BrightLearn UPDATE for Dec 27, 2025 - BRIGHTLEARN MIND MAPS!
|
Time
Text
Well, welcome to this BrightLearn update.
I'm Mike Adams, the developer of the BrightLearn engine.
And just got a quick technical update for you here, giving you some insight of what we're doing behind the scenes.
We ran into an issue that the popularity of the site has really exploded very rapidly.
We're rapidly approaching 10,000 books being published on the site, which is great.
We love it.
Over 3,000 authors, over 107,000 downloads of books.
I mean, it's a very popular engine.
People are loving it.
People are sharing the books.
It's a format that really makes sense to a lot of people.
And people are totally impressed by not only the quality of the cover art, but also the quality of the writing and the research.
And I got to tell you, the research is only going to get better because I've had some other major breakthroughs on the data pipeline of, well, books and science papers both that are going to go into the index soon.
I've been able to increase the processing speed by about 40x for those books.
And very interesting reasons why, but it all has to do with AI breakthroughs and more efficient systems and lower cost of compute, things like that.
But anyway, what you need to know from a practical standpoint here is that for the next several nights, around midnight or after midnight Central Time USA, we're going to be doing some maintenance, which is we've had to expand to a larger, more robust external database.
And the reason is the database service that we were using was fine at small scale, but it could not handle the load.
And the load has been pretty intense with hundreds of database connections and a lot of data in and out, using a lot of CPUs to drive the database, not just the data itself, but a lot of compute on the SQL statements.
And that's because of all the things that we're doing in terms of regenerating covers and regenerating chapters and queuing books.
And the whole process is very data intensive, as you might imagine.
And as a result, we've blown away this smaller database.
And it has had a couple of hiccups that cause a few interruptions.
So I'm in the process of migrating to a larger enterprise scale external database with features like auto-scaling and better connection pooling capabilities, things like that.
So that migration is pretty complex.
Fortunately, I've got a lot of experience in this area.
And I know the process of doing this without breaking it.
Well, I should say without breaking it too badly.
It will break temporarily from time to time.
But I have to make all the code changes and everything.
Well, I mean, my AI agents are doing the grunt work, but I'm directing it.
And I'm doing this without bringing the service down.
So the service will stay up during this entire migration.
And the service will coexist with the new database that we're putting in place.
It will seamlessly take over with full synchronization.
Although, just as a caveat, I should say, if you are using this engine after midnight central time, there could be some glitches due to some restarts.
So just use with caution at that point.
Make sure you save your prompt so that if the book somehow, if it's in the process and it just gets clobbered, you can just restart it with a prompt.
And I also want you to know that if you're using a token with the BrightLearn engine, that token does not get used until the very last step after the book is created and published and posted online and the email is sent to you.
Only then is the token marked as used.
So if the book somehow fails during any of the creation and processing and anything, if you've used a token for that book, that token is not used up.
That token is still good and you can restart the book with the same token.
So we did that on purpose so that people don't like lose tokens when bad things happen to books in the middle of database restarts and things like that.
So yeah, again, that's all by design.
And wow, as I'm watching the database migration here, it's really something.
Over 4,000 database connections are active right now just for the migration.
And it's cranking the hundreds of megabytes of data into the system right now.
We're doing a full migration.
And it's working.
So anyway, once this is all done and all the code is good, what you'll find is that the system is probably a little more responsive and it will have less downtime.
And it may speed the book creation time just slightly, but it's probably not noticeable as the main wait time on book creation is really waiting for LLMs to do the writing.
And there's only, you know, you can't just make that 10 times faster.
You have to wait for it.
So the other thing that is going to happen in 2026, which is going to be a really exciting year for BrightLearn, is that the, you know how currently there are 10,000 books and I don't know how many millions of other pages of other documents that you can use as reference material when you're choosing to create the book, you know, you can choose, I want to use the following as references.
It could be articles, it could be podcasts, could be interviews, or it could be books.
And the books are currently 10,000 books.
That number is about to take a huge, huge leap.
From 10,000 to, well, it just depends on how much time we want to go forward.
