Linklist: July 16, 2024
recommended: all except 2, 3, 6
links (links with excerpts below):
California AI bill becomes a lightning rod—for safety advocates and developers alike
Unintended consequences of changes to Open Access and the Impact Factor
A scientist is likely to win Mexico’s presidency. Not all researchers are rejoicing
The importance of distinguishing climate science from climate activism
On Being a Science Writer and Managing a Mental Illness (2017)
Nuclear energy could power the AI boom—but only if proliferation risks are minimized
links with excerpts:
"It’s okay to ramble into Myriad Intimacies, Mani tells us. She presents the book as a mandala or yantra, a diagram shaped by Hindu-Buddhist tantric traditions and possessed of multiple entrances. Readers who begin at the beginning will find acknowledgments, followed by a poem, followed by an introductory essay. The poem speaks for itself, balancing in its lines the facets of the collection as a whole: “Analytical Contemplative.” The lengthy gloss of the subsequent essay is also vital. It clears a path stripped of old habits of critique. What would it mean, Mani asks, to “cede” rather than “capture,” to make an “offering” instead of an “intervention,” to embody the “modest explorer” and not the “omnipotent researcher”?
This is a tough shift to make. Those of us who are scholars have been trained to cultivate mastery over our material, to calibrate our value in terms of the sharpness of our arguments, and to approach social problems of colonialism, labor exploitation, misogyny, racism, and war, which elicit our full and embodied selves, as abstract problems for the mind. Tantra enables this shift, Mani suggests, by offering “post hoc compilations of wisdom yielded by practice,” by emphasizing meditation rather than book-learning. There are multiple openings here, making it possible to behold the sentience of the universe; the coequality of all life forms; conceptions of triadic intelligence (body, heart, mind); and the triangulation of self with the other and with the divine. For the incredulous reader, Mani substitutes “laws of creation” with “laws of nature,” enlisting ecocritical, ecopoetic, and environmental humanities-based conceptions of interspecies co-dependence, planetary consciousness, and more-than-human relationality that incorporate (without always acknowledging) what Indigenous epistemologies and vernacular spiritualities have long known."
"At its core, SB 1047 aims to hold developers of large-scale AI models legally accountable for providing reasonable assurances that their models do not have hazardous capabilities that could cause critical harms. The bill’s definition of critical harms includes the creation of chemical, biological, radiological, or nuclear weapons that could lead to mass casualties; cyberattacks on critical infrastructure that cause at least $500,000,000 in damages; or actions by an autonomous AI model that cause the same level of damages, harm humans, or result in theft, property damage, or other threats to public safety. To avoid such harms, developers must be able to fully shut down their models, effectively building in kill switches.
The proposed legislation applies to AI models “trained using a quantity of computing power greater than 10^26 integer or floating-point operations [FLOPS], and the cost of that quantity of computing power would exceed one hundred million dollars ($100,000,000).” In other words, the bill primarily targets the largest future AI models (likely larger than the majority of even today’s advanced models) built by companies that can afford some of the costliest model training. It uses the same metric for training computing power (10^26 FLOPS) referenced in the Biden Administration’s AI Executive Order to delineate models covered by the bill. Under the law, unless a model receives an exemption, its developers would need to submit an annual certification of model compliance, and to report AI safety incidents involving their models, to a newly created Frontier Model Division within California’s Department of Technology."
"While the movement to OA has allowed progress to be made on breaching paywalls, it has triggered some unintended consequences. The new and evolving financial models threaten many journals from learned societies (Johnson and Malcolmson, 2024). That is a negative outcome as disciplines need to support their specialist areas and the best peer reviewing often comes from scientists committed to a society of like-minded researchers. However, a greater threat to the scientific endeavor has come from pop-up publishers who identified a new business opportunity from the combination of increasing OA and the decreasing relevance of IFs. Without IF as a guide, the name of the journal becomes largely irrelevant. In the absence of the Journal IF, an author-paid paper in a journal with low-quality standards is as visible and has the same legitimacy as one in a journal driven by high standards for its content. Moreover, these opportunistic publishers increase their profits by higher acceptance rates of submitted papers and hence a lower quality bar for publication. Weekly invitations to submit papers or join the editorial board of journals with random and often irrelevant titles are the signs of a whole segment of the industry that will accept any submission, with very minimal peer review or no review at all, provided that the author pays for it. A similar industry has grown in the area of scientific meetings where CVs can be readily bolstered by accepting invitations to be a keynote speaker at a self-proclaimed world-leading conference.
