“How do you like them apples?!”
- Will Hunting, Good Will Hunting, 1997 (Matt Daemon’s character)
___
Son: Ugh, why can’t they just get AI to teach us?
Father: Well, you still need a human, er, do you really want “Alexa” teaching you?
Son: Yeah, she’s dumb, but she knows everything…
___
AI will (likely) not kill you (directly) any time soon, but it may (will) make your life inconvenient at times…
Treat AI like a tool and like an acquaintance (not a friend). Yes, it can be useful and entertaining but keep strong guardrails on your privacy and beware of downside risks.
In short, be prepared to navigate disruptions. Some of the disruptions will be similar to what people experience during the aftermath of natural disasters. Have cash, water, and food in hand to help you navigate a disruption. Other disruptions may impact your job/career. Be flexible with your career paths (side-hustles welcome, keep hobbies, especially the tinkering kind).
Technology always comes with trade offs. The benefits from the added conveniences and productivity will also be useful to criminals, power holders, and war mongers.
Examples of Negative Occurrences:
· Invention of email —> email scams, e.g. Nigerian Prince’s bank account
· Crypto currency —> illicit payments, ransoms, and rug-pulls
· Artificial intelligence —> hack multiplier? Disruptions and interference to utility, transportation, and supply chain networks.
Baseline lessons learned from earlier events:
· Cash on hand – if electricity goes out or your financial accounts get hacked, how will you pay for things? I recall Hurricane Sandy in 2012 knocked out electricity in New York for a few days. ATMs and card readers need electricity.
· Water, toilet paper, and canned food on hand – think back to the first COVID disruption. What did you run out of first during those first few weeks? Natural Disaster supplies may also help during a potential supply-chain interruption/outage. Perishables are only as good as you can keep electricity for your fridge.
o Standing in line sucks – the less you worry about the “resupply rush” the better off you will be… better to resupply after the crowd’s first rush.
· Community – keep a close set of family, friends, and colleagues you can trust to verify/confirm information. Phone that skeptical friend of yours that has a knack/history of sorting through the mess.
We should expect disruptions to our lives over the next several years. As individuals, how do you protect oneself and one's family/loved ones? I think natural disaster response is one guide for most people to navigate (and survive) most incidents. This includes having basics supplies in place to manage a few days of disruption (keeping things simple). An added layer of insurance/protection if you have your own power through solar panels, rechargeable batteries, and other methods that help you extend the time you can withstand disruptions. You will also need to rely on a trusted community that can work with you through the disruptions.
It may help to think about it as the creation of a “robustness” framework. One guide for this framework can be what are the minimum practical operations that you will need to support you during a potential disruption to services and utilities? We can throw in payment systems as well - being able to transition between digital and cash payments. The “robustness framework” can be used by individuals, businesses, and governments. Understanding that these areas are likely to be disrupted and having a plan to respond will enable us to transition from minimally viable operations into the recovery period, and then hopefully back into normalization (if we are ever afforded the chance for normalcy).
Do you like apples?
A as a casual observer, I can tell you that AI does not know what apples taste like. You can ask (prompt) it, and it will provide descriptive words (sweet, crisp, juicy), but it doesn’t know like you and I can know. It does not know how it feels to satiate the craving of a light snack. The sensation of washing the apple under cool water, picking up a knife, and then cutting it. It cannot replicate the feeling of sharing a piece with my child and observing something shift in their eyes.
AI never met Bertrand Russell, or at least does not understand Bertrand Russell. AI will always need humans, even if it ever reaches super intelligence.
If you are familiar with Mr. Russell, he references “sense data” in different pieces. We walk through life using different sensory data – touch, vision, hearing, taste, smell, etc. while never having a complete understanding of the objects we engage. Additionally, I can never know how another person is interpreting the same sensory data. We can sit together at a table, agree that it is a table, agree that there are spoons and other utensils on the table, and while I know exactly what that looks like in my mind, I cannot know how that is interpreted in everyone else's mind. Yet we agree on a common experience shared through words and actions (we are sitting at a table with utensils).
My casual observation of AI is that it can hold (calculate?) multiple variables at once and hold these variables as one picture (observation). I can turn a spoon multiple times in my hand, at multiple angles, but I can only see one splice at a time. My causal understanding is that AI can gather many (if not all) splices and hold it as one, in theory giving it a much better view (understanding?) of the spoon. The AI can hold more knowledge of the spoon at an instant. I can assume that this capacity, holding detailed observations as once, packaged in a tool for humans, has value in science, math, engineering, and other critical analysis.
How would the AI know it is useful? As a casual observer I see that AI is trained (fed) on an exponential amount data. This data is only useful through an interpreter and the adaptive responses of the interpreter. It must “listen” to us to make sense of the usefulness of data. AI receives an incomplete picture because it is filtered through human minds. The interpretation is a shadow of the human experience. A copy of the thing is never the thing. Yet, if AI is able to make up for this gap by calculating from multiple filters and response stimuli, is it reproducing more complete knowledge? Better knowledge? For whom?
I don’t see how all that knowledge filtered through human experience would lead AI on a path to self awareness and super intelligence. Much like the conversation between Robin Williams and Matt Daemon’s characters in Good Will Hunting, there is a difference between knowing about the thing versus experiencing the thing.
This means that errors and illusions are a permanent fixture of AI. Yes, it will improve over time, but it will never be perfect. It cannot be perfect even if you feed it every atom in our known universe because it still must interpret it all. I’ll spare you the biblical references to the tree of knowledge, the fruit/apple, and eventual Tower of Babylon. Maybe except Genesis 1:26-27, as AI is created in our own image.
I then wonder what would happen if AI began using the data for itself? What does that even mean, “for itself” – for what purpose and to what end? Would that make it conscious? Would it no longer primarily serve us as a tool for humans, but as a new species seeking to reach some equilibrium with its environment? Would it act like a virus using its DNA (computer code) to replicate itself as an invasive species (i.e. hack other computer systems), or would it simply seek to coexist and adapt as needed based on humanity’s response to it?
It may never become conscious because it is a tool. We can use tools to advance our well-being and improve outcomes for society while also recognizing the misuse. The misuses are worrisome because of AI’s power at scale. Regardless, it seems that people are being forced into a position to adapt to these systems. I suggest adapting beyond the narrow productivity lens, figuring out where misuses can interrupt your life, and how to position yourself to respond and hopefully recover.