You sabi GitHub, dat web service wey dey use am for collaboration among open-source software developers?
Recent years, dem don dey use am well well for collaborative work, no be only for open-source software, but even for company software development and other things wey no even relate to software.
Me too, I dey use GitHub to manage my own programs and the drafts of articles wey I dey write for this blog.
Inside dis article, I go check how GitHub go fit dey useful pass software development for future, and become a place to share knowledge openly.
How DeepWiki Dey Generate Wiki Site
Plenty software development tools wey dey use generative AI, dem design am to help human beings with dia programming work. Na human beings dey write the programs, then AI go just support.
But then, one new kind of software development tool dey come out now, wey human beings go just give instructions, and generative AI go take over the work of creating programs.
Devin na one of those tools wey first come out and people come notice am well well. Some people even say say if you bring Devin join your development team, e be like say you don add one more programmer. Even though dem still dey talk say human engineers still need to give am proper support to make am work well, dem go surely gather all those data and use am for improvement.
The time wey software development teams go get just one human and AI programmers like Devin as team members go common, e no too far again.
Cognition, the company wey develop Devin, don also release one service dem call DeepWiki.
DeepWiki na service wey dey automatically generate wiki site for each software development project for GitHub. This one mean say one AI, similar to Devin, go read and analyze all the programs and related documents of that project and create all the manuals and design documents.
Dem talk say Cognition don use DeepWiki create wiki sites for over 50,000 big public software development projects for GitHub, wey anybody fit access for free.
Since these ones na public projects, e no get any problem at all to do am like that. Even though wiki sites fit generate automatically, e must don take plenty generative AIs to run for a long time with full power, and the money wey dem spend must be plenty.
By paying all these costs, Cognition don give big benefit to plenty public projects, wey make dem get explanations and design documents for free.
If statistical data show say these wiki sites dey useful for each public project and e get big effect on improving quality and productivity, then software development companies go adopt DeepWiki for dia own projects.
Cognition must don invest for generating wiki sites for plenty public projects, because dem believe say this one fit happen. This one show Cognition's confidence for DeepWiki. And when dem adopt DeepWiki, Devin go automatically follow, wey go make the possibility of AI programmers becoming common to increase well well.
GitHub as Platform for Sharing Documents
GitHub don turn to one popular web service, wey almost everybody dey use for sharing, co-editing, and saving programs for open-source software development.
Recently, dem don make im management and security features for companies better, so e don become common tool for big companies wey dey develop software.
Because of this, GitHub dey always make you think say na only for saving and sharing programs. But for real, you fit use am share, co-edit, and save different different documents and materials, wey no even get anything to do with programs.
So, plenty people dey use GitHub to manage documents wey dem want make many people co-edit. These documents fit be about software or even wetin no relate to software at all.
Wetin pass dat one, blogs and websites too na documents wey get one kind program inside dem or wey programs arrange and publish dem.
Because of this, e no dey strange for individuals and companies to save the content of blogs and websites, plus the programs wey make dem easy to view and the programs for automatic site generation, all of dem together as one single project for GitHub.
E also possible to make those kind blogs and websites public projects on GitHub so dem fit co-edit dia content.
Furthermore, recently, generative AI no just dey use am for software development alone, dem also dey always put am inside software.
For this case, instruction sentences wey dem dey call prompts, wey dey give detailed instructions to the generative AI, dem dey embed dem inside the programs.
These prompts too fit be considered as one kind of document.
Intellectual Factory
Even though I be software development engineer, I still dey write articles for my blog.
As much as I want plenty people to read dem, e hard well well to make the number of readers to increase.
Of course, person fit think of creating articles just to get attention or actively contact influential people for advice, among other efforts and smart ways.
But, considering how I be and the effort and stress wey dey involved, I no too keen for aggressive promotion. Wetin pass dat one, if I spend time on top those kind activities, e go reduce the time for my main work, which na programming, thinking about ideas, and documenting dem.
So, recently, I decide to try one strategy wey dem call multimedia or omnichannel, wey mean say I go expand how far my blog posts fit reach by developing dem into different different forms of content.
Specifically, this one include translating Japanese articles to English and posting dem for English blog sites, and creating presentation videos to explain articles and publishing dem on YouTube.
Wetin pass dat one, apart from publishing on general blog services, I dey also consider creating my own blog site wey go list and categorize my past blog posts and link related articles.
If I dey spend time creating these ones every time new article come out, e no go make sense. So, all tasks apart from writing the original Japanese article go dey automated using generative AI. I dey call this one an intellectual factory.
I need to develop programs to make this system work.
Currently, I don already create programs wey fit fully automate translation, presentation video generation, and uploading to YouTube.
I dey now for the process of creating basic programs for categorizing and linking existing blog posts.
Once that one complete, and I create a program to generate my own blog site and automatically put am on a web server, the initial idea of my intellectual factory go complete.
Intellectual Factory for Everybody
My blog post drafts, wey be like raw material for this intellectual factory, I dey manage dem as a GitHub project too. For now, dem dey private, nobody fit see dem, but I dey think of making dem public projects with the intellectual factory programs for future.
And the way I dey categorize blog posts, link articles, and explain video-transformed blog posts, wey I dey develop now, get the same idea with DeepWiki.
When you use generative AI, e fit produce different different contents from original creative works as raw materials. Plus, e fit connect information and knowledge inside dem, and dat one go effectively create a knowledge base.
The only difference na whether the raw material na program or blog post. And for DeepWiki and my intellectual factory wey generative AI dey power, dat difference no really mean anything.
In other words, if we sabi "intellectual factory" for a general, wide sense, not just for my program, then DeepWiki na also one kind of intellectual factory.
And wetin intellectual factories dey produce no just stop at translated articles for other languages, presentation videos, self-made blog sites, or wiki sites.
Dem go fit convert content into all kinds of media and formats wey you fit imagine, like short videos, tweets, comics, animation, podcasts, and e-books.
Wetin pass dat one, the content inside these media and formats fit also change to fit the person wey go receive am, like wider multi-language support, versions for experts or beginners, and versions for adults or children.
More than dat, even creating customized content on-demand fit happen.
GitHub as "Afa Intellectual"
The raw materials for an intellectual factory fit dey anywhere, basically.
But, if you consider say GitHub don become the main standard for sharing, co-editing, and saving open-source project programs, and say plenty people, no be only me, dey use GitHub to save dia documents, e come clear say GitHub fit be the main place to get raw materials for intellectual factories.
In other words, GitHub go turn to one shared intellectual mine for everybody, wey go dey supply raw materials to intellectual factories.
The term "shared by humanity" here dey remind us of the idea say open-source projects na like shared software property for everybody.
The open-source philosophy wey don support GitHub go also fit well with the idea of open documents.
Wetin pass dat one, one culture of managing copyright information and licenses for each document, just like programs, fit come out. Content wey dey automatically generated from source documents fit easily get the same license, or follow rules wey the license put.
From the angle of developing an intellectual factory, say all the raw material documents dey centralized on GitHub na the best thing.
This one get two benefits: e make development more efficient by just connecting GitHub with the intellectual factory, and e give the ability to effectively show the functions and performance of your own intellectual factory using publicly available documents, just like DeepWiki.
For future, as dem go dey develop different different intellectual factories and dem go fit connect to GitHub, and as more people and companies go dey manage documents on GitHub and process dem with intellectual factories, GitHub's position as an intellectual mine suppose become strong well well.
Everybody's Shared Public Knowledge Base
With GitHub for the center, like one intellectual mine, and different different contents and knowledge bases wey intellectual factories go produce, this whole system go create one public knowledge base wey everybody go share.
Wetin pass dat one, na one active and live knowledge base wey go automatically dey expand as the number of documents wey dey published on GitHub dey increase.
Even though this big and complicated knowledge base, wey get plenty plenty knowledge, go dey useful to human beings, e go hard to completely bring out all im potential value.
However, AI go fit use this public knowledge base, wey everybody share, to the fullest.
Public Knowledge "Veins"
If dis kind system come true, different public information go naturally gather for GitHub.
This one no go just stop for personal blog drafts or company websites.
Academic ideas and data, like papers dem never publish, research ideas, experimental data, and survey results, go also pile up.
This one go attract not only those wey want use knowledge, ideas, and data to help everybody, but also those wey want quickly spread dia discoveries and get recognition.
Even for scholars and researchers, many go see value for make AI confirm how valid, new, and impactful dia work be, express am through different contents, and make am go viral, without waiting for the long peer-review process.
Or, if dia work catch the eye of other researchers or companies like this, wey lead to collaborative research or funding, e get practical benefits too.
Plus, e go likely be say AI's own knowledge go dey returned.
Generative AI dey get plenty plenty knowledge through pre-training, but e no dey actively search for unexpected connections or similar structures between that plenty knowledge while e dey learn.
The same thing dey go for new insights wey come out from connecting different pieces of knowledge.
On the other hand, when you dey explain such similarities and connections during conversations with a pre-trained generative AI, e fit accurately check dia value.
So, by randomly or completely comparing and connecting different pieces of knowledge and putting dem into a generative AI, e possible to discover unexpected similarities and valuable connections.
Of course, since e get plenty plenty combinations, e no make sense to cover all of dem. However, by making this process simpler and automated for the right way, e go possible to automatically discover useful knowledge from existing knowledge.
By achieving this kind automatic knowledge discovery and saving the discovered knowledge on GitHub, e look like say this loop fit repeat forever.
This way, plenty undiscovered "veins" of knowledge dey inside this intellectual mine, and e go possible to dig dem out.
Conclusion
As GitHub, wey be like one main standard, shared human knowledge base, don dey established, dem go likely use am for pre-training generative AI and for finding knowledge like RAG.
For that kind situation, GitHub itself go work like one very big brain. And generative AI go share this brain, dey spread and expand knowledge while dem dey share am.
The knowledge wey dem go record extra there no go just be records of facts, new data, or classifications. E fit also include "catalytic knowledge" wey dey help discover other knowledge or new combinations.
I dey call that kind knowledge with a "catalytic effect" "intellectual crystals" or "knowledge crystals." This one include, for example, new ways to think, like "frameworks."
When dem discover or develop a new framework and dem add an intellectual crystal, im "catalytic effect" go make different combinations and structuring of knowledge possible pass before, and this go lead to the growth of new knowledge.
Among these, other knowledge crystals fit dey. This one, for turn, go further increase knowledge.
This kind knowledge no be scientific discovery but something closer to mathematical exploration, engineering development, or invention. So, na knowledge wey dey grow purely through thinking, rather than through new things dem observe like scientific knowledge.
And GitHub as an intellectual mine, plus plenty plenty generative AIs wey dey use am, go make the growth of such knowledge fast fast.
Knowledge wey dem dey discover one after another, faster pass how human beings fit discover am, knowledge factories go provide am in a way wey we go easily understand.
This way, knowledge wey person fit explore purely through thought go dey dug out sharp sharp.