Deep Breath Before Keyword Histogram Clustering
As I explore the potential of machine learning to optimize my blog content, I am considering changes to my domain URLs to appeal to search and AI systems. Additionally, I plan to create an MLGL license and a heartbeat tech system, and use Linux commands to keep my laptop and server instances in sync. By leveraging my laptop computing power to analyze server data, I am confident I can create a more effective blog post.
Exploring Machine Learning to Optimize Blog Content
By Michael Levin
Sunday, January 22, 2023
How does one organize one’s own content in a blog like this. Maybe I headlined, and thus titled, and thus keyworded and URL’d things correctly to help Search and AI systems notice and pick-up and digest and learn from, and maybe I didn’t. Why not let some machine learning decide? I have either this file I’m editing right now or the URLs on which they’re published, via the Gihub Pages (previously and often hosted at github.io) and the Jekyll static site generator system that takes a bunch of markdown files I just edit with vim and keep in private repos I push to Github.com and over time a website rife with fairly well optimized come into existence.
But are the afore-crawled pages really optimized? What if you suddenly didn’t care about how old URLs are or how often they change. You’re always trying to appeal to, keyword index-wise, some six-degrees of separation real-time crawl system under the guidance of an AI-navigator.
If you had no fear, what would you do? I would change the URLs of my long-running domain to be perfect reflections of their content, rather than some static-state fear of changing URLs because the ocean always gets saltier, and all that. Because AI, the ocean doesn’t always get saltier. Real-time Maxwell’s Demons are desalinating web-results or bot chats.
It is in this environment, I’ve got to get serious about that Machine Learning General License. Do I put a P in there to match GPL? MLGL or MLGPL? MLGPL is too much of a mouthful and I don’t need to associate it that closely with GPL. The meaning of the license will hold true with the shorter usage.
Okay, I really want to focus on these tuples for life for a moment. I want to make a fun heartbeat tech system.
Okay, let’s get started with Linux command like like you’re going to make something run regularly on your laptop system. It’s just now since your laptop can be made to look like it’s a running Linux server, you can keep something running in the background in the language of your choice, refine the app there, and when you sync that repo with a cloud-server or an LXD instance running on your home NAS, you have your own scheduler or task queue running.
It all becomes very easy to synchronize because the WSL instance on your desktop is almost exactly the same as the server instance. You’ll keep your Linux instances of Python upgraded and in sync the same way and everything. You can write thing so that their path to their data is the same, desktop or server instance, so you can always run analysis of server data using your laptop computing power. I resist setting up dedicated databases for this. It still may be achievable with SQLite.