Is The Rise of AI The Fall of SEO? (systemd snapshot machine)
As a passionate FOSS SEO professional, I remember the early days of search engine forums and giving advice on how to optimize websites. With the rise of AI, I recognize that my skills are becoming less applicable and understand the anxiety this may cause. I discussed my idea of collecting data to gain competitive advantages with Jo, and am now working on a systemd snapshot machine.
Gaining Competitive Advantages with AI: My Journey with Jo and Systemd Snapshot Machine
By Michael Levin
Friday, March 10, 2023
This is a topic I need to keep very much on top of.
I remember very plainly the rise of SEO and search engine forums which are no longer there. One we called SEF and it predated almost everything but Danny. And I remember my advice to others in 1997, 98, 99 as Google was just starting to be really important.
That advice was “view-source”on search results. Look at the HTML code. Do it often and try to learn to read and understand what you’re looking at and how it translates into a graphical web page. Maybe even save it and look at how it renders in a browse separate from the Internet.
Piece the output part of the search process together. Why this output from your input. Search and view-source again. See how it changes. Much is subjective in the field of SEO, but not this.
Those bytes are objectively there as a response to a query you made on such-and-such a date from such-and-such a location from such-and-such a device. This is good data. It’s a snapshot of things were in that moment in time, space and platform. And those bytes can be saved and used for later purposes.
So there I was teaching people how to SERP-track, touching even on the issues of selecting good multi-column primary keys for your database, just rolling onto the 2000s. I recognized and used many other data types useful to an SEO such as performing website crawls or hitting APIs, always spinning my own “in-house” non-vendor dependent secret weapons for years.
When I was in-house, this was highly useful. I created web-tracking systems and flexible Intranets with a precursor to Ruby-on-Rails that I wrote on VBScript and never released. But if served me well and worked as the basis for HitTail, a keyword suggestion tool i created in 2006 that had a 15-year run. But as I went Agency-side, these skills became less applicable. Based on Active Server Pages (the old .asp files) as my skills were, I found myself suddenly obsolete tool-wise and more importantly, muscle-memory wise.
Skills build on skills build on skills. This is
Take note of parts you don’t understand.
Trading effort for economic product up to this day had been in an area that predictive deep learning models will clearly intermediate: They might feel anxious, insecure, threatened or obsolete. They might worry about losing their job or income to a more efficient and accurate machine. They might feel frustrated, angry or resentful towards the technology that is disrupting their industry. They might also feel curious, hopeful or motivated to learn new skills or adapt to the changing environment.
- Predictive deep learning models are a family of learning algorithms that can be used to learn complex prediction models, such as multi-layer neural networks with many hidden units⁴. They can be used to accurately and reliably calculate various properties or outcomes for different domains, such as environmental properties², customer behavior³, image analysis and speech recognition⁴.
Source: Conversation with Bing, 3/8/2023
(1) An Introductory Review of Deep Learning for Prediction Models With Big Data. https://www.frontiersin.org/articles/10.3389/frai.2020.00004/full Accessed 3/8/2023.
(2) Predictive deep learning models for environmental properties: the .... https://pubs.rsc.org/en/content/articlelanding/2019/gc/c9gc01968e Accessed 3/8/2023.
(3) What Is Predictive Behavior Modeling? | Built In. https://builtin.com/machine-learning/predictive-behavior-modeling Accessed 3/8/2023.
(4) Predictive modeling, analytics and machine learning | SAS UK. https://www.sas.com/en_gb/insights/articles/analytics/a-guide-to-predictive-analytics-and-machine-learning.html Accessed 3/8/2023.
(5) DeepLMS: a deep learning predictive model for supporting online .... https://www.nature.com/articles/s41598-020-76740-9 Accessed 3/8/2023.
Yeah, it’s about stepping back to that objective level. Of course all our experiences are subjective, and the field of SEO is very subjective indeed. However, the best we can do to remain objective for the competitive advantages it is likely to confer is to observe, record, take pictures. Collect data. Do screen snapshots. Maybe apply visual learning to the .png files (or whatever) later. But collecting the data is easy. Very mechanical. Very ideal first project for FOSS SEO in many ways that benefits both MOZ as my employer and my passion for FOSS as much as possible. There are some ideal best of both world projects here, and I talked with Jo who likes where I’m going (I think), and so the systemd snapshot machine is coming.