With the rise of ChatGPT's capabilities and the hype building on social media, its critics (read: "AI doomers") have become louder too.
For the discussion to stay balanced, between utopian AGI communism and total machine-fascist destruction, while before AI's bull run, I had seldomly heard of the AI alignment community leaders, although names like Eliezer Yudkowsky and Timnit Gebru always rang a bell, they usually didn't stay available too long in memory.
But, it must have been with OpenAI's improving marketing talent that those side characters started coming to our rescue, shielding us from imminent death and chaos - with arguments!
Figureheads of a domain that claims to confidently forecast the future of human progress, not through experimentation or data, but through conviction and being less wrong. And sure enough, the world decided to amplify their voices as to bring nuance to a currently unhinged techno optimism.
There is, however, an incredible value and power in controlling both sides of a story and even more, in having the audience "bear with you for just a moment."
That is, what I think, the secret sauce that make AI doomerism and OpenAI's marketing work so well together, and a strategy which I want to unveil today.
Jason Calacanis is a smart dude, in case you had doubts. In a video uploaded 13 years ago to Youtube, and titled "Private tour of tesla model s," with a young Elon Musk giving details on the car, Jcal is supposedly behind the camera, vlogging the tour and asking an un-identified bystander if "they ordered one yet," with him re-affirming that "he got two."
And although it is seducing to claim that Jason surely was just lucky, I think he's someone that truly understood how to leverage others' emotions for reach on social media.
In the culmination and collective confusion that was the COVID-19 pandemic and the on-going climate catastrophe, Tesla stock, the original hedge against the environment-destroying car industry, reached its all time high with Greta Thunberg giving speeches in front of influential industrialists.
It is this neatly set up game-theoretic riddle of wanting to colonize Mars, while also wanting to save Earth from imminent destruction, that catapulted Musk into the social media's stratosphere. A hedged bet of either escaping Earth on a self-built rocket to a virgin planet or saving planet A with electric vehicles. A marketing strategy so grand it's impossible to ignore.
And still, while already this set up is fascinating enough, it isn't what truly fueled Musk's success. Although very impressive and much credited, Musk's secret sauce isn't engineering or empire-building. It is, what always hid in plain sight, namely the accompanied threat of death, the anxiety of drying out or drowning - Tesla's and SpaceX's most valuable asset is a hyperobject itself: The climate change crisis.
The early recordings of YCombinator's start up school, with Paul Graham talking about Airbnb's journey to product market fit, are a master class in story telling.
They galvanize us because, seemingly, Paul weaves a story from something repetitive and tedious as starting a company, as someone truly inspiring and exciting. If you've ever seen Justin Kan's YT channel intro, you'd know what I mean, as he's essentially using blog post screenshots like Polaroid pictures of friends to symbolize important moments in his founder story.
But knowledge work mostly happens in our heads, on paper and on computers. Seldomly, it is so exciting that it can compete with more tangible stories, like two people falling in love or someone making it in a foreign country.
Still, Graham's story about Airbnb is engaging, with a key myth constantly being repeated, namely that of the investors' confusion after Chesky's pitch presentation and their thoughts racing around the doubt if Airbnb guests will be comfortable with staying at strangers' apartments. Paul and others say this openly, too, that "it's such a stupid idea at first, moving into the place of a total stranger. After all, they could be a serial killer."
However, I think it is this very controlling of both sides of the story that makes Airbnb's value proposition, because few of us truly ever stayed in a real person's apartment.
Personally, while I have had these experiences, most places intentionally looked "home-y," but were indistinguishable from regular bed and breakfasts booked through other sites like Booking.com.
But Brian Chesky and investors knew that hotel chains were a rather recent invention, as Garry Tan admits. And besides, on my trip to Colombia and on another one to the south of Italy, I always ended up getting the impression that Airbnb is really just a clever re-branding of BnBs (the name should have made that obvious sooner, I know).
And still, this reveals the importance of cleverly controlling both the risk and reward side of a marketing proposition. That is because simply selling someone the idea of staying at a friend's place, an ocean away, wouldn't sell. It'd come with a subconscious thread of anxiety and uncertainty. And it is, hence, the explicit marketing of the existential risk of being cut into pieces and fed to the fish that makes Airbnb's marketing work and original.
It's because, through Airbnb, you might stay at a stranger's place, get to know them, and leave as a friend. Still, you are taking the existential risk (the one with the fish), but we all know that it's incredible ridiculous to catastrophize!
