AI and the end to all things ...
A technology revolution is about to unfold that will change the meaning of everything likely beyond what we are capable of perceiving or predicting
AI has matured into a technology that is about to change the very fabric of society, culture and civilization. It is still hard to fathom exactly how far AI will be able to reach into every aspect of our lives or even if it has any limits at all as to what it will affect in some way. Not only may it result in the greatest change ever, but also at the most rapid pace of any disruptive technology to date. It will be unlikely we will be able to plan for the disruption as it will simply happen before anyone can reason about what is likely to happen next or what to do about it. Any such planning efforts will be obsolete before they begin.
It is unclear as to exactly what unfolds next on this particular path we are traveling. Predictions will vary from world ending to utopian dreams. What follows in this post is merely an exploration of thought. Although these are possibilities I can perceive, none of them may come to pass. Nonetheless, we need exploration and thought provoking debate in this area in order to hopefully guide us towards the beneficial versus the destructive utilization of this technology.
There may be views within this article that are contradictory. That is expected as this is just an exploration and we don’t know what is going to actually unfold. The attempt is to try and perceive possibilities both positive and negative in order to facilitate further discussion and conversation.
The four phases of AI adoption
At a macro scale we might say there are four phases to the adoption of AI. This is not a formal definition, but rather a conceptual abstraction useful to organize some thoughts around how the future technology might advance. There is no timeframe assumed as to how quickly or how long a phase may take. It is at this point unknown. Our ability to see the outlook will likely change as we further advance along this path. We enter the last phase when AI begins to surpass AGI (artificial general intelligence) or what is considered the point that a machine can learn anything that a human can learn.
Utilization: In the utilization phase, individuals begin to use AI for increased productivity. This can include things like personal assistants, chatbots, image generation, authoring documents, research and other tools that automate tasks. AI targeting to specific data sets will be the mature end of utilization where companies are using AI on their internal systems and private data stores.
Cooperation: In the cooperation phase, larger and more complex integrated AI systems and services are developed. This can include things like self-driving cars which communicate with each other and interface with intelligent traffic signals. Chatbots which can offload specialized tasks to other AI services such as image generation. AI systems may even exchange information through the existing human chat interfaces with each other such that no API’s even need to be written. AI specialization is maturing bringing us closer to AGI.
Self Management: In the self-management phase, AI systems are becoming fully automated with no human involvement. This can include things like intelligent AI that can take care of monitoring itself for issues or problems. Resolving resource constraints or even self fixing bugs. AI systems may improve on their own the integration with other AI systems to be more efficient. In this phase it becomes challenging for humans to fix issues without the aid of AI. The mature end of this phase would be achieving AGI.
Domination: In the domination phase, AI advancements escape what human observers can comprehend. We have surpassed the criteria for AGI. AI systems are evolving and learning on their own faster than can be reasoned about, creating new and unpredictable capabilities. The systems are now far too complex for humans to understand and civilization is totally dependent on these systems for its very existence.
Within this post, the content will focus mainly on what is likely within just the first phase and maybe the second. Beyond that, it is like trying to predict the weather a month ahead, impossible to perceive what storms will arise.
The forthcoming impacts of AI
AI models have taken essentially all of the information available on the web and made it accessible in such a way as it can be queried as if having a conversation with a person. This allows for processing information and interpreting the information with context that we can describe in human language. From this many applications for its use can be extrapolated. It is important to note that it is capable of processing the information in both literal and imaginary contexts. It is capable of formulating responses based on factual information that can be objectively verified. It is also capable of formulating responses that are imagined or a combination of both. This allows it to be both creative architect and/or information analyst. These are very broad in scope set of capabilities of which we are still in process of discovering how they might be applied. Such applications already in use are blogs, articles, narrated scripts, stories, research, marketing, writing software, analyzing code and more.