The risk of this development is that the scientific literature becomes polluted by papers that would not survive analysis in a journal club. It then becomes difficult for those outside the system, who look to “the science” for guidance on policies or treatment, to know where truth is. Contradictions and corrections are central to the scientific process and to increase knowledge. However, if papers can be published without the stamp of peer review, this scientific debate culture becomes a convenient cloak for those who use science to disrupt. The steps towards science becoming opinion-based rather than factual and shifting from objective data to a social construct are facilitated by the unhindered publication of shoddy work merely because the invoice has been paid."
"Even more so than the explosion of mass media and print culture in the late nineteenth and twentieth centuries, the internet today trades in ephemerality. And so how exactly is fiction—a kind of writing that aims at its very core to tamp down and preserve a particular place, a time, a mood—supposed to approach this?
When faced with a variation of this question amid the rapid technological shifts of his own era, E. M. Forster cried out “Only connect!” in the pages of Howards End. A renewed focus on human subjectivity—the ghost in the machine—offered a path forward, and it still does. For all of the namechecks and memes replicated in the first half of Lockwood’s No One Is Talking About This, the book found its footing in the second half, when the tragedy of a child’s mortality—real life—interrupts the endless scrolling. And then there are the books that have used the development of the internet as a chance to explore the muck of human existence, such as the Ecuadorian writer Monica Ojeda’s novels Jawbone and Nefando, with their repellent explorations of child abuse and the Dark Web. Or the French writer Delphine de Vigan’s Kids Run the Show, on child influencers, that plays with the internet’s favorite genre—true crime.
“Internet writing” has become a category broad enough to mean essentially nothing because the internet is a technology in the same way that a book is a technology. There is an immersive quality to the internet, the Wikipedia rabbit holes and the endless link trees, that the affectless writing that has become the house style of online life fails to capture. We turn the page; we scroll. An endless deluge of information, our existences online are now defined by a complex tangle of memes and references and rhetorical quirks. But it’s in the sorting of the information, the understanding of how this information gets filtered into the very structures of our language, that the art lies."
"Among the heights, Shepherd gave lyrical weight to a widespread intuition: that being in the mountains precipitates awakening. It is common for walkers and climbers to report experiencing reveries in high places. People talk about the radiance of the light, of pregnant silences and quickening skies. Forever prone to anthropomorphizing the mountain, we refer to its brow, shoulders, flanks and feet. Though it is absurd to accord sentience to a pile of rock, we often perceive mountains as mercurial, places of moods. At times they are beneficent, river-bearers, all-seeing guardians; at others, they transmute into implacable godheads conjuring wrathful weather.
This idea that the mountain is a repository of esoteric energy tends to be central to the various religious doctrines that venerate them. Recently, I spoke to Tim Bunting, an adherent of Shugendō, an ancient form of mountain worship still practiced in some areas of Japan. Originally from New Zealand, Bunting moved to Yamagata Prefecture to teach English in 2010 but threw himself into Shugendō in 2016 after the sudden death of his father. “Most people come to Shugendō at some sort of turning point in their life,” he told me. “We see mountains as a space to reflect and gather, get direction.”
Soon, Bunting was initiated as a yamabushi, literally “one who prostrates before mountains.” Based around the three sacred peaks of Dewa Sanzan, his sect follows a Shinto-adjacent form of Shugendō and incorporates elements of Buddhism and nature worship. Devotees walk for miles through the thick-forested slopes dressed in white vestments and take part in weeklong ascetic retreats punctuated by prayer, chanting and rituals, the contents of which are closely guarded. Upon arriving at a place of meditation, a conch is blown to greet the Shinto divinities, the kami, “to let them know that we’re entering their realm.”"
"It’s not clear whether that vision will persuade scientists skeptical of Sheinbaum Pardