	So what's the big trend that's happening that you need to be aware of?
	Well, there are a couple of them, but first let me start with the fact that the cost of tokens is plummeting.
	So what are tokens?
	Well, tokens are representations of either words or images.
	Tokens are the unit of communication and also of the internal thinking of reasoning models.
	So for the most part, we see language models producing words.
	Well, those words consist of tokens.
	And the tokens are what are represented in the large vector database, which is actually a massive multi-dimensional vector database that describes the relationships between words and tokens based on larger context of potentially thousands of words and tokens.
	And I can't go into all the details of transformers and how it all works, and I don't even know how all of it works.
	But all you need to know is that tokens are what the language models are generating.
	When they're outputting text or they're outputting images, they are essentially creating tokens, or you could sometimes say they're using tokens.
	All right.
	Now, every graphics card can only produce so many tokens per second.
	And although this varies, a lot of language models today might produce 25 to 75 tokens a second, let's say.
	That's pretty typical for a relatively high-end prosumer graphics card running a common language model.
	Now, the bigger your language model gets, that is, the larger the parameters are, and some models are 70 billion parameters or more, the slower the models get.
	And then they can produce fewer tokens per second.
	Whereas smaller models, such as a 7 billion parameter model, which is very common, a 7 billion parameter model can produce tokens very quickly.
	And that's why running a smaller model is less expensive per token, because the tokens come out faster.
	All right, now, the trend is that the cost of tokens is plummeting.
	Over the last two years, the cost to produce tokens is just plunging so much that tokens are getting so unbelievably cheap that users of AI models are finding that they can just sort of throw tokens at a problem and burn a bunch of tokens and get a much better answer.
	And one approach to this is called recursive reasoning.
	It's where you tell a model to take its own answer and then to reprocess that answer as a new input and then to rethink about that and refine it and maybe correct it if it's off base.
	And by looping like a feedback loop of the AI model's own output, looping it back into itself, again, that's why it's called recursive reasoning, you can burn a bunch of tokens and get a much better answer.
	It turns out that AI models are just so-so with a one-shot answer.
	A one-shot answer means like, this is my first answer.
	You know, it's kind of like walking up to a person and saying, you know, what's the capital of Argentina?
	You know, and they're like, they just give an answer because not everybody knows that.
	I mean, I don't know the capitals of most countries around because it's just not important.
	I can look it up if I have to, right?
	But if you give a one-shot answer, it's not as good as if you think about the answer.
	Like, wait a second, let me think about this.
	Or let me sort of research it.
	Let me ponder it.
	And then you can get a better answer.
	That's true with humans.
	And it's true with AI models.
	And so more tokens means better thinking, better output.
	And the reason this is critical, as the cost of tokens is plunging, and as the number of tokens thrown at projects is rising, it allows AI models to more effectively replace the cognition of humans.
	That's why the human replacement curve is accelerating.
	Because tokens are getting cheaper, but the cost of human cognition is not getting cheaper.
	So the cost of human cognition, if you were to think about a person's output in a work environment, think about what the person does.
	Oh, person's writing emails.
	The person is, I don't know, moving around documents and spreadsheets and entering numbers and reading this and trying to understand that and writing reports, whatever they do as a middle manager.
	If you think about that as using tokens, the human brain is a really amazing model of neural networking and intelligence, but the cost of the human tokens is not going down.
	It's still about the same.
	The human employee needs a salary, needs health care, needs time off.
	Actually, you know, most of the 24 hours, that person is not working at the office, obviously.
	That person needs vacation days, sick days, etc.
	And then when that person is actually at work, they're shooting the breeze at the coffee machine with the co-workers or whatever.
	The actual number of productive hours in a typical office day is something like two or three, maybe four if you have a really well-run office.
	And that means that human tokens are incredibly expensive and incredibly slow.
	Whereas machine tokens are getting cheaper and cheaper and faster and faster.
	And I know some of you are going to say, well, you can't reduce humans to tokens.
	They're not robots.
	Humans bring something special.
	They have innovation and creativity and inspiration and compassion.
	And yes, that's true, but that's not what your company is hiring you for.
	Your company, I mean, most companies hire workers to function as automatons.
	They want you to be a thinking machine.
	They want you process these reports.
	Or what was the joke in that movie, The Office?
	Was it the TPI reports or whatever?
	It's like, we need the TPI reports.
	Nobody ever knew what those were.
	Just some paperwork garbage.
	Well, most jobs require humans to become machines.
	Most employers are not rewarding humans for your humanity.
	They're not saying, oh, we love your creativity.
	We love your vision.
	No, produce the TPI reports.
	That's the issue.
	So if you're lucky enough to have a job where you get to be creative and you get to express your innovation, hold on to that job because those are pretty rare, actually.
	And those will be very difficult for AI to replace.
	Whereas the more automated jobs, oh, here's a spreadsheet, you know, or here's a stack of grant applications.
	You know, process these grant applications.
	Oh, my God.
	You know, that's turning a human into a machine to process the grant applications.
	Well, that can be done by machines more cheaply and usually with better accuracy.
	So cheaper tokens for machines, but more expensive tokens for humans.
	So what do you think employers are going to do?
	They're going to cut the humans.
	And that's exactly what's happening.
