2025 Important Updated Perspectives on AI
Notes From the Desk: No. 36 - 2025.02.13
Notes From the Desk are periodic informal posts that summarize recent topics of interest or other brief notable commentary.
2025 Updated Perspectives on AI
Two years ago, I published my first essay, AI and the end to all things, It was shortly after the release of ChatGPT. I had not put any thought into AI prior. I sat down and wrote down all of my thoughts on the subject at the time. Importantly, I did not read anything about AI. This was purely thought exploration on my own. I wanted to work out meaning in my own head without influence.
Now, 2 years later, much of it has still held up in principle. Which is astonishing, as it is rather difficult to write something about AI that has persistent relevance. If you have never read it, I think you will find it of great value. I would love to hear your thoughts if you have opportunity to read the essay. This was the introduction I wrote in 2023.
What Is The State of AI Today?
After watching the evolution of AI over the past two years, some of my perspectives have evolved, and others have remained the same and unchallenged. I currently see the likelihood of the current AI labs and LLM architectures becoming AGI as not possible. However, most importantly, this does not negate many of the concerns around AI.
So, what does the AI landscape look like today? It is still comprised of many contradictions, as much of the discourse intermixes future AI that might never exist, AI hype, and, finally, in the minority, AI reality. Arguments are often misaligned, with each side defending a totally different context of AI.
The combined set of perspectives below you might find surprising, as they are unlikely to be found together from any typical discourse found across the internet. This is likely due, in my opinion, to most arguments singularly guided by either utopian optimism or world-ending fears.
The Current 2025 AI Landscape
Alignment is impossible. Not difficult, impossible. This is the position I started with and where I still confidently remain.
None of the AI Labs are building AGI. They don't know how. It is hype.
My position continues to move toward doubt of all AI Lab proclamations.Billions of dollars are flowing into AI. Maybe someone will figure it out, but current paths are not it.
Who would have thought randomly correct answers would become a multi-billion dollar industry?
It might be impossible to build AGI. Potentially not possible in silicon. I have more thoughts on this to be written.
X-risk from AGI/ASI is absolute. If we can build it. My position here is unchanged except with greater skepticism of our ability to create AGI/ASI.
We should continue to articulate why we shouldn’t build it in case someone figures it out.
Even if mythical alignment were achieved, a paradoxical despair of utopia would ensue.
Note, AI Labs continue to redefine AGI. It will be redefined as whatever keeps the money flowing.
p(doom) is an absolutely useless metric to determine risk. Ignore it.
The current true risk of AI is The Cartesian Crisis, Dead Internet, war, manipulated social behaviors, and surveillance state dystopias.
Same concerns I started with, but are more apparent now.We spend too much time on X-risk and not on these issues happening now.
AI is the dream capability for authoritarians and utopian dreamers. e/acc is the acceleration of 1984 and Brave New World.
AI will result in significant social and human behavioral disruptions. Social media 2.0.
AI has disproportionate capability for nefarious uses than legitimate uses.
It is easy to fill the world with AI spam, scams and deceptions. However, it is difficult to detect and safeguard against these uses.
AI art, music, and creativity: the non-artists are having the time of their lives. The artists are in complete dismay, and the rest of us are losing our ability to know what's real. Watch this sector closely; it is the canary in the coal mine for what may come.
It is pointless to be principally opposed to all AI. It will become the foundation of nearly all of our technology; it is the new code, the new computer, and is equally as unavoidable. We must focus on the nefarious uses.
There are productive uses for AI, but they don't justify the enormous investments pouring into AI Labs. If they fail to achieve the wish-granting machine, likely there is going to be an epic bubble collapse at some point.
I don't live in fear of AI despite the concerns. We witnessed how fear itself is so destructive during the period of 2020/21. People lose their sanity. We won't find the best course of actions in a state of fear. It doesn't help.
The intent of my message is mostly "informed consent of the future".
What’s Coming in 2025 for Mind Prison?
I have spent the last two years posting concerns and warnings about AI, and they still remain mostly the same. AI has now established a foothold in the base layer of productive tools, and it exists in some part in everything we use. All apps, services, companies, and products are integrating AI.
Therefore, there will be more focus on separating hype from reality. How do we navigate this space that already surrounds us? What is AI actually useful for and how do we remain aware of the deceptions and problems?
Just as we are challenged to keep our sanity using social media, we will have increased challenges with AI. But we can’t discard it no more than we can social media. If you want to communicate with the world or sell a product, social media is an essential platform. AI is here, we will complain about all its negatives just as we do for social media, but we must use it to some extent, just the same, to function within the world.
Some topics coming soon include:
Practical information and guides around AI
What works and what doesn’t work
AI Primer and the hidden ideological landscape behind AI that few talk about
New analysis into AGI, reasoning, and human cognition.
Hints as to why AGI might be theoretically unachievable.
Philosophical reasoning guides
Templates for thought processes and methodologies used by disciplined critical thinking masters
How to avoid blind spots
Please let me know which of these topics you are most interested in or what information you are in search of that you haven’t been able to find elsewhere? Suggestions welcome!
No compass through the dark exists without hope of reaching the other side and the belief that it matters …
(Your usual unbiased) practical information -- i.e. what works, what doesn't -- would be very helpful. Thank you.