But by the end of this next year, it'll be 100,000 books easily.
It might be a couple hundred thousand books.
And the reason for that is I've achieved some really strong breakthroughs in book processing for the data pipeline.
And you might be able to hear some of my computer fans here.
Actually, it's funny because I'm running one, just one of the pilot dashboards on my desktop computer to keep an eye on it.
But even it is doing a ton of work.
So anyway, I've been able to do really vast multi-threading of the normalization workers.
So a lot of steps to take book files and actually prepare them for reference material.
A lot of steps.
And it's not as easy as you think.
Especially when you don't even know what book it is at first because the file's not named correctly.
And you don't even know if it's fiction or non-fiction, etc., right?
And you don't know, was it scanned via OCR?
If so, it's full of OCR errors, things like that.
Or it may be a PDF that's just a bunch of OCR scans with no text, just images, and then you have to carry out OCR on it.
And then you have to fix all the OCR errors, et cetera, et cetera.
It goes on.
This process is something that I've really refined.
Through most of 2025, the data pipeline processes have become really fine-tuned.
Really well done, because since the summer of 25, I've been writing the code using mostly Opus 4.5 or whatever version was available before 4.5, but it's Opus, and Opus has been writing all the code both for Linux systems and Windows systems, or a combination some Windows and some Linux systems.
Mostly Linux for the hardcore processing, because Linux systems are much more reliable and they don't they don't auto update and the GPUs work better in Linux too, by the way.
So if you want to do what I'm doing, you should get Linux and don't even mess with Windows.
But anyway, we still have some legacy Windows systems that are running, and so whenever I write code for data pipeline tasks, I have to write it in both Windows Python and Linux Python, which is not the same.
There are differences.
So anyway, the bottom line is, you are about to see, over the months ahead, you're about to see tens of thousands of new books coming online as research material that can be cited by your books.
So when you create a new book, instead of instead of it just searching through 10,000 other books, it might be searching through a hundred thousand other books, and that's, um, that's a pretty big database function too, all by itself, and just the fact that we'll be adding hundreds of thousands of books over time to the system for the research engine.
Uh, that that alone is is a pretty big deal.
But then on top of that, we are preparing millions of science papers.
So even though more books will come online first, we're also processing science papers.
And the last count I took of the science papers that we have partially processed is 10 million.
10 million science papers across all the mainstream journals.
And we are currently classifying those papers.
And shortly we'll be putting those through normalization and another layer of classification.
And then we'll be adding those to the engine.
very soon within a few weeks, I expect and that means then you'll be able to start selecting a checkbox for science papers and that way, when you publish a book or you submit a book, the research engine will cite.
You know it'll go through millions of science papers in addition to tens of thousands of books, but then just that alone means that the database that we use for those tasks has to, of course, be upgraded, not only to hold all of that material, because now we're starting to talk about Terabytes of storage, okay.
You know, terabytes of storage.
We're not, this isn't small-scale stuff, right?
We're talking enterprise-level databases and enterprise-level storage.
And then, when your storage gets bigger, your compute has to get bigger to sort through it all.
So, it's kind of a domino effect.
When you start shoving millions of new documents into the system for indexing, then you have to upsize everything around it.
The whole infrastructure has to be upsized.
And even though you, as the user, you don't see any of this behind the scenes, I do.
And I, I mean, I'm, you know, I have to monitor a lot of things to make sure it's all working well.
And so far it is working well, but as we scale it, I want to make sure it continues to be fast.
And this, by the way, this is also the secret sauce of what makes the books so great.
Because anybody can go out and ask some AI engine, you know, Chat GPT, write me a short story or whatever.
Yeah, sure, it can do that, and it will suck.
And it won't be based on anything except the engine's internal knowledge, which is probably very badly biased if it's Chat GPT.
Whereas what our engine does is it does the research and then cites the documents that are related to each subchapter.
That's why our books are the best.
That, you know, of any book creation engine that you could possibly think of, our books are the best books with the best references, etc.
And this is the reason why, and it's only going to get better.
It's because of all the citations, you know, all the documents that we have indexed.
So that is continuing.