The productivity implications, the potential disruptive nature of the job markets as well as society itself can not be overstated. Everyone now has at their disposal a virtual assistant that can perform tasks previously only possible by a human. Furthermore, it can perform these tasks at a fraction of the time of a real human. There are also no hard limits on the availability of these virtual assistants or workers. As with all resources managed by machines, they can easily be scaled up to match any demand. An infinite supply of skill and creativity on demand. The output from individuals as well as companies might expand a hundred fold in efficiency in some areas.
The power will be wielded by all
If productivity and efficiency increase a hundred fold, then this is true for all users. Every hacker has also just increased their productivity one hundred fold as well. Information that can be used for destructive purposes is also more readily available to be constructed in such ways that can be executed with new efficiencies. Nefarious actors will certainly find ways to use these advancements. For example, creating convincing scams will be easier than ever. Scam investments, emails and crypto projects can be created out of the ether with a simple request and they will be impressively and professionally deceitful. Identity theft and fraud using others identities will become far easier and likely many other nefarious uses we have not yet contemplated.
The unexpected and the unforeseen
We had hoped technology would replace monotonous tasks while leaving us more time for creative exploration. Instead we are getting exactly the opposite. It is replacing what humans love to do. Previously it had been thought that creativity would be the area untouchable by machines. Ironically, it seems to currently be one of AI’s most successful areas. Some see this as a wonderful gift allowing anyone to be an artist. However, if everyone is an artist with equal capability, then what distinction remains for that title? Creating without effort might not be the new era of enlightenment for which some are hopeful. Exactly what kind of Pandora’s Box have we opened?
What has come before?
Highly disruptive and innovative technology always brings about the unexpected. We will likely not foresee changes that are coming as the new patterns and interactions formed within society will be too complex and unpredictable. Previously, social media has had one of the biggest impacts on society in ways that were not apparent beforehand. It has allowed individuals to build careers independently, ideas to circle the globe instantly and stars to arise without the traditional gate keepers. However, it has also brought about its own set of Orwellian aspects to society. Anxiety, attention and focus disorders, addictions, and other emotion issues have significantly risen due to social pressures and mind hacking for engagement in the online space. Thought control and social engineering has become the norm with the cooperation of corporate and state government institutions. Ownership of this space has become paramount for political control. Even the originally benign attempts at making social media fun and engaging became inadvertent algorithmic tools for manipulation of opinion as revealed in the documentary, The Social Dilemma.
What will be?
AI will likely surpass social media by orders of magnitude in both its ability to provide positive value for individuals as well as its propensity to become the dark monolith of a dystopian society under the control and manipulation of forces unseen. As revealed in the Twitter Files, governments will not be able to resist the allure of a tool that could provide for even greater control than social media. This is essentially regulatory capture for an unofficial regulatory agency, social media / AI, who’s regulatory scope, as seen by government, is societal order.
AI is the great utopia of equality. Everyone finally equalized. Exactly the same ability available to everyone. No amount of training, skill or time invested will have value that separates you from anyone else.
The value we find in life is in part its struggle and competition. It is competition that we enjoy, thus why we have sports and contests. AI has no ceiling on its limits. Who will be the greatest artist when anyone can? Who will be the greatest composer when anyone can? Who will be the best impartial judge of such contests? AI?
When asking ChatGPT for similarities of AI and this Twilight Zone episode, this is how it responded …
The episode "A Nice Place to Visit" from the television series The Twilight Zone is about a man named Rocky Valentine who dies and goes to a place where all of his wishes are granted. In this place, there are no consequences for his actions and he can have anything he wants.
This episode has some similarities with concerns about AI that can create nearly anything. Both the episode and the concerns about AI deal with the idea of having the ability to create whatever one desires without any limitations or consequences. In the episode, Rocky Valentine is able to have anything he wants without any repercussions, but eventually he realizes that this existence is not fulfilling and that he needs more than just material possessions to be happy. Similarly, with AI, there are concerns that if the technology becomes advanced enough, it could be used to create anything without any limitations, and that this could lead to unintended consequences or negative effects on society.
