You are probably overthinking AI
here's what to do instead
Artificial intelligence is neither doom harbinger nor salvation herald.
And every minute you spend convinced it is either is a moment wasted.
Worse, it is a minute spent complicit in suffering.
Why is this?
Because as with every moment in history, we have no way of predicting the future.
"But this time is different!" Of course it is. The only way it wouldn't be different is if history had already stopped. This moment, by definition, is novel.
It is unlike every moment in our shared past.
And yet, every moment so far has been the very same.
We have always existed on the knife’s edge between advancing too far, and not advancing far enough.
Thinking about how well perched we are on the knife has always been vital. But it comes at a price.
Worry too long about which way we are currently leaning and you leave the leaning to others.
And it's too easy to waste time looking for a clean answer.
We don't need a clean answer. We don't need to resolve to only shift our weight in one direction. We need a real-time recalibration and rebalancing driven by as many of us as possible.
We need everyone to experience and draw conclusions. Not sidelined in armchairs making hypothetical doomsday vs. salvation calculations.
Quit wasting time speculating.
Instead, start solving as many problems using AI as possible. As fast as possible.
This includes solving problems inherent to the solutions themselves! In other words: use AI to not only fix things but to FIX AI itself!
You can do both at the same time. That should be our conscious goal. It is our responsibility either way.
So jump in. Realize that these are harmonious repairs. The saying is “if it ain't broke, don't fix it.” And if you only use one rule of thumb for AI, that should be the one.
Use AI to solve what's broken. Even if what's broken is AI.
Broken areas of our lives to fix WITH AI would include:
war
disease
famine
confusion
ignorance
poverty
and all other forms of meaningless, useless suffering
Broken areas to fix WITHIN AI would include:
faulty reasoning
limited human control and transparency
lack of empathy for living beings
potential for corrupt people to corrupt it
tyrannical applications
easy devastating use in warfare
potential to concentrate wealth too much
extreme energy usage
hidden prejudice
perverse values (e.g. kill humans to stop environmental damage)
There's no need to be a saint about it either. Working to fix things with AI will also help you in many ways.
For one thing, engaging with AI improves your ability to think rationally about it.
This is because engaging with it alleviates your anxiety. Engaging shows you that you have control and agency over your own life and its course. Even when it comes to AI.
More so, by engaging you are better able to realize where its shortcomings lie. This means you will begin to feel able to steer its development in a more positive direction.
And even if you are unsuccessful, and AI causes problems. Or even catastophes. By engaging and being less anxious day-to-day, you will have enjoyed the journey more.
But we don't need to assume the worst. Very likely, by jumping in, you can save some number of people from needless suffering. You can save lives and improve the quality of life on earth all by choosing to engage with AI.
This has been very obvious in my own life. Growing up, robots and artificial intelligence fascinated me. I feasted on books about the subject. My favorites were hopeful stories. I especially loved the Robots in Time series and iRobot, all written by Isaac Asimov.
In 2017, season 1 of HBO's Westworld shocked me off my ass. Robot consciousness having moral value and intrinsic dignity was more romantic than Shakespeare. Ramin Djawadi's mind-blowing soundtrack and Anthony Hopkins' terrifying acting didn't hurt either.
This was so moving to me that I taught myself to program in Python. I self-taught data structures, algorithms, and a splash of machine learning too. I dreamed of helping to spur on the advances in artificial intelligence.
But this optimism and glee was short lived. I soon slid into AI Doomerism. I became convinced that AI would replace all humans. That we had no future in any area. And that we would soon spiral into meaninglessness and implode.
In existential terror, I started an organization called Vocation Amid Automation. It focused on preparing humanity for our inevitable replacement. This was during the covid-19 pandemic. I had driven across the US from Virginia to live alone for 9 months on a frozen lake in Washington state. I spent many cold covid lockdown days searching for answers. "What will we all do when our problems begin to fade away?”
All told, I'm glad I had both of these reactions, because they have led me to where I am now. I had to feel through these emotional states. I had to embrace the distorted worldviews they enabled.
This was the price for finally reaching higher ground. Somewhere with a view and clearer thinking. I hope at least that I can claim a foothold there now.
I want to encourage you to remain humble in your prognostications about the future. I hope more of us can set aside these impossible questions. Otherwise, how free are we to face the realities here and now in front of us?
Accept whatever reaction you currently have to AI. But make sure to start using AI regardless of how you feel about it.
Here is a place to start: identify a boring, undramatic, but repeated challenge you face in your daily life.
For me, this is the infestation of mosquitoes in my apartment.
The problem isn't dramatic.
I am.
Without AI, I would have already spiraled into frustration. I would be swatting with rage and feeling powerless against my pointy adversaries. They are far too nimble.
But with AI, I am mighty. (And even more dramatic, apparently.)
AI whips up instant systematic evaluations of my situation. As many times as I need!
It matches new solutions to the problem as I tackle it. Every update refines the approach.
I'll explain how. While I do, think about doing this for one of your own problems.
First, I asked ChatGPT for help with my mosquito infestation. I gave it a description of the bathroom area and sink. I listed the number of drains and the quantity of mosquitoes present.
"GPT, Chati, I'm scared. There are TWENTY mosquitoes in a 5 foot radius. They're right above my toilet and yesterday's dinner is cresting."
Dramatic but accurate.
Chati did its best to help me, suggesting that I pour bleach and hot water down the floor drain, and then cover it. (It had no suggestions for keeping number 2 in. I walked to a restaurant and violated their facilities. Desperate times.)
After this made no discernable impact, I prompted it again. I noted how the second drain, in the bathtub might have more water in it at any given time. I did not think this was the culprit. How could they come out from the bathtub drain???
My understanding of mosquitoes and drainage pipes was quite limited.
Chati immediately insisted that the bathtub drain was the culprit. Insisted!
This created a paradigm shift for me and humbled me, as it destroyed my earlier assumption.
After these 2 ChatGPT interventions, the number of mosquitoes dwindled from 20+ to 2 or 3. For me, this is still an unacceptable number of buzzy pricks in my living space.
Therefore, I am implementing a maximally thorough anti-infestation protocol.
Here is what is called for:
an exhaustive overview of context
what I have been experiencing in timeline format
threat sources
a full description of my living space. this should include all possible standing water sources and entrances to the apartment
attempted solutions
everything I've tried and what it has yielded
status over time
how many pricks have been present each day
arsenal (weaponry to destroy them with)
everything from bleach to hot water, RAID, and bug spray
other options not considered
what else can be done, Chati?!?
One thing has become abundantly clear. AI can improve our thinking. I'm ready to bet that it can already improve our thinking better than most other stimuli in our lives.
Exceptions include trusted authorities and confidants. And anyone else with the ability and desire to be more selfless when helping us. Other than that, AI has a potent blend. It's dispassionate avoid offending you. But it's also hyper personalized and attentive in its responses. It will rephrase and re-strategize as much as you need.
If we use AI to improve our own thinking, it will do exactly that.
But it can't do that when we overthink it. When we fret, we lose.