	And you got to understand, human tokens are not going to get cheaper.
	But machine tokens will continue to get lower and lower in cost.
	I mean, dramatically.
	And machine tokens are going to get faster.
	And at the same time, the quality of the machine output continues to improve dramatically.
	And this is not a linear progression.
	This is an exponential increase in the functional intelligence of AI models.
	I mean, we're in the middle of an intelligence explosion that very few humans can even properly understand.
	They're just not tracking it.
	It's hard for humans to understand exponential phenomena, but that's what this is.
	And as a result, it's going to have runaway intelligence so incredibly rapidly that even CEOs of corporations are going to find themselves answering to the AI rather than commanding the AI.
	Yeah, that's going to happen pretty rapidly.
	So the only way to make human tokens less expensive and higher quality is for humans to work for less money, fewer benefits, et cetera, or for humans to upgrade their skills and their knowledge, especially domain knowledge in your particular industry, so that your output, or that is your thinking, is higher quality than AI.
	And for the moment, that's possible in many areas, but that's not going to continue to be possible very soon because AI will have more domain knowledge in every domain.
	AI will be the smartest entity in the room by far.
	And compared to AI, every human will look like an idiot.
	And so in 2026, you're going to see a horrifying acceleration of layoffs of human workers.
	Now, it has already begun here in late 2025, but it will only accelerate and it will absolutely surprise many people who have been in denial about this.
	And the thing that's different about being laid off or fired from your job today compared to any other time is that it's now extremely difficult to get another job.
	Extremely difficult.
	Whereas, you know, previously, especially an engineer, let's say, who got laid off from Amazon or Google or whatever, they could go out and get a new job pretty easily because they've got a skill set.
	You know, they learned how to code.
	They know a lot of things.
	They've got value to a lot of tech companies, et cetera.
	They would go out, look for a new job, land a new job in a matter of weeks.
	Today, the people that get fired, they are looking for jobs for months.
	And in some cases, people have been looking for jobs for over a year.
	And you're seeing a sense of desperation set in for a lot of people.
	For example, you're seeing people launching GoFundMe, you know, fundraising donations.
	Online begging is really what that is, begging for money to be able to continue to live, to buy food, etc.
	You're seeing more and more of that.
	And in some cases, people who have very accomplished college degrees.
	And in this economy, by the way, it turns out that having a degree from a university is not necessarily any kind of a significant advantage.
	It's just that that's almost irrelevant at this point in many areas, or maybe most areas.
	Because again, everything that a person learns in college is already known by almost every AI engine.
	Every possible thing that you could have learned from your college books.
	Well, all those textbooks have already been read and ingested by AI.
	You know, college is less and less relevant to the workplace.
	Maybe for a long time, it hasn't been relevant because of all the woke universities and things like that.
	But it's even less relevant today.
	So you're going to see a wave of unemployment that you have never seen before, not in our lifetimes.
	And in time, it will exceed the Great Depression.
	You know, you're going to see unemployment rates first at 25% and then 50% and then even higher in certain sectors.
	You're going to see now labor will be safer for many years to come because robotics is difficult.
	But cognition, which is what I'm focused on here, and tokens in the workplace for desk jobs, white-collar jobs, et cetera, you're going to see unemployment rates in some areas like that of 70 or 80%.
	For example, customer service.
	You will see 80%.
	This is my estimate.
	80% of current customer service jobs that are conducted by humans will be replaced by AI.
	And probably that number will be even higher, but I'm being conservative, just to say only 80%.
	So four out of five of those people are going to be looking for jobs.
	Where are they going to go?
	Well, you can't drive an Uber because all the taxis are going to be automated in the coming years, or you can't drive an Uber for long, maybe for a year or two, and that's about it.
	So what are you going to do?
	Well, society is going to have to address this question because it's a massive, a massive issue.
	Widespread unemployment, massive defaults on consumer debt, which includes home loans, car loans, medical debt, and student loans, etc.
	This is going to collapse financial institutions.
	This is going to cause widespread homelessness.
	And this is going to drastically reduce consumer purchasing behavior, which will dramatically impact online e-commerce giants like Amazon and also retail outlets like, I don't know, Home Depot and Best Buy and whatever.
	This is going to upend everything in our economy.
	And almost nobody is ready for this.
	And very few people realize it's even happening.
	Everybody's in denial.
	And that, you know, denial is just not going to work.
	So people had better get a grip on this.
	Anyway, you can use our AI engine.
	It's completely free of charge.
	And it's available at brightu.ai.
	That's the word bright, followed by the letter u.ai, brightu.ai, and it can help you with lots and lots of things.
	So take advantage of that.
	It's free to use.
	And then also check out all the latest censored news.
	And we do news analysis trends, emerging trends, all kinds of things at censored.news.
	So check that out.
	And thank you for listening.
	I'm Mike Adams here.
	God bless you.
	Praying for all of you to have success in this new economy.
	Adapt, learn, train.
	It's going to get very difficult for a lot of people.
	So be prepared for that.
	Thanks for listening.
	Take care.
	Craving something sweet, but want to avoid processed sugars and candy?
	Our delicious and nutritious soft dried apples and mangoes are non-GMO, certified kosher and organic.

▲