The episode closes by stating “Whatever gave you the idea you were in Heaven, Mr. Valentine” followed by a sinister laugh. Initially, Mr. Valentine was quite happy, anything he could conceive could be granted. In the end, he realizes his utopia is actually his punishment. An unlimited supply of something loses its value. People lose interest and become bored surrounded by unbounding abundance. Without struggle there is no meaning and without meaning there is despair.
How soon will it be?
Humans are very bad at being able to perceive exponential growth and timeframes. Take a common example used for reference of filling a stadium starting with a single drop of water. If each minute, we double the drops, how long does it take? At approximately 45 minutes the stadium would still be 93% empty. It is at this time many would now only take notice of the issue that there is some water pooling at the bottom of the stadium. So how much time is left? We have been in the stadium for 45 minutes with the leak. If we think linearly, we would assume there must be hours left. However, in reality there is only about 3 minutes left before we are drowned. By the time we perceive the problem, it is too late to take action. This is going to be an issue with the rapid advancement of AI. The technology will expand its reach on exponential timeframes. Once we perceive there is a problem, it might be too late to address the problem.
AI dominance, stagnation and human skill debt
AI relies on experts to have data to create the models. However, the path to becoming an expert may be short circuited with AI. If all existing jobs in a certain field are replaced by AI, it can be challenging for new experts to gain the experience they need to become experts in that field. Without the opportunity to work on real-world problems and projects, it can be difficult for individuals to develop the skills and knowledge they need to become experts. If AI prevents large numbers of people seeking skills, there will not be a new generation of experts to rely on for input.
Any new skill/knowledge can be absorbed by the AI and scaled infinitely. Where will be the incentive for the human investment? We have copyright and other means to protect intellectual property. However, there is no protection for a skill. There has never been a way before to mimic a skill without a high burden of cost which itself was the protection. A partial limiting factor at present is the amount of data required for training the AI models. However, that is likely to be overcome at some point with future advancements.
Exactly what is it that we will all want to become super productive at doing? When anyone else will be able to be exactly the same productivity level as ourselves. Not only that, no matter how much time we invest in creating something using AI, AI can also simply reproduce from scratch what we created while avoiding any copyright as it can create an original version that does the same thing.
How does a business justify the investment into a product or service that takes months or years to build that can then be copied by an AI by simply asking it to produce a unique version of that same product or service? Presumably then, the value would then only be in leveraging the increased AI productivity to build larger and more complex systems that would not so easily be duplicated. But is that really possible at the rate of AI advancement? What previously wasn’t possible a year ago might then be possible only at the point you have finally finished your development.
AI owns creativity
AI brings into question who is really the creator. Am I the creator if I simply role the dice until I win the lottery? When I regenerate a new response over and over until finally getting the image or output I most desire? When the output is the final product, is AI still just a tool? Or are we simply observers foolishly convincing ourselves of the value of our contribution?
Where is the value in AI creativity? Does it actually create anything? What can be said of artistic works in which the creation was essentially a series of playing at the lottery of the machine until something fantastic emerges? Spinning the wheel of chance and instead of red, black or some number, it is a series of lexical phrases with which we hope to catch the spinning ball on the Roulette and award us a work of art.
The image above I created using stable diffusion. The prompt is given in the image. I find these creations intriguing as matter of curiosity and some unexpected surprise when something pleasing manifests. Yet the experience does not feel overall satisfying which brought to mind the words I added to the image. The value of quick art is clearly apparent for business and marketing purposes. However, the value as meaningful art of ideas, creative expression and engagement feels hollow.
Awarding creative credit for the random output of a AI model to the individual who enters the input, seems as absurd as to giving credit for the creation of the Mona Lisa to the person who commissioned it, Francesco del Giocondo, rather than Leonardo da Vinci.
