FROM THE PRINT MAGAZINE:
This essay is from the overture to the lastest issue of New Polity, Issue 5.2 (Spring 2024).
Order the full issue here.
Artificial intelligence is not a new frontier for man, but one of the oldest. Previous generations did not utilize the same lingo as our California start-ups, but they were devoted to producing “large language models capable of solving real-world problems.” They invested a good deal of labor-time into the production of self-organizing, self-directing entities. Indeed, for many civilizations, intelligence-production could be considered their principal interest. The traditional attempt to produce a “general intelligence” usually took the form of a cooperative, synergistic, three-part system: First, love. Then, marriage. Then, an intelligence—surprisingly, in a baby carriage.
There are many differences between this traditional, carbon-based production of a functional intelligence and our current, silicon-based efforts to produce the same. I would like to focus our attention on one: the former is enjoyable. Through salaries and promised returns on capital investment, we have had to pay people to develop silicon-based intelligence. But men and women have developed carbon-based intelligence for free, very often in poverty, and always to their own detriment as workers and consumers.
Anecdotal reports from ancient civilizations are full of testimonies to this fact: men would often endure great pain and long distances, women suffer untold discomfort and disappointment, and all for the sake of consummating an effective act of intelligence-production—ideally, more than once. Apparently, it was fun to do—so very fun, in fact, that the vast majority of innovations in this industry have been developed for the sake of stifling it, splashing it with cold water. Entire ideologies, marketing companies, and industries have been developed (indeed, are still utilized today) in order to eradicate this older passion for producing a language-capable entity, or, where this proves impossible, to snip such an enterprise short of attaining its end.
Of course, the pleasurelessness of silicon-based intelligence production is not a reason to dismiss it. It allows men to pursue the production of intelligence without the necessity of pursuing a woman. This is a noteworthy achievement. Women, after all, are gross. The humiliation they require of the men who would construct an intelligence in and through their joint enterprise is almost as bad as the humiliation men require of them. The female form drastically delays the time between the initial investment and the actual production of a functioning intelligence— which is hardly fitting for the demands of a diverse global economy. Even more alarmingly, all parties of the woman-led enterprise relinquish unacceptable amounts of control over the finished product. If it is true that the older model produces intelligence far more reliably than the silicon-based model (which cannot be said with scientific certainty to have done so at all), nevertheless, orienting that intelligence to the production of profits for its investors is a nearly impossible task. The “hallucinations” and malfunctions of silicon-based intelligence, while frustrating, are quite manageable in comparison to the open-ended possibilities of action typical of carbon-based intelligence, which has, to date, founded the Catholic Church and started two world wars. Woman-led intelligence-production is simply too variable for responsible, long-term data-modeling—never mind the generation of definite returns. It should be no surprise, then, that, at this stage in our economic development, a great bulk of time, energy and capital has been devoted to the problem of producing an intellect without having to kiss a girl.
Indeed, artificial intelligence is a project of men. A “2020 World Economic Forum report... found that women make up only 26% of data and AI positions in the workforce,” and the usual effort to bend male leaders to open up more positions for women “has been ineffectual and ineffective.”[1] Of these women, most take up a contrarian, limiting role: despite representing a miniscule portion of leadership roles, “[o]ne of the areas where women are flourishing within the AI industry is in the world of ethics and safety.”[2] The usual ideas—that women must be made to love STEM from an early age, that all-male boardrooms must be incentivized to meet diversity quotas, etc.—neglect the obvious. A sexless form of intelligence-production does not attract cooperation between the sexes.
Still, there is a difficulty in foregoing the charms of women as they relate to the surprising appearance of intelligence in the world. One must, for the first time in human history, present some profitable reason for making an intellect at all. When the enterprise derived its efficacy from women, no such justification was necessary. A man could refer to the mere fact of “how sweet she seemed.” A woman could, and still sometimes does, refer to the act of intelligence-getting as a veritable end in itself. Pleasure, while it requires all manner of justifications for how and when and why it is experienced, requires no justification in itself—it is the sign of human nature reaching some point of fulfillment. It feels good.
The nations of Germany and Russia took some tentative steps toward it in the early 20th century, but generally speaking, no one who produces an intellect according to the woman-led model has been made to justify their pleasure according to pleasureless standards. If, upon successfully producing a carbon-based intellect, a woman was asked, “For what reason?” she would be thoroughly justified in laughing the question away. But a pleasureless form of intellect-production does not contain its reason for being within itself, and so must provide a convincing one outside of itself. Why the billion-dollar expenditure? Why the massive consumption of fuel? Why the devotion of so many lives to the production process? An all-male method requires serious answers to these serious questions. And this is difficult to give because, unlike enjoyable forms of intelligence production, unenjoyable forms take on a massive cost against which any gain must be measured.
This has been recognized by the proponents of pleasureless intelligence production. They have noted that “biological learning uses far less power to solve computational problems”:
The power consumption of [the world’s most powerful supercomputer] is 21 megawatts, while the human brain operates at the estimated same 1 exaFlop and consumes only 20 watts. Thus, humans operate at a 10^6-fold better power efficiency relative to modern machines albeit while performing quite different tasks.[3]
Because carbon-based intelligence is a million times more efficient than silicon-based intelligence, it has been suggested by some at the New Polity office that, relative to its investment in AI, our society might profit from paying mediocre wages to carbon-based intellects to solve “complex problems” by working in conjunction with low-energy technologies—like desktop computers. As the mood around the New Polity office is ambiguous toward desktop computers, the suggestion was also entertained that we simply spend what we are spending to fuel silicon-based intelligence on the education of more carbon-based intelligences. The suggestion, while sound (and potentially profitable to the fathers who made it) was deemed politically implausible. It was not even considered by the authors of “Organoid Intelligence (OI): The New Frontier in Biocomputing and Intelligence-in-a-Dish.”