And by the way, that alone, I don't know if you've been involved in any of the using LLMs and purchasing hardware or using APIs or any combination of that.
But it's not inconsequential in terms of the cost to process 10 million science papers.
It's probably tens of thousands of dollars of ultimate cost in order to do that.
And the same thing with books.
So this isn't, it's not free to do this, but it's something that I feel is really critical to the success of the engine.
And the other thing is that all this research, this goes into the same shared resource database that's used by our AI engine at BrightU.ai.
So BrightU.ai also does research in order to bring you answers to your questions.
Well, that research uses the same collection of documents.
So every time I put 10,000 more books into this engine, it benefits two applications, the BrightU.ai chatbot engine or research engine, as well as the Brightlearn.ai book creation engine.
And in effect, it benefits some other services that we have internally and some we're going to roll out externally as well.
So it's a single source of curated knowledge that's used throughout our ecosystem to be able to bring you better answers, better books, better writing, etc.
And one more reason why this really matters is because in 2026, not only will there be a lot of auto translation of books into first Spanish and then French and then Chinese, I think will be next.
But also we're going to be doing a lot more audio books, full-length audio books in some cases.
And then as the technology becomes available and affordable, then mini documentaries.
So that might be summer of 2026.
Who knows?
It might be late 2026.
Whatever it takes, that's on my list, is to be able to create three or four minute mini documentaries that are great video with narration, with full presentations, the whole thing.
based on a book.
And those would be generated when your book hits a certain number of reads.
Just like currently when your book hits, well, you may not know this, when your book hits 300 reads, there's a mind map image that's generated for your book.
And that image, which is awesome, appears on the book homepage now.
And you can see it.
You can see examples of that by going to the site, books.brightlearn.ai.
Click on the most popular button there on the homepage and click on any of those popular books.
And you'll see the mind map on the book homepage.
Well, those get generated automatically when you hit 300 reads.
And then when you hit 500 reads, then an audio podcast is generated for your book of about 15 minutes duration.
And currently it's set that when you hit 1,000 reads, your book gets auto-translated into Spanish, although I haven't yet turned on that feature.
There's some debugging and things like that to work on there.
And then when it hits another 1,000 views or reads, it'll be translated into French, etc.
And then perhaps when it hits, I don't know, 5,000 reads or whatever number we decide on, then a mini documentary video will be created for it.
And that will be viewable from the book homepage as well.
And ultimately where this is going, which is probably a couple of years away, just depends on AI breakthroughs, I suppose, is full-length documentary films based on your book.
So yeah, imagine taking your book, which was built based on your prompt, and then taking the whole book and turning it into a movie.
Yeah, a documentary.
And then having that documentary completely rendered from AI with everything from, you know, actors and dialogue and all the scenes and everything just completely created by essentially AI agents that are filmmakers and directors and producers, etc.
That's going to be a really interesting project.
And, you know, ultimately, Hollywood is obsolete.
I'm sure they see that coming at this point.
But, you know, people aren't, you're not going to have to go to a studio to make a movie.
All you need is competent AI engines and some breakthroughs in technology and some cost efficiencies and things like that.
But anyway, we'll be among the first to really explore that from the point of view of creating full-length films that teach people things.
So, you know, my goal is to is to teach people.
The Brightlearn.ai book website, it's about spreading knowledge and bypassing censorship and empowering people all over the world in different countries and different languages in different modalities, you know, text, audio, video, or podcast, audio books, mini documentaries, full-length documentaries, mind map images, etc.
So every different modality that human beings can understand, that's what we want to use, all of them, to help bring information to people.
And as a result, consumers will be able to get this information and learn these things in whatever way they want, in whatever format they want.
But you won't need to go to Hollywood.
I mean, you won't need to buy an audio book per se, at least not for instructional skills.
Now, there'll still be a market for certain audio books.
And there will still be a market for certain books published by people like, you know, autobiographies and things like that.
Or special analysis or celebrity books.
Or, I mean, there's all kinds of things, you know, photography books, right?
Lots of things where there's still going to be a market for regular books.