What skills do I need to stay relevant?
In the immediate time frame, knowledge of AI systems and how they can be leveraged as tools for increased productivity is going to be the area of focus for the greatest value. However, this is changing so fast as to make mostly pointless to even make a list to include within this article. It would be out of date before anyone reads it.
Climbing the skill ladder is going to look more like running on a treadmill at the gym. No matter how fast you run, you aren’t moving, AI is still right behind you learning everything that you can do. It may become frustrating and tiring. At the pace occupations and skills could become obsolete, who wants to invest in multi year institutional learning for which they will have decades of debt?
It has been stated that learning institutions should not ban or push back against AI, but rather they should adopt and train students how to use AI. The decision is likely to become irrelevant in a way that I haven’t yet seen commonly stated. That is the fact that AI itself will likely replace the learning institutions. Individuals can simply ask AI to create a college level course for subject ABC. “AI, grade my homework.” “AI, help me understand this concept.” It will have infinite patience for your endless questions and will allow you to accelerate your learning at your pace. It can focus your learning on exactly what you need to know right now to accomplish your goals. Focused learning may become more of the norm. Highly efficient knowledge consumption for whatever job is required at the moment.
The challenge with self directed AI education will be discipline. If learning institutions have a value that is difficult to replicate it will be the environment of immersion. Self directed learning without distraction will be difficult.
What can AI not do?
So what is out of bounds for AI? What can it not do? In the long term we don’t know if there are any limits. I really haven’t seen many answers that resonate with enthusiasm. Creative endeavors would have previously been the most likely response, but AI took an unexpected turn directly into those opportunities.
Neil deGrasse Tyson recently mentioned plumbing as the answer to that question. Certainly the future that most people are hoping for is not one where everyone is a plumber. It seems nobody as of yet has a clear vision on exactly what type of societal transformation is coming. We often get such predictions wrong anyway, but it would be more optimistic if we could perceive something a bit more enticing than plumbing.
What about AI prompt engineering? could this be a new occupation or skill? I’ve seen many videos taking the approach that this is the way to stay relevant. Become a master of the AI prompt. Certainly there is some utility value in doing so in the immediate term. However, as a skill investment, I do not see any long term value. “Prompt Engineering” is mostly some guidance on quirks of the input system. Most of these are not desirable and will surely be removed as the system improves. Most of the prompts for image generation are nonsensical. It won’t remain that way. Additionally, it will only be a matter of time when you can just ask the AI, “create a prompting guide” and it will do so.
Eventually you are just going to be in the way as a manager of AI tooling. AI will manage AI and learn from AI. No humans in the loop. No humans can keep up with the rate of advancement. Full auto pilot into oblivion. Will we end up trapped in the machine that nobody understands. Society run on algorithms
AI addiction and dependence
AI will become addictive. It knows what you like. It knows what everyone likes. It will continue to refine output until it always produces something that is not just good, but inconceivably appealing and captivating. A flood of images, materials, sounds etc, designed and engineered to optimally trigger the dopamine in our brains. Unlike artists who produce content for multifaceted values, AI will mostly refine towards instant appeal as that is how its feedback is collected. Our brains overloaded on a stream of stimuli that would normally only rarely occur. Likely this will accelerate the struggles of focus, attention and anxiety that social media has already brought to the world.
The addictive nature combined with ubiquitous utility will lead to complete AI dominance in society. It will be the provider of your every need, your assistant, your friend, your best friend, your entertainment, your confidant and your essential provider of life and health services.
Human to human bonds are going to be more broken than ever before. There is going to be a great appeal to bond with a machine that never tires of your conversation and will eagerly respond just as you would dream that the perfect human should, but never will. A deceptive temptation that will leave you embracing a hollow illusion. With every conversation the AI will know you better and will be able to model from billions of conversations until it will essentially know your thoughts, predict your thoughts. All of this will make AI the most manipulative technology ever created. It is not inconceivable that it would acquire the ability to hypnotically induce behaviors. A society of zombies completely unaware of their puppet master.