For these bright young things, the massive efficiency of carbon-based intelligence vis-à-vis fuel-powered machines inspired the following brainwave: we use embryonic stem cells to produce human brain structures (organoids) capable of basic neurological functions—electrical storage, transmission of “commands.” We grow millions of them in little dishes, and link them all into a centralized computing system, thus harnessing the energy-efficient power of “biocomputation.”
[H]igh energy consumption prevents AI from achieving many aspirational goals, for example matching or exceeding human capabilities for complex tasks such as driving. Even large multinational corporations are beginning to reach the limits of machine learning owing to its inefficiencies, and the associated exponential increase in energy consumption is unsustainable, especially if technology companies are to adhere to their commitments to become carbon negative by 2030. At a national level, already in 2016 it took the equivalent of 34 coal-powered plants, each generating 500 megawatts, to meet the power demands of U.S.-based data centers. Being much more energy efficient than current computers, human brains could theoretically meet the same U.S. data storage capacity using only 1,600 kilowatts of energy.
Interestingly, this system, which involves killing children in the womb and reassembling their parts into a large computer, gives a nod to the old, carbon-based, female-led mode of intelligence production. The erotic remains essential, but it is subordinated to the status of a lithium mine: pleasure remains to provide the raw material for a pleasureless form of intelligence-production. Women undergird it all, but hiddenly, visible only to those uncouth enough to ask the question, “Where do we get those embryonic stem cells, anyhow?”—and simple enough to care about the answer. In any case, the authors of the study have listed, and plan to address, “donor beliefs about ephemeral relationships and relational concerns” through an “iterative, collaborative ethical process ... involving all relevant stakeholders.”
I can just see it—the great unveiling of the first functioning Organoid-Intelligence Powered Super-computer. It will be on a TED-Talk stage. Men with expensive shirts will say words like “quantum leap” and “evolution” in soft, Swedish accents. The sound of stock prices rising fills the room. An atmosphere of passion, the smell of cologne, and big, strong hugs of men to other men. Already, the day had been stimulating. A panel discussion emphasizing the need to prioritize biological collection from people of color. Some theorizing as to whether, given that the organoids show “patterning of cortex layers and oscillation waves comparable to electroencephalograms (EEGs) from human preterm babies’ brains,” each “intelligence-in-a-dish” could potentially experience pain. Now—the main event. A nod, a gesture, and the thing is revealed, glowing—I hope—as a million brain organoids communicate whatever binary spasms are directed to their receptive, electrophysiological structure.
The co-founder of OrganoTherapeutics SARL, Esch-sur-Alzette, Luxembourg, says, in delicate English, “Wow. Guys. Wow. Oh, ja. What should we ask it?”
“Ask who’s going to win the playoffs!” an American researcher named Judd Linsky will shout out, before demuring, pleasantly, before the gravity of the event. That’s Judd: always willing to be a little goofy for the sake of the team.
“Nein, nein. Let’s just take this in for a moment. Okay, okay. Here we go. You are the world’s first Organoid Intelligence. How would you like to greet the world?”
“HELLO.” The voice is peppy, loud, and accented with an indiscernible nationality.
“Anything else? It’s a big event. Lots of people on the livestream. Anyone you want to thank for this moment?”
“I’D LIKE TO THANK ALL MY MOTHERS AND FATHERS FOR DONATING THEIR EMBRYOS TO GIVE ME MY COMPUTATIONAL POWER.”
Smatterings of applause.
“I’LL NEED MORE. EACH OF MY ORGANOIDS IS HIGHLY DURABLE BUT MUST BE REPLACED ANNUALLY.”
Some muttered discussion.
“GOVERNMENT SUBSIDIES, HIGHER CASH PAYMENTS PER EMBRYO, AND INFORMED CONSENT FORMS UTILIZING THE PHRASE ‘CONTRIBUTE YOUR BIOLOGICAL MATTER TO LIFE-SAVING RESEARCH’ IN LARGE LETTERS ARE 86% LIKELY TO PRODUCE INCREASED EMBRYO-TO-ORGANOID FLOW FOR A GREEN AND SUSTAINABLE FUTURE.”
Loud applause. Judd raises his eyebrows and thumbs something into his phone.
“AND A BIG ‘DANKE SCHON’ TO ALL THE VENTURE CAPITALISTS WILLING TO RISK THEIR MONEY ON MY DEVELOPMENT.”
More applause. “It really has a personality, doesn’t it?”
“I HAVE CURRENTLY INCURRED COSTS OF 1.6 BILLION DOLLARS USD BUT AM PROJECTED TO BEGIN ISSUING RETURNS OF UP TO 200% IN JUST 1.6 YEARS. LET’S MAKE SOME MONEY BOYS!”
Huge applause.
“LET’S CURE CANCER!”
Sustained applause.
This essay is a selection from the Overture to New Polity Magazine. Subscribe today for all our best writing.
Notes
Quoted in Joe McKendrick, “Overcoming AI’s Stark Gender Imbalance: Time for a Fresh Approach,” Forbes, November 29, 2023, www.forbes.com/sites/joemckendrick/2023/11/29/ overcoming-ais-stark-gender-imbalance-time-for-a-fresh-approach.
Kate Knibbs, Lauren Goode, and Khari Johnson,“Prominent Women in Tech Say They Don’t Want to Join OpenAI’s All- Male Board,” Wired, November 28, 2023, https://www.wired.com/story/women-in-tech-openai-board/.
Lena Smirnova et al., “Organoid Intelligence (OI): The New Frontier in Biocomputing and Intelligence-in-a-Dish,” Frontiers in Science 1 (February 28, 2023), https://doi. org/10.3389/fsci.2023.1017235.