Buon Natale/ 2022 Science Breakthrough of The Year/ The Brief History of AI/ Sam Altman on DALL-E2/ TikTok’s Secret Sauce
Massimo Portincaso
Founder & CEO at Arsenale Bioyards, Industrial Romantic and Antidisciplinarian Stoic
Buon Natale. It is the last Antidisciplinarian of the year, as I will be back in early January while taking a short Christmas and New Year break. It has been surely a very tumultuous year, with the first war on European soil since WWII, an energy crisis, and a gloomy economic outlook, just to mention a few.
Yet, as the year comes to a close, I am feeling more and more optimistic and bullish about the future. One just needs to read this and the past Antidisciplinarian issues, to see that a lot of breakthroughs are happening, all pointing in the right direction.
Clearly, it is not going to be smooth or easy, but, no doubt, we are moving toward a different, hopefully better world. We are just about to undergo a profound and systemic transition, which, like all profound and systemic transitions, will not be painless. But I am confident we are going to come out on the other side of the transition in much better shape. The alternative is extinction… so we’d better make the transition happen.
And I see more and more beacons of hope that this is going to be possible, that the energy transition will be possible, that the shift to a generative paradigm is going to happen, that we will have a more distributed and sustainable economic tissue…
It is not going to be easy, but it is certainly possible. The current surge of Generative AI is actually a very good sign. It does not come without issues (see below) but is a very important cornerstone of the new generative paradigm upon which our society can be built. I am going to come back on this in the new year.
In the meanwhile, I wish you all a very merry Christmas and a happy new year, and in case this does not apply to you, a nice holiday season. I’ll be back in 2023.
Buon Natale a tutti!
In a year of tech news dominated by the rise of generative AI like Stable Diffusion, ChatGPT, and DALLE-2, a decidedly more old school technology has topped Science’s Breakthrough of the Year category. The James Webb Space Telescope “ is the most complex science mission ever put into space and at $10 billion the most expensive.” 344 critical steps “could have doomed the mission had they gone wrong.”
A “tennis court-size sunshield” lies behind the eye of JWST and once it was successfully deployed. “researchers began to find galaxies more distant than any previously documented.” US president Joe Biden unveiled the first images from JWST calling it “miraculous” and “hard to even fathom.”
News items:
Fusion’s big moment may be overrated - and far from the renewable energy silver bullet we’ve been waiting for. According to Dr. Chris Cragg: “I am prepared to bet that a true fusion power station is unlikely to be running before my grandchildren turn 70. After all, it has taken 60-odd years and huge amounts of money to get this far.”
AI is a history of megaflops. petaFLOPs actually - up 1 quadrillion of them. But FLOPS aside, “Research and data [continue] to make progress against the world’s biggest problems.” Max Roser of Our World in Data maps out the recent history of AI to help give more context to where we might be going. Fun fact: the first “AI” system defined by Roser was Theseus, a remote-controlled mouse that was able to find its way out of a labyrinth and could remember its course. built by Claude Shannon.
News items:
DeepMind, a company acquired by Google, has developed a new artificial intelligence system called AlphaCode that can write code for certain tasks. The system is able to write code that is both efficient and easy to read, and it has the potential to revolutionize the way software is developed.
Recommended by LinkedIn
Even Open AI co-founder Sam Altman acknowledges that DALL-2’s impact “is not all positive.” In an interview with MIT Tech Review, Altman says that the explosive launch of GPT-3 and DALLE-2 offers “an important set of lessons for us about what the next decade’s going to be like for AI.” Generative AI like DALL-E 2 “is going to impact the job market for illustrators. The amount one illustrator is able to do will go up by, like, a factor of 10 or 100… At the same time, there’s huge societal benefit, where everybody gets this new superpower.”
One of the “most interesting things” about DALLE-2 “is that this was the first AI that everyone used,” and that can be the starting point for the potential benefits (and detriments) of generative AI - “a very hard societal conversation” that we need to have.
News items:
Alphabet-owned Waymo’s driverless dream took a giant leap forward recently by “expanding the size of its service area in both Phoenix and San Francisco as it seeks to send the message that despite all the recent dour headlines about the future of autonomous vehicles, its robotaxi business is still going strong.”
Those transfixed by the omniscience of TikTok’s recommendation algo may be in for a shock. “Many people say TikTok knows them better than they know themselves. Some users think of the algorithm as a divine force that guides them.” But, in fact, “there’s no truth to the idea that TikTok’s algorithm is more advanced than its peers. From everything we know— TikTok’s own description, leaked documents, studies, and reverse engineering efforts —it’s a standard recommender system of the kind that every major social media platform uses.”
If the secret’s not in the algo, what’s TikTok’s “secret sauce”? TikTok feeds the niche. YouTube and Instagram are relatively unadventurous. “The app is far more successful in converting content consumers into creators, in part because its creator tools are superior and more fun.”
News items:
China has thrown some cold water on the generative AI revolution. As of January 10th, 2023, China will require that AI-generated text, image, voice, and video bear watermarks that identify the products of “deep synthesis” tech.
“Real-World Robocops” just got shut down in San Francisco, “but the technology is out there and it may be just a matter of time before it’s used by local police departments.” Ryan Jenkins, ethical scientist at Cal Poly, believes that “Anytime we have an opportunity to protect police officers from unnecessary harm or more harm than is required for them to accomplish their goal, we usually think that that’s a good goal to accomplish.”
Blue lives matter.
News items:
Stable Diffusion and DALLE-2 don’t just borrow, they steal. Perhaps the creation of a “high tech solarpunk utopia in the Amazon rainforest” isn’t as utopian as it sounds? Particularly if you’re the human artist that got ripped off?
DALL-E is not so great. It requires a lot of iterations to get anything meaningful