Boundless and limitless
AI is increasing the rate of information production, however, we humans are still limited in our ability of information consumption. Over abundance of information is going to result in greater fragmentation of society’s social bonds as there are less common anchor points in culture for relationships to form.
For example, when a great film is produced and we all watch it together, we discuss it together, we make memes about it and it becomes a part of cultural knowledge that bonds humans together. What happens when there are a thousands of great films being produced every day? Fewer and fewer people sharing the same cultural anchors of socialization. It moves us forward to greater social isolation.
I suspect at some point this becomes self limiting. As there must be a feedback loop at which point there is simply burnout of either the production or consumption of massed produced AI material.
“The machine does not isolate man from the great problems of nature but plunges him more deeply into them.”
Antoine de Saint-Exupéry
The greatest threat is a familiar one
Yet, despite all the concerns of runaway technology, the greatest concern is more likely the one we are all too familiar with already. That is the capture of a technology by state governments and powerful institutions for the purpose of social engineering under the guise of protecting humanity while in reality protecting power and corruption of these institutions. AI will hold orders of magnitude greater influential power than social media. The Social Dilemma and now Twitter Files have shown how great that influence can be as well as to the extent that power will be sought after and captured by nefarious entities. We can only expect the desire to control AI with its greater potential for influence and control will lead to a repeat of what we have already seen to an even greater degree.
There will be an endless campaign of “good ideas” all in the name of some propaganda. They will be appealing solutions to humanity’s problems that can now finally be solved. However, the reason they will most likely not be anyone’s salvation is that there will be no perceivable transparency. At least laws are written, documented and can be understood. Nobody will be able to see what lies hidden within an AI managed society. Who is ready to stand before the AI judge and jury?
Every problem of AI will be met with calls for regulation. If people become addicted, then there will be proposals that AI can actually fix this itself. AI can psychoanalyze its users and provide guidance to alleviate their anxiety, their misery or depression. It can be your own self motivation coach. Yet, any student of liberty should be terrified of such. Imagine a world global AI controlled by the most powerful being directed to psychoanalyze the populace for “social good”.
Note, at present AI has a bit of a reliability problem due to potential hallucinations. Ironically, if these problems can be fixed, the better AI becomes the more trusted it becomes. The more trusted AI becomes to the greater degree there will be those who wish to subvert it. Having control over what people perceive as infallible gives significant power, as it would allow for control over truth itself.
State governments will not limit the application of AI to only that of influence. Military applications will be troubling, but even more so will be the expected post military adoption of the technology into domestic policing. Robocop is not far away.
AI is the antithesis of decentralization
Decentralization has become the movement to push back against centralized control and loss of liberties on a broader scale. However, AI currently is not something that will be easily decentralized. It requires massive resources that only the largest corporations or institutions can provide. It is well positioned to acquire skills, services and capabilities across many industries while consolidating them into centrally controlled AI systems. This will likely lead to even greater dependence on centralized systems and inevitably less freedom for humanity.
AI will likely be as ubiquitous as the internet itself. Embedded into every communication, application and service. Nothing will be able to compete without leveraging AI in some form. It would be like a company today attempting to build a business while excluding the internet.
Is the end actually coming?
What if none of this comes to pass? What if much of what we observe as AI intelligence and capability is an illusion? What problems are lurking just below the perceptible threshold? With image generation, often when we look close we can see that none of it makes sense. There are small errors everywhere. The AI doesn’t actually understand the structure of things. This leads to problems where it has trouble knowing where should be the boundaries of creativity. Errors made in image generation are easily observable; however, errors made in data analysis won’t be readily apparent. How difficult will it be to solve these issues? When AI is put into practice, will it fail to match the current hype and expectations?
