What the heck was I thinking? A link dump from. 2024.03.2
I’m going to try to do a better job of capturing my thinking, reading, and writing in a more searchable, and public format. I’m stealing this from one of my favorite thinkers, Cory Doctorow:
This is a ridiculous theory that most sci-fi can be understood through “the magic dick.” These stories are about a young man, lost and afraid, going on a quest to find a magical and magically phallic object. A sword stuck in a stone. A lightsaber. A feather. Or in the case of Dune- giant penisy worms. In short, a lot of these stories are allegories for puberty and masterbation.
And maybe the theory is not that ridiculous. The publishing industry definitely overindexes on cis gender male stories and experiences. Maybe this tracks. This is definitely making me re-think Star Wars. And makes me question what George Lucas was thinking about when he wrote the strange interactions between Luke and Leia. Eww.
Vampire man wants to live forever. Believes AI will kill everyone expect for 20 million people. Those people get to live forever. And I want to understand the structure of this new world. How do these 20 million people eat? Does the AI feed them? What is their relationship with the AI? Is it their slave? Or their master? I also want to know how these 20 million people were selected? Is it because they also chose to follow the AI to design their diet and their sleep schedule like this weirdo?
It seems that so many of the beliefs around AI, and especially AGI are religious. The tenets seem to be, if you give over your life to the system, then the AI will give you eternal life.
It takes nearly 500ml of water to process a handful of prompts on OpenAI ChatGPT. That’s an incredibly inefficient architecture. It seems to make one image in Dall-e-2, their system uses half a liter as well.
My friend Anand asked in our group chat recently, What % of video we are delivered will be AI generated?
That question was in response to this tweet from Marques Brownlee
I responded:
“Less than the percent of video that comes from places like Shutterstock (which is where this demo is drawing a ton of its training data).
It is very cheap architecture for Shutterstock to return a query. Very litter energy is expended.
For OpenAI to render one of these AI videos, if it takes .5L of water to make 1 AI image, and a video is 24 images per second, and the average length of these demo vids is about 20 seconds, then one request takes OpenAI about 240 liters of water.
In today’s water cost that works out to .12 cents per prompt. It probably cost stutterstock a fraction of a fraction of a fraction of a cent to return a search query today.
So today at least gen AI video is a SUPER expensive recommendation algorithm for shutterstock videos. I don’t know what the future holds, but the current architecture (if real-I don’t trust no demos), is not gonna be replacing a ton of video creation. But in the short run, while it’s subsidized, we’ll see some people make a few toy videos.
I am super curious about this deal and the calculus the shutterstock team is doing. If they believe that Sora will make videos cheaper than the $80 they charge per download, then they’re basically helping train their assassin. But, if they think that outcome is less certain, taking OpenAI’s money seems like a brilliant move to sucker a patsy. It’s a really interesting business case.
The open thread for the phd marketing job market has started. Like last year I will try to refrain from commenting (I make no promises.) Already the thread is concerned with what “top” school places their students at some other “top” school. The focus on prestige is fascinating and infuriating. I did a phd at two different schools. I finished at one of these “top” places (Kellogg) and I started at a non-“top” place (FIU). And no one believes me when I say this, but I really learned how to do research at FIU. Moreover, the difference in the day-to-day work is negligible. The biggest difference is in expectations. In terms of job prospects- no one expected anything at FIU. People are happy to get tenure track positions. At Kellogg the expectations do not match reality, and thus, the stress is much higher and endemic.
The police were not defunded. The IRS was. So was USPS. Public service funding is a measure of priorities.
No masters above. No servants below.
I think this is my core personal politic.
I love this dude’s art. He bets on himself. And his vision of Blackness reimagined in pop culture.
Cory Doctorow talks about cutting out the middle man as much as possible. But it seems that Amazon has made this impossible. They are all retail’s middleman now. They are essentially a monopolist is a ton of small categories. And they are part of the inflation problem.
The market is not always dumb. And Amazon is already priced as a monopoly.
Drake is lightskint Kevin Samuels.
Calling it now, Google will shut down Gemini before the end of 2025. It is a garbage product. This is the output for “List the most successful African-American private equity investors.” Click on the links
Buzz feed tried to replace its writers with AI. The market was excited for a couple weeks. Then reality set in. AI is not a great writer.
See also: https://www.buzzfeed.com/aglover/ai-quiz-whats-my-perfect-job
There are two camps that have a battle for the soul of AI, particularly Artifical General Inteligence (AGI). There are Doomers- who believe AGI will kill all humans (see- Skynet), and Accelerationists who believe AGI will save all humanity (see- Star Trek). I think we need a third camp- let’s call ourselves Realists- and our core belief is this, “We are not on the path to AGI. Cars and trucks are not going to drive themselves. We are not going to have machines write scripts or make movies of art without human involvement. Because AI is so much dumber (and more expensive) than either of these two groups believe- or admit.”
