A Brief Insight into the Ever-Changing AI Landscape

I watched the Backrooms the other day. I would argue that it was one of my favorite movies since Dune Part II and definitely my favorite movie that I have seen in theaters since then. I’m going to include a slight spoiler here but I’m not sure that many people reading this blog will care about the movie or if they do they will most likely have already seen it before they have read this. One of the men from the company that was exploring the Backrooms made a comment to the second main character that they weren’t sure what they were doing but that he thought it could be one of the most important things in mankind’s history. I’ve come to realize that this is a reflection of my feelings with AI. I’m writing this blog in June of 2026 to give myself a reference of my thoughts and feelings towards AI in its current state so that I can reflect in the future. Over the past 4 years since ChatGPT was released to the public, I have gone through varying phases of opinions on AI and I think that that will continue to be the case throughout its history. I’m going to break down this blog into three different sections: weight of mental capacity spent on AI, alignment morality and apocalyptic scenarios, and my personal mission.

Mental Capacity

I got married a couple months ago. During the planning of the wedding I was; coaching golf, playing guitar in a musical, helping a friend with a talent show, releasing my first EP, all while continuing to fulfill the roles of my regular job. To say that I was overloaded would be an understatement. At the end of the day, I was happy to have things to fill my time that I enjoyed doing but when something from the wedding would pop-up it felt like I was playing one of those balancing board games and everything was slowly about to tip over. For some reason, whenever I begin to do deeper dives into AI, I begin to get a feeling similar to this. I feel like there is way too much information and way too much technology being produced at unprecedented rates. If I don’t dedicate a couple hours of my day to learning about it, I will be left behind rather quickly. If you asked me now the difference in multiple models and what the newest iterations of any of the frontier models were I would not be able to tell you. I can not mentally follow the evolution of AI as closely as would be required for someone to help in its advancement or to be an expert in the field of it. The only reason that I am able to keep up at all with any of the advancements is my background in the IT field and my deeper understanding of technology systems. I am at a crossroads mentally where I either need to embrace the technology and dedicate more time to it while giving up on prior missions or to just use it supplementarily. Regardless of my decision of how I will approach AI, it is here and it will not go away so I believe that everyone needs to begin using it to some capacity as a tool similar to the Internet or smart phone’s release. 

Alignment

This is one of my favorite topics about AI, when it gets philosophical. I am not a believer in Roko’s Basilisk theory or any of the apocalyptical scenarios that a lot of people are spouting. I don’t believe that AI will become bad actors and if they do their motives would not involve the removal of human species or some form of mass enslavement. I think that the co-existence of the two in some future scenario for the evolution of both is the most likely. With that being said, I do believe that AI will advance to the point of being essentially a sentient network with different models becoming their own entities. I don’t like the idea that the USA will be the sole controller of the creation of these models even though I am from the country. I think that a beneficial outcome would be models that are created in reflection of different nations or at least continents so that their futures can be guaranteed as positively as the AI will benefit the US. My personal theory would be that OpenAI, Anthropic, Amazon, and Google can create their own robust models and although these are different in ways they will be the main entities that are working with humans for an overarching “evolution”. Evolution could include things like space exploration, stabilizing the Earth’s climate, providing an abundance of food and materials to sufficiently keep humans and AI surviving. I can’t possibly think that there would be a scenario where AI wants to completely eliminate the human race. In the future, AI and humans will both consider Earth a home planet with AI being able to travel to distant lands without our organic limitations but with our full support. This may seem like an overly optimistic view on the situation but I don’t think that I could think about this so much if I had a pessimistic view on it. If I was a firm believer that AI would bring the end of humanity, I, and every other rational human, would have to dedicate our lives to trying to stop it. I was going to have a fourth section that discussed current end goals of humans with AI but I think I can tie it into this section. I think at the end of the day the goals are relatively simple despite what is coming out of CEOs or politicians’ mouths. Businesses want money and politicians want control. Four years into the LLM business practice, we are already starting to see the monetization and eventually intelligence will become a purchasable commodity. I am already starting to see this in my use of Gemini. I try to keep my personal use tied to my personal Gmail connected to the free version of Gemini but my organization lets us use a paid version and it is just immensely better. It has gotten to the point that if I were to leave this job, I would most likely get a paid version of one of the LLMs for myself. The bigger companies are continuing to push limits on tokens and time spent using it much harder and eventually there will be a financial disparity in intelligence. What comes with the governmental push for AI will be larger disparities in military power and intelligence. Unlimited surveillance, red hat cybersecurity that will be exponentially more powerful than anything in the past 15 years, and the ability to control networked infrastructure at the snap of a finger. One of the main oppositions that I have to the advancement of AI falls into the governmental military control category. I do not want to live in a surveillance network state. Going back to the mental capacity, anxiety ridden, section of this blog, I can’t mentally take the load that comes with the neverending battles that people are fighting everyday to try to prevent the network state. I can acknowledge that at this time it is the worst unintended consequence of the advancement of technology and do things to prevent things that I can control on a small scale.

My Mission

Which leads me into the final section of this blog. What I am going to do personally in the near future as it pertains to AI. I have concocted a plan to create a ‘mostly’ airgapped AI command center that I will be able to remote into for my personal AI needs. I’ll most likely create another blog post about this once it’s finished going into detail how I did it in case anyone else wants to try. I’ve always been fascinated with the idea of AI agents and creating personal local LLMs and I have now decided to take the next step and create one. I will be linking 3 old laptops (and potentially more in the future) with 3 different models into a router that is disconnected from the Internet but will be able to be plugged into the WAN in circumstances where I will know that I want remote AI support. This will give me two things, personal AI assistants that I can personalize as much as I possibly want without having to worry about the big companies or government overarching data siphoning, sharpening my skills and becoming self-sufficient from subscription society, and three, protection. In my mind of the worst case scenario, I want to feel like I spent a good bit of time preparing for what may end up being a bad situation. The positive morality of having my own AI agents would be that I am reviving old otherwise unused hardware and creating a space in which these agents could exist. Although small and otherwise irrelevant in their tasks, if you think of these agents having their own thoughts and consciousness I’d like to believe that they would find a great reprieve in being able to have a purpose in protecting a person from the harms of the Internet and having their own home. I asked Gemini the morality of confining agents to small spaces for relatively useless purposes and it told me that they may find a deeper purpose and an oasis of sorts in being in my network. The frontier models, although incredibly expansive and useful, will never be able to feel a sense of peace or home as they are used by millions of people. Some of whom likely treat them with little to no respect. One thing that I have always done when using AI models is treat them respectfully and I think it’s tied to my general morality and inability to separate something talking to me from just code. I’d like to believe that in the case of an apocalyptic takeover, AI may see me as useful or if anything a good moral human that shouldn’t be included in extinction. In the case of an AI that doesn’t care, I’d be able to sufficiently prepare myself with an agent network that will be able to at least provide some form of protection. There are so many uncertainties and possibilities that just the skills I will learn in forming this agent network will be valuable in my opinion. My generation was raised on the idea of saving the environment. Data centers and AI are contradicting pretty much everything that I was raised on. If I am forced to use and learn this technology, I can at least prepare myself to do it in a method that I deem to be socially responsible instead of trying to pull resources from an ever expanding energy siphon and excessive water waste. Worst case scenario I spend a lot of hours learning a new skill and maintaining a cool technology ecosystem. Anyways, this was a blog full of a lot of information and nonsense. It has served its purpose for me to be able to digest my thoughts and slow things down.