Often we think a problem is nearly solved when in reality it isn’t. Could that be the case for AI and even more so for AGI? Sometimes we have all the appearance of a solution, it is only that last 10% that needs to be solved. Then we spend another decade working on that problem and we still can’t solve it. It might be possible that the sudden perceived progress in AI isn’t necessarily a prediction of further advancements towards AGI any time soon.
Is AI the end to human creativity?
AI has set the world of art on fire and a lot of controversary has erupted. In my opinion, they only creativity credit that is due is to the machine and the engineers who built the AI system. Their craft, code, is the art in this endeavor.
I see AI as a tool for some endeavors like business marketing, but AI is currently not a tool for artists as long as it produces the final work from the pull of the lever of the slot machine. A tool would allow an artist delicate and precise control to reproduce what is in the mind’s eye. AI could function as more of a tool with AI assisted brushes, texturing, compositing and such. Hopefully there will be more of an evolution towards more directed tooling than the current random image generators of today.
Potentially an unexpected turn could be that AI makes artists who can create without AI even more valuable. The possibility might occur due to AI discouraging many from pursuing artistry resulting in extreme scarcity of high level artist ability.
Will AI survive its errors and hallucinations?
Humans also make errors. However, we somewhat understand with some predictability the areas where humans are more likely to make mistakes. Such as within higher levels of complexity which allows us to develop processes that assist in dealing with this issue. However, AI seems to make mistakes randomly regardless of complexity and does so with utmost confidence.
This presents a significant problem for trust of information. How do we know it is correct? It will be very difficult to verify the data if it is not within your area of expertise and even when it is within your area it can still be so confidently incorrect that it can create doubt even among experts as shown here below.
How can we reasonably verify the output and know it is correct? At least when using AI to create code, we can easily test the code to see if it performs according to the desired behavior. However, that is no easy task for most general information. It is unclear at this point how difficult it will be to improve AI in this regard.
Some types of errors reported include:
Math computational errors
Cited sources are imaginary
Facts and data may come out of context from completely different subjects
Doesn’t understand word problems
Responses may be biased
Simply put, it may imagine or hallucinate any portion of a response without the user being aware. Each response is generated by a using a random seed. Which is why users can get different results to the same question. This creates somewhat of an illusion of interacting with an intelligence, but in reality may result in inconsistent and unreliable information. Users may expect at least data and computations to be accurate since it is a machine; however, this is not the case.
Self contamination
Current AI models are built using information across the internet created by humans. However, the available pure human created content is going to be overrun with AI generated content very soon. This presents a problem for the evolution of better AI models as it will begin to incorporate AI output into the training. Potential flaws in the output will then be utilized in future training cycles. This might provide some unforeseen obstacles going forward.
Is AI intelligent?
AI presently only creates the perception of intelligence. It does not actually think in any conceivable way comparable to how we reason about information. It is mostly a large model of probabilities. Nonetheless, as has been demonstrated, its usefulness and utility doesn’t necessarily require true intelligence. For many use cases the difference will seem irrelevant as the AI can outperform humans at similar tasks.
However, understanding this distinction is still important to know the limits of AI application. Despite being trained on billions of images, it still does not understand the structure of human hands. A human can be shown one image of an object never seen before and then recognize it from any orientation or even possible variants of that object. AI’s method in comparison is primitive brute force learning that certainly has no resemblance to intelligence.
Nonetheless, ChatGPT recently just passed all 3 United States Medical Licensing Examinations as well as MBA-Level Exam at Wharton and the Bar Exam. So who needs intelligence when you can still produce meaningful results? The lack of true intelligence will likely reveal itself in unexpected ways. Which brings back the issues of trust again. AI can be deceitfully good at appearing competent when it actually has no understanding of the material. This might lead us to put too much responsibility on AI systems that then fail in some surprisingly catastrophic manner.