But nobody's going to need to buy a book to teach them something like, you know, how do I build farm tools?
Or how do I preserve food?
You know, because you'll be able to generate a book that tells you all that information expertly and at zero cost.
And that's what BrightLearn does.
And that's the whole point of it is it's not to, you know, not to replace the book industry, but rather to make the cost of learning zero so that everybody in every language, in every country, all over the world can learn everything they need without having to spend money.
Because previously, the cost of an education, you know, college degree was very expensive.
And even myself, you may not know this about me, but I was accepted into MIT coming out of high school or actually after I took the college entrance exams in my junior year in high school.
I was accepted into MIT.
And actually, I had acceptance and invitation letters from all over the country.
I mean, you know, California schools plus MIT, Georgia Tech, I mean, all kinds of schools, but I couldn't afford any of them.
So I know what it's like to, I mean, so I just went, I went to a university that I could afford because it was in my own state and I had scholarships there, etc.
But I couldn't afford to go to MIT, which at that time I was even told by an MIT advisor that met with me, I was told that that would cost about $50,000 a year and, you know, for room and board plus the tuition books, the whole thing, 50 grand a year.
And that's in like 1988 or whatever the year was.
50 grand was a lot of money then, you know.
So, no, I couldn't afford to get an MIT education.
Today, because of the technology that I've developed with the help of AI, we can give everybody any education they need on any subject that they want at zero cost, whether it's chemistry or physics or mathematics or philosophy or history or engineering or business or whatever.
Or even if you just want to learn how to use AI, anything you want to learn, you can now learn for free.
So this is a game changer for our world.
For the first time, it means there's zero barriers for people to learn anything they want to learn at zero cost.
Well, I guess the only barrier is they have to be online.
They have to have access to the internet.
But that's the vast majority of the human population right now.
And those numbers are also increasing.
So if you have access to the internet, you can now learn for free.
You can get a college education for free, if not better than a college education, given the decline of Western universities and how little they teach people these days.
It's like woke ism 101.
You know, that's what people graduate with.
Like, that's kind of pointless, kind of useless in the real world.
If you want real-world skills, then you can get them all for free now.
And so that's why I appreciate the fact that many homeschoolers are using our engine.
And I'm aware of some homeschooling organizations that are starting to advocate this engine for their teachers and students.
So think about it.
If you're a student, you can generate your own book.
Or if you're a teacher, you can generate a book for your students.
And then you can legally give all your students that book completely free because of the Creative Commons attribution licensing that we have.
So you can use these books commercially or non-commercially as long as you give credit to BrightLearn.ai.
So that empowers teachers.
It empowers students.
It empowers moms and dads who are doing homeschooling.
You can generate the coursework for an entire semester.
You know, I mean, the possibilities are limitless.
So anyway, there you go.
A little bit of an update behind the scenes and some of the vision for 2026.
I appreciate your support.
And if you want to help support us in this project, it is currently fully funded by generous donations from HealthRangerStore.com, which is an online store of clean foods, health foods, superfoods, and supplements.
Oh, and clean lab-tested personal care products.
It's a store that I founded.
So that's why it's called HealthRangerStore.com.
And we conduct lab testing and expert formulations completely free of synthetic fragrances or artificial colors.
We don't use GMOs.
We don't use aluminum in the deodorants.
You know, everything is ultra-clean.
So if you want to clean up your health and your products, shop at healthrangerstore.com.
And your purchases there not only will give you loyalty points that you can swap for book credits that you can use at BrightLearn.ai, but also whatever profit we manage to earn helps us donate to brightlearn.ai so that we can keep this service free for you and everybody else who uses it.
So thank you for supporting us by shopping at healthrangerstore.com and thank you for using brightlearn.ai and spread the word and keep on using it.
And together we can inform and empower the whole world with free but valuable knowledge.
That's my vision.
So thank you for listening.
I'm Mike Adams, the founder of BrightLearn.ai.
Take care.
Organic heavy cream powder provides the same creamy texture as regular heavy cream without the limited shelf life.
HealthRanger's select organic heavy cream powder is convenient, portable, and shelf-stable.