Tangent #23 - The AI Rat Race
Plus, a Coca-Cola ad, essays on books and envy, and a groovy song for your weekend.
Hello and welcome to the first Substack edition of Tangent and welcome to new subscribers! š. (Iām sorry Convertkit, itās not you, itās me).
Being the first Substack post, this one is a bit longer than my usual editions. Letās dig in!
Iāve spent most of my week in bed as my body attempts the Guinness World Record for the most consecutive colds a body can experience in a month. Between my coughing bouts, Iāve had a plenty of time to think about a couple of fears that were frequent visitors in my mind this week.
Iāll discuss one of them this week, and the other one next week.
The AI rat race š šāāļø
As part of my day job at Google, Iāve gained basic proficiency around AI and ML. What ML models are, how to train them, what computing hardware to use, terms like transformer models, neural networks, etc.
Itās been fascinating to see how the industry has advanced over the past 4-5 years. In the last 6 months, weāve taken a hyperspace jump akin to an escape scene from The Mandalorian.
This week was overwhelming. Announcements galore. And not just any announcementsāthese were āholy $&!#! this is wild!ā announcements:
Midjourney can now model hands (with five fingers!).
Google Workspace and Office 365 can now be verbose on your behalf.
GTP-4 can now pass the Bar exam, but it still cannot write an AP English essay explaining why Watership Down is a great book (to its defense, no one can).
The narrative is that all this progress is good for humanity. Youād be insane to criticize the progress made in AI! We can now detect tumors better and cheaper, you can measure the carbon capture of forests, and companies can conserve massive amounts of energy and reduce waste thanks to AI.
In our daily lives, you can now craft better emails if you are a non-native English speaker, you can write into life any image you dream of, and you can even use AI to āchatā with your favorite podcast.
Something doesnāt sit right with me about all of this.
Itās been covered ad nauseam how our silly mammal brains are simply not wired for most of the ways we live life today. Spending so much time looking at our screens, drowning in content overload, communicating with far more people than our ancestors did, etc. We know itās all too much. Patience and moderation now come at a premium.
Yet, the world is being taken for a ride in the battle of egos (mostly male). See the language thatās popping up: āThe arms race for AI supremacyā āMicrosoft and Google trading blows.ā These are all conflict metaphors. The media eggs on this conflict (it gets eyeballs), and we start buying into the narrative that we are in a race, that there must be a clear winner, and that speed is the primary metric.
Less so is said about the stakes, the negative externalities, what āwinningā actually looks like, and whether we are treating an infinite problem like a finite game.1
Technology is exciting. Technology is helpful. Technology can be good. But, weāve opened Pandoraās box and no one in this planet quite knows what this means for our civilization.
And thatās what I fear. We are all being taken for a ride and not enough questions are being asked by powerful people. For instance, how is it ok that OpenAI, a company which started as a non-profit, all of a sudden transitioned to a ācappedā for-profit and was essentially acquired by Microsoft? If you look at the release notes of Chat GPT-4, it is anything but open.
There are many other fundamental unanswered questions around bias in training data, attribution of copyrighted material used in training data sets without consent, and the incentives of profit-generating entities in providing broad access to these technologies (and the tradeoffs).
I read Raika Sarkettās latest Substack post, which asked this question:
How might we balance the speed + time AI (and new tech) give us with the importance of time spent wrestling with ambiguity, mastering a craft and working with people vs machines?
These are the types of questions we need to be asking.
To techno-optimists2, asking these questions will feel like a bureaucratic attempt to slow down progress. They will call you a luddite, curmudgeon, or worse, an EU parliament delegate.
Would slowing down impact progress? Perhaps. The larger question is what is progress for? How do we measure progress? Is progress strictly bottomline growth of a few companies? That would be a myopic way to see the world.
And we have recent examples of a rat race gone wrong. The last 20 years saw a rat race for social media platform dominance. At first, it all seemed cool: You could talk to your friends all over the world, you could connect with strangers over common interests, you could add moody photo filters to your bathroom selfie. Fantastic!
As we now know, all the goodness of social media has dark side effects: Itās made us phone addicts, turning our smartphone into an extra limb. Itās wrecking younger generations, being a contributing factor in opioid use, suicidal behavior, and other mental health issues.3
I canāt help but to think we are going to have similar negative externalities with the AI rat race we are in. That doesnāt seem like progress to me.
I wish I had an optimistic, rousing, āwe shall overcomeā-type speech to end this with. The best I can do is offer you my imperfect method for dealing with existential dread. Accepting the duality of conflicting ideas (which create tension), and understand my agency against the fear.
Accepting the duality: Advances in AI will bring harm, but they will also create a lot of novel and interesting things and save lives. These two happen simultaneously.
Accepting my agency against the fear: I have both the responsibility and the privilege to become well versed in these technologies (or try to keep up). It is up to me to continue challenging the myopic pursuit of speed over temperance, of conquering over collaborating, of advancing humanity at the risk of losing it.
Worst case, my conscience will merge with ChatGPT-8 and I will never have a cold again.
Media Worth Consuming
Coca-Cola released a stunning, art-inspired commercial this week. It features art work from Warhol, Hiroshige, Picasso, among others. If you are an art lover, you will be delighted by this ad.
A Pilgrimage for Book People: Charlie Becker wrote a riveting and personal essay on a massive book sale event in Chicago in the 90s. But beyond this story, lies a dedication to books, those who love books, and the marvelous ways in which words can connect us and create rituals.
beauty: Elaine writes a delightful short piece on how a visit to Frida Kahloās Casa Azul helped her challenges the idea of envy and how we see ourselves.
Photo(s) of the Week
I do this thing sometimes where I try to align things into yellow-blue-red after the Colombian flag. Itās my subversive way of preserving my Colombian identity. I love these pictures, taken at the Milan Triennale Exhibit (a boon of inspiration for me last year).
I chose these photos because Iāll be traveling to my homeland for the first time since January 2020 for a close friendsā wedding in beautiful Santa Marta. Iām also starting Steven Fosterās Photography for Creatives course, so Iāve been revisiting old photos that bring me joy.
Song of the Week
La Primera Vez (The First Time) by Manuel Medrano & Juliana: Iāve been obsessing over this song by Colombian artists Manuel Medrano & Juliana. Manuel has some of the smoothest R&B vocals Iāve heard, paired with Julianaās velvety vocals, this song is a groovy, sensual throwback that makes me wonder whether I was born in the wrong era (probably).
This song is part of my latest favorite Netflix show La Primera Vez (Eva Lasting in English). Someone referred to it as a Latin 500 Days of Summer, and Iād have to agree.
The show takes place BogotĆ” in 1976 and follows the story of a young man named Camilo (not joking), who falls in love with the first girl admitted to his all-boys school (Eva). The screenwriter clearly saw high school Camilo in action and used it as inspiration to make a show set in 1970sā Colombia. If you want to get an idea of how I was in high school, Iād say this is 70% me (minus the fighting. Iāve always been a lover, after all).
Until next time!
Infinite Games is an idea popularized by Simon Sinek which asserts that organizations should pursue infinite games, which are not played to win. Playing infinite games emphasizes the importance of collaboration, long-term thinking, and taking the long view in order to create lasting success. Whereas finite games are short-sighted, winner-take-all endeavors that do not create sustained value over the long-term.
I am not referring to a group of people that are very optimistic about techno music.
Americaās Teenage Girls Are not Okay by Derek Thompson from The Atlantic addresses the nuances of social media being the primary cause for declining mental health. However, I personally believe it is a significant cause, though not the only one.
Tangent #23 - The AI Rat Race
Love this piece. It aligns with a lot of my thinking with AI. I mostly had been optimistic about it as āAI-assisted creativity,ā but the latest evolutions are going in more aggressive directions. More than anything I resonate with the question you shared about the importance of ambiguity in creative process. My latest āsketchplorationsā post made me think about how much I enjoy that phase of it, and our feedback sesh (thx again!) is another example of all these aspects that are really important. AI skips over that stuff.
All depends on how ppl use it, but the tool often dictates that for majority of ppl, which is the scary part. I feel many folks will simply lose that practice, skill, patienceā¦
"Latin 500 Days of Summer" !!
Say no more. Sold.