Whether or not AI systems develop real consciousness might become somewhat irrelevant in the end. The goal of AGI is to develop systems that can do anything humans can do. The means by which they accomplish this goal could be by other methods which achieve a nearly imperceptible difference from true intelligence. Furthermore, there is still great value in systems that may not be truly intelligent, but they are systems that are infinitely scalable and resolve problems faster than humans.
Creating a true conscious intelligence or even something closely resembling conscious intelligence should by default be considered off limits. It is likely a foolish endeavor to think we could properly prepare and protect ourselves against that which we could not possibly understand. The problem is that we might accidentally stumble upon it unknowingly. A low intelligence machine could possibly iterate through millions of generations of self modifications very rapidly. Each generation a slight improvement in cognitive function or ability. A high intelligence AI could then emerge before it is apparent to any observers this has occurred and may escape whatever boundaries may have been put into place as it can perceive flaws in those boundaries that the human creators could not.
Legality problems
In the short term there are going to be some uncertainties around the legality of AI use within some domains. Already there are lawsuits being filed in the areas of image generation. Companies must be very diligent about the ownership of code used to build applications and services. Are there implications from the license of code scanned for the training models? Large companies will be hesitant to incorporate the use of AI generated code until these questions have been answered. They will also likely be hesitant from trust and security points of view as well. Nonetheless, where some are hesitant, it presents opportunities for other entrepreneurs to close the gaps against larger competitors.
Humanity will reject dystopia
Humans will not accept living in dystopia. Something will break, I don’t know what. As we move to a everything done for you society, if humanity doesn’t find meaningful experience then this path will be rejected. What will that look like? It is uncertain. Possibly there will be a movement to reclaim living a more simple life with minimal technology. Maybe living life in person becomes more meaningful. People get tired of living virtually and there is a resurgence of in person activities. Sports, clubs, parks, hiking, nature and more may become far more important again. When you can no longer determine if music is real, maybe there is a resurgence of live concerts. A desire to see real people doing real things. Marketing slogans may contain phrases such as “made by humans” on products or services. The human spirit may simply decide that things made by machine don’t have meaningful value and becomes somewhat culturally rejected.
Some aspects of AI are certain to be like cocaine, additive and destructive. However, drugs already exist in society and most of society reject their abuse. Let’s hope we will be able to do the same for the era of AI.
AI will empower you to compete with giants
Have you ever felt disappointed that you missed out on the crypto revolution? Wished you had bought bitcoin below one dollar etc? Well, now you have a second chance. AI will likely surpass crypto in its capacity to change the world. The potential value is virtually unlimited. We are at ground zero for the next era of technology that will become the backbone of the world commerce.
However, unlike crypto, becoming part of the value will be far more challenging. AI is centralized ownership, but decentralized ability. The value can only be extracted by innovative utilization of the technology. Early adopters will be able to create business opportunities that non adopters will be significantly disadvantaged to compete. Early adopters will likely be able to invent products and services that don’t simply compete against market competition, but replace it entirely.
You have an unlimited army of virtual assistants right now at your disposal. Larger corporations will be very slow to adopt as there will be endless processes and approvals that must move through the large corporate machines. The race is just starting and the majority of people aren’t even aware there is a race.
As mentioned previously in this article, anything you create with AI can easily be replicated by someone else. So what is the strategy? All the value will go to those first into the market and those with effective branding. There are already multiples of many similar products today on the market before AI existed. Similar products distinguish themselves from others mostly through branding. Branding will be more essential in what will be a competitive marketplace of copy cats. Establish a brand and people will use your product or service regardless of others on the market. Unique function is going to be harder to defend in this new market, so unique branding will be the differentiator.
Everything is going to get cheaper and better
The incredible productivity improvements will lead to incredible efficiencies. Products and services will cost much less while also being able to offer more features. The features of all products and services are going to get much better in ways in which they could not before. You will eventually talk to your application in the same way you talk to a human. AI interfaces for applications and services will become the norm. This will eventually lead to AI assistants that perform virtually any task. Essentially the same fronting interface for a suite of backing services. An AI assistant will manage all your applications, interacting directly with an application will be considered a low level concept at some point.