Recently, Jason posted a video of an accidental alpine rescue that made the rounds on social media. It shows a person skiing on an unpaved slope and the ensuing coincidental rescue of a snowboarder stuck immovably in a tree well.
The video isn't for the faith-hearted, and if you're an avid skier or snowboarder, maybe it's wise not to even watch it. Still, to drive my point, and having been involuntarily exposed to it, I'm linking it below and I'll be discussing it in the following sections.
Most amazing video I’ve seen in a long time… snowboarder saved by a skier — seems both were skiing alone in the deep powder. Always ski with a partner or two! pic.twitter.com/8ybvr8u56F— @jason (@Jason) April 3, 2023
Assuming you've now seen the video, I'm guessing you also experienced the same feelings I had. A gut-wrenching sensation and anxiety, fear for life, fear for my family and friends' lifes. Being forgotten and left to suffocate.
As a skier myself, my thoughts immediately went to my brother and father, who had just recently sent me photos of their trip in the Austrian Alps. I felt terrified and very uncomfortable.
In reaction to it, I hence started reading the replies to the post and oh!, what do I conveniently see there? A woman advertising a neat wearable device for "outdoor adventurers," quasi-made to avoid this dreadful situation. Dear Mrs Zenner shut up and take my money!
The skier is a hero! Thankful both are ok. 🙏@Jason agree to partnering up, + adventurers need Milo Action Communicators! “Go far stay close.” Stay connected w/your group! No Wi-Fi needed. Hands free. Walkie-talkie 2.0!— Joyce M. Zenner (@CompellaSearch) April 3, 2023
And so it is once again, the control and intentional exposure to fear and its potential resolving agent, and hence, the ultimate freedom of shredding fresh powder but also the lonely tree-well-suffocation, that make me want to expose my family to the dreadful experience of watching this video, just to teach them about the dangers of tree wells, and to entrust them with the task of figuring out better safety practices for their next trip.
Now, by no means do I want to claim that Jcal is promoting his investments here: But I found it awfully convenient that this device was advertised in his replies.
And I also want to say, having skied on potentially unsafe slopes as a kid, teenager and adult, with my parents or friends, that, as with any risky hobby or activity, catastrophizing usually doesn't serve a too useful function.
Yes, it highlights the problem of getting caught in tree wells and educates skiers. But beyond creating awareness, up-keeping basic safety practices is anyways vital to do during this sport off-slopes.
You don't end up alone in a tree well because you ski together. And still, alone or together, this scene will now forever be burnt into my memory, surely available to wet my palms the next time I pass a tree on the way down.
In an alternative version of this essay, the main point I had been driving was to convince you that Sam Altman and Eliezer Yudkowsky play on the same team.
See, Eliezer making the media tour to proclaim that AI doom is inevitable serves OpenAI greatly, in that to refut him, even his most stubborn critics will have to "at least for a moment" put themselves into his shoes and accept the possibility of a universal AI-god's existence.
It is then playing into OpenAI's hands, as were they to convince us of their chatbot ushering in an era of abundance with no strings attached, that'd simply come across silly.
So it is as with Airbnb's serial killer risk, Tesla and SpaceX's looming climate catastrophe and Jason's tree-well doom, that having the consumer face their existential fear over the product's potential and in the process, leading them to imagine the unlimited potential of the dooms-day scenario, has a similar effect as hypnosis.
Any reasonable human, faced with a potentially threatening existential risk, reasonably assumes its correctness for just a moment to think it through empathically. We can't think of AI doom without maximizing AI's future capabilities. But my argument is that this is the evil trick, as what remains in our heads is that burnt-in image of the risk, which is a derivative of us deliberately unleashing the horse men in our minds. And so it happens that we give small risks with terrible outcomes a lot of attention.
In the case of OpenAI's ChatGPT, it really is just a website creating plausible-sounding text responses. Let's not forget that. Yet, easily by listening to Eliezer's gut-wrenching stories about the apocalypse, it makes ChatGPT seems so utterly important that we must not, under no circumstances, ignore it. We shall try it and observe its mind.
But I'm not here to say that Eliezer is a paid shill of Sam's OpenAI. I have zero evidence of that. But he sure does a great job convincing us all that modern AI is a really big deal and that we should all pay attention to what is happening.
So I can't help but wonder if it isn't the same doomerism marketing playbook of controlling both sides of a story that now recently also Balaji Srinivasan has started using for scaring people about de-dollarization and in an effort to pump his Bitcoin bags.
published on 2023-04-08 by timdaub