Trump is selling $20 Chinese Sneaks for $400. It’s just another scam. And the people buying them are just hoping to flip ie scam someone else. It’s a synecdoche of the whole Trump ethos.
The Suit Guy on Twitter/x had a good breakdown:
Killer Mike is a phenomenal rapper. And a fool.
I think we’re at the tail end of generative AI, not the beginning.
Barbara Pravi’s Voila bangs.
Other than Jolene, I hope Beyonce picks some songs from this list and Yonce-fy’s them for her upcoming country album:
I copped a pair of these Reebok run.r’s for less than $100 on eBay.
Camron seems like a happier person now that he’s embraced life is a team sport.
Betting on yourself wins in the long run.
The new Bird and Magic.
This research finds support for the hypothesis that the racial wealth gap is systemic.
The talented tenth will not save us.
This short story is amazing. I can’t believe they made this into a movie, just to kill it.
Virgil Abloh was the same age I am now when he died. Seeing his old weblogs makes feel two thoughts simultaneously. I should more. I should do less. Those thoughts are clearly contradictory.
Arrogant lady rappers are running hip hop. I love it. I made a playlist last year to honor this fact. Still slaps.
Coleman Hughes is either a patsy or a shill. This review of his book is spot-on. Hat tip.
I just discovered that UGGS are different from Uggs since 1974. Viva La Australian Uggs.
I want this coat. I do not want to pay this cost.
Throw Vince in the trash. And everyone around him. Bin him. Ban him.
“The authors propose that the content of certain sociopolitical ideologies can be shaped by individuals in ways that satisfy their social motivations. This notion was tested in the context of color-blind ideology. Color blindness, when construed as a principle of distributive justice, is an egalitarian stance concerned with reducing discrepancies between groups’ outcomes; as a principle of procedural justice, however, color blindness can function as a legitimizing ideology that entrenches existing inequalities. In Study 1, White people high in antiegalitarian sentiment were found to shift their construal of color blindness from a distributive to a procedural principle when exposed to intergroup threat. In Studies 2, 3A, and 3B, the authors used manipulations and a measure of threat to show that antiegalitarian White people endorse color blindness to legitimize the racial status quo. In Study 3B, participants’ endorsement of color-blind ideology was mediated by increases in their preference for equal treatment (i.e., procedural justice) as a response to threat. In the Discussion section, the authors examine implications of the present perspective for understanding the manner in which individuals’ compete over the meaning of crucial ideologies.”
This paper fits into my white blindness hypothesis.
Vouchers seem to be a regressive tax. A way for rich parents to steal from the public school system.
“Value capture occurs when an agent’s values are rich and subtle; they enter a social environment that presents simplified — typically quantified — versions of those val- ues; and those simplified articulations come to dominate their practical reasoning. Examples include becoming motivated by FitBit’s step counts, Twitter Likes and Re- tweets, citation rates, ranked lists of best schools, and Grade Point Averages. We are vulnerable to value capture because of the competitive advantage that such crisp and clear expressions of value have in our private reasoning and our public justification. There is, however, a price. In value capture, we take a central component of our au- tonomy — our ongoing deliberation over the exact articulation of our values — and we outsource it. And the metrics to which we outsource usually engineered for the interests of some external force, like a large-scale institution’s interest in cross-con- textual comprehensibility and quick aggregability. That outsourcing cuts off one of the key benefits to personal deliberation. In value capture, we no longer adjust our values and their articulations in light of own rich experience of the world. Our values should often be carefully tailored to our particular selves or our small-scale commu- nities, but in value capture, we buy our values off the rack. In some cases — like de- creasing CO2 emissions — the costs of non-tailored values are outweighed by the ben- efit of precise collective coordination. In other cases, like in our aesthetic lives, they are not. This suggests that we should want different values suited to different scales. We should want value federalism. Some values are perhaps best pursued at the larg- est-scale level, others at smaller scales. The problem occurs when we exhibit an ex- cess preference for the largest-scale values — when we consistently let the universal metrics swamp our quieter interests.”
A beautiful philosophy paper on the unintended consequence of rankings.
The IMF seems to overestimate the impact of Gen-AI. Will be a good cite however For the AI and Trust project.
I absolutely love this paper. They estimate which job tasks. can we done by AI according to employees themselves. We are building on this methodology in our AI and trust project.
https://dspace.mit.edu/bitstream/handle/1721.1/120302/pandp.20181019.pdf?sequence=1&isAllowed=y
This paper seems to suggest that automation leads to wealth inequality.
https://dspace.mit.edu/bitstream/handle/1721.1/120302/pandp.20181019.pdf?sequence=1&isAllowed=y
All romcoms seem to be about rich people. Even my favorite romcom, Boomerang, was about the richest Black people Eddie Murphy could imagine. Same with coming to America. Are there any romcoms about poor people? Or can poor folks not be in love and laugh?
https://www.digitalspy.com/movies/a867107/rom-coms-diversity-wealth-income/