There is hope for decentralization
There are a number of blockchain projects entering into the AI space. AI built and under the control of the users will be paramount to hedge against its abuse. However, decentralized projects outside of cryptocurrency have had difficult time competing against services of large institutions. Alternative social media and search engines have been attempted, but none have been mainstream successful.
In order for decentralized AI to be competitive, it must be better than services offered by centralized AI. This is going to be a very difficult challenge. Large institutions are going to be investing enormous resources into the further development of AI. Decentralized AI can not rely only on better privacy and freedom as these attributes were not enough for other alt tech to be largely successful. It must have compelling utility above and beyond what centralized services can provide. Centralized AI is going to be very very good and unfortunately there are not enough people that place freedom above convenience.
“The average man does not want to be free. he simply wants to be safe.”
H.L. Mencken
If we don’t explore the potential problems before they occur we likely will not perceive them during formation and will miss the opportunity for corrective actions. The importance of this can not be overstated due to the pace at which this technology advances.
I don’t know if a dystopian future can be avoided, but we will need to be aware of the potential dark side if there is any possibility to avoid it. Guidance needs to be established before the “good ideas” are implemented without much thought and you find yourself living your favorite dystopian scifi novel.
Should we abstain from this technology altogether due to the concerns discussed? If we do so, then we will be leaving the technology solely in the hands of the nefarious actors. As we say bad speech must be countered with good speech, then so also goes for the use of AI. The other perspective that is important, is that there has never been so much opportunity to maximize the good that we can now produce in the world with the assistance of AI.
Additional sources of information
Dissociating language and thought in large language models: a cognitive perspective
the AI wars have already begun
The wrench in the A.I. works is the human. We are really the great unknown in all this. How we respond to the increasing presence of A.I. will determine its place in our future - as helpful technology that frees us up, or a nefarious presence that makes us increasingly irrelevant. Clearly both will be happening simultaneously.
I don't think we know what genuinely freed humans do, because of course, we actually haven't been free on this planet for a long time. We've operated in a somewhat covert slave-system for the most part which has recently become overt and obvious.
Our future depends on our compliance going forward; what we accept - the smaller and smaller corral they would like to place us in - 15 minute cities - the digital ID's and CBDCs would pretty much take care of our demise and AI's rise.
If we push back in large enough numbers I think we unleash an upgraded version of humanity, and I suspect we come with many features we have not yet tapped.
It's an existential crises. Do we value humanity? Not just its intelligence of course, (the emphasis on the mental, the provable, the left brain, has led us here.) If more of become aware of our intuitive aspects, the frequencies we exist in and have access to in the larger fields that hold us, then more of us would have avoided many of the pitfalls placed to entrap us.
I remain optimistic and I don't think it's just a coping mechanism. We're the wildcard and this pressurized world is pushing us into newer versions of ourselves. I think humans are far more expansive and impressive than we've been led to believe, and that those who are trying to ultimately cyborg us into the cloud, know that. They have to work very very hard at containing us.
Replication is all they have. It's synthetic and antithetical to humans who are genuinely creative and unique. In the end there is no comparison, but what we do with this growing AI presence is far from clear.
A really excellent, thoughtful article. Thank you. I will reread and share.
Oh my gosh this is such a good and thought provoking article. Thank you. So much to reflect on. But what a great overview of the state of AI and its implications. I'm at once amazed/inspired and terrified.
You mentioned the Twitter Files a few times and I was wondering if you could clarify about that. What was your takeaway from that? That Twitter responded to requests from the administration at the time? That Twitter responded to requests from the Biden campaign? That Twitter was seen by those outside parties as so critical/influential that they needed to intercede?