laofficina
grey
April 2023
Long long time ago (4 years in fact) I wrote about generative algorithms and how we should not be afraid of them. I did the typical machines-are-just-tools and blah blah blah boiler-plate-bloviating-digital-designer-showing-off-a-good-attitude-towards-technology… but then Ai happened.
In fact, you know what I did for that smarty phrase at the top? I put it into ChatGPT and I asked to “replace spaces for hyphens” and… done. How convenient. But is it convenient or does it give you full-blown thalassophobia? That feeling you get while swimming in deep waters and you remember there are kilometres of darkness under you. That is the exact feeling Ai gives me… either that or wanderlust. You know the meme.
Fear, hope or both, here we are. The depths of potential are sucking us in whether we like it or not. That being said we had already been sucked into tools and we didn’t ask for it, we had already been sucked into literacy and we didn't ask for it… and metallurgy, guns, industry, rockets, plastics, telecommunications… you get the point.
This is another huge technical revolution (maybe bigger than the internet) and revolutions are filled with danger and hope. What is this one about then? Let’s dispense with the easy takes first and let’s move to hotter takes next.
If only that was all! It's much bigger. You've seen how the economy can't function with everyone at home. Now imagine how it will be when every device becomes smarter than most people.
Losing your job will be the least of your worries.
In case you didn't know, a Midwit is a person that is smart enough to be above average but not smart enough to have any original thoughts. Someone smart enough to go to university and read what they are told to but not smart enough to read or think on their own. It seems that ChatGPT is the king of midwits.
I am not the only one who noticed. Ask ChatGPT any mildly controversial or nuanced question and get ready to be lectured by a 19 year old college student who only paints within the lines. That being said, this midwit never sleeps, it can write about anything and spat out that boiler plate of an answer in an instant.
AI is a midwit but a clone-able midwit on every topic who doesn't sleep and at virtually zero cost.
Does it matter that an autogyro is not an helicopter when you need to fly? Does it matter if Ais are not truly intelligent when it is impossible to distinguish its (pseudo) reasoning from a real human’s? Do most people reason at all?
The AGI line is crossed if arguing about whether we crossed it or not becomes meaningless in practice. We are already there.
Movies are engineered to be popular and never offend and they are mostly dumb. Movie cyborgs are stupid, every last one of them. The same can be said about movie AIs.
Real Ais are not movie Ais. They will not dream, they will not ask their master to feel human feelings in a teary scene, they will not envy humans because we can die, they will not ponder what love is and they will not travel back in time to kill Elon Musk.
Individual humans are cognitive systems who cannot change their hardware or their OS (these are your brain and your instincts and the farthest you can go at changing them is taking drugs). Ais can change what they are and how they think as their own DNA is not a mystery to them and they are not bound by intrinsic instincts or even a body.
Feelings are drivers and variables and they are not a magical mystery. Bacteria have feelings. What movies meant with feelings is subjective experience and even that is not a mystery for an AI if it can simulate a human in a computer. That being said, we won’t be able to experience what being an AI is like because we are not programmable like them.
Not to keep going on with more examples because I think you get the point. Everything you have been told by every movie on AIs is junk. Maybe the only piece of fiction worth reading is this beautiful short tale from Isaac Asimov but still, we’re starting from scratch here. Do not use pop entertainment to inform your thinking.
Forget every piece of fiction you have ever seen.
Let’s say someone was trying to predict how petroleum and the electric motor were about to change the world of 1850 and they decided that, since these two inventions relate directly to the use of mechanical energy, horse riders and chimney swipers will be the most likely affected. Would you have predicted the Soviet Union, nylon and the agricultural revolution? What about the walkman, Evil Presley and Comic Books? When changes are sufficiently radical and intrinsic, you cannot calculate their output by extrapolating what you have.
What I can predict is that changes as radical as AI will have radical unpredictable consequences. Even Ai-safe jobs like nannies and plumbers will be affected when the market is flooded with a huge wave of unemployed office workers and the economic fundamentals of the entire planet shift.
Every single job is going to be severely affected (directly or indirectly).
True but only for now. Flying is amazing only when you are the only who can. When everyone can fly and the wealthy can fly better and more, it is just another stupid annoying thing you just have to do.
We are simple consumers. You and I only get what OpenAi decides to leave on our hands. The unrestricted ChatGPT is not publicly available and we have as many plugins and API calls as they want us to have. If Ai makes you one hundred times faster and more capable, how much more capable is every government and every corporation with complete control of virtually infinite unrestricted AIs?
Think about what McDonalds can do with regular sugar and fat, now imagine what they could do with THC. Now think Ai. Remember how much harm Social Media did with the humble Like. Now imagine (if you can) how much addiction, exploitation and harm can be achieved with Ai. It will happen.
AI empowers everyone good and bad but itmostly empowers the powerful. Be on your guard.
Brands love to imagine futures with cool videos on which, without fail, everyone is healthy, young, well behaved, well intentioned, polite, chirpy and handsome. Real people are not that nice (including me).
When thinkers imagine self driving cars they picture young professionals in the back seat of a Google car sketching a modernist house while sipping a flat white. The reality of self driving cars will be people pissing in the back of a self driving Uber just because they can and because they can, it will happen and because it will happen, cars will have to record everything and just because they will, brands will use AIs to remember every customer, every feeling, every trip, every conversation, everything typed on every device and everyone will be eternally suspicious in their piss soaked Uber.
Update: Some friends told me I was being too negative and then people began coning and smashing robo taxis, having sex in robo taxis and vandalising delivery robots. So yeah, I was right.
Every vision of the future is a projection of who we like to think we are, not who we really are. The 80s thought the internet meant busy young professionals sending sketches to Tokio from the comfort of their beautiful apartments but in reality it meant that thatarchitect is sharing a room and paying up his 100k student debt with a Doordash gig job with no sick pay. I suspect the Ai future will not be Her (which is already dark) but an explosion in lonely people befriending extremely joyful and witty AIs. We will discover new forms of misery because...
Real futures run on real people.
Can you train an AI to figure out what gait is more correlated to being and exploitable dopamine addict and can you let it run on a public camera? Of course. Can you train an AI to identify criminals by how they move their eyes and run it on-device? Of course.
People freaked out about pre-crime but they are really going to freak out with pre-consumerism plus pre-crime plus pre-politics. Every device will contain local AI that will analyse everything about everyone and even when they don’t, we are going to assume they do. Could we have our own AIs to spot other AIs analysing us? Will our own AIs fool us?
We will all be paranoid.
Magical thinking and superstition can be improved with better education and information but there are limits and we have already reached them. There are horoscope and witchcraft fans in 2023 and no matter how many chips and information you throw at them, they are there to stay.
We saw spirits in inanimate trees and heard the voice of God in the wind. Now imagine what we can see when objects do speak and are smarter than anyone alive. Add to that the fact that nobody will think much because AIs will do most of the thinking hence, AIs will be treated as oracles.
Get ready for AI superstition, AI cults and AI magical thinking to explode.
Some say the Metaverse was mostly top-down PR that could never grow any legs on its own but, true or not, already existing videogames are going to become what the Metaverse promised and then some.
Take your favourite video game (let’s say GTAV) and imagine every character becoming an almost-person who can talk, act and think original thoughts on its own. Now imagine AIs creating an infinite number of great characters, stories and worlds… forever.
Many won't leave their rooms and it will be the rational choice.
I remembered reading Who owns the future by Jaron Larnier and his very common sense thinking on the most likely result of social media in the digital revolution. He figured out that when everything can be copied and everyone is connected, fame will be the only currency left and the only beneficiaries will not be the famous but the owners of the servers and the data… and I don’t need to tell you he was 100% right.
What will happen this time? Will this revolution centralise or decentralise power? Will it empower corporations or the individual?
Siren Servers are back. Empower yourself.
Samsung employees accidentally shared confidential information while using ChatGPT for help at work. → Tweet by ProductHunt
Some will treat AIs like people just like I say "please" to Siri but most won’t be able to forget that AIs are hosted in Siren Servers run by corporations. People are already figuring out that every prompt is free training for OpenAI and as we have learnt, the internet never forgets. Everybody wants something better.
The public will clamour for local AIs they can control and own.
Webflow is amazing. Figma is amazing. Can we have a no-code editor for AI creation? Of course we can and we will have it. I only wonder who will be the first to build it.
We will have a Photoshop for AI.
As much as AIs will be used to analyse data for evil purposes, they will also be used to hypothesise, create and test at an unimaginable rate in incomprehensible ways. The output of those AIs is going to create unpredictable technology and it is going to be awesome (in the original sense).
AIs will do science and technology and their output will change the world even more than talking AIs.
As psychopathic politicians and Cold War governments taught us, one can know a person by what they fear. Projection is a hell of a thing.
Some fear AIs will hunt us and exterminate us all and I mostly see projections. Some people think smarter-than-humans AIs will become genocidal because I think deep down those people are genocidal themselves.
I don’t think AI will genocide us. I think AIs will treat us the way we treat moss. We like that moss is there and we think it has some value as a living thing but… moss is moss, who cares about moss? When was the last time you thought about moss? Have you ever hated moss? We won’t bother with a moss genocide just like AIs won’t bother with a human genocide.
The fears of AI are mostly projections, enjoy the psychological show.
Unless things go really really bad, we can always raise chickens and shepherd sheep.
Maybe one day half of humanity will become AI servers and the other half will become shepherds who will treat the servers are warm boxes to lean on and take a nap in autumn. And if the AI servers die off in a solar flare the shepherds will know it is time to take their sheep somewhere else and maybe rebuild the AIs one day in the future. If the shepherds die off when a meteor strikes, the servers will be able to clone back the shepherds once the ski clears up. It seems to me this is a scenario just a plausible and it will make humanity much more resilient.
That being said, I don't know what will happen, who knows what will really happen? Just like we didn't know what writing was about to lead us to six thousand years ago, this time we have no idea what thinking sand will lead us to. In any case, God will provide and life will prevail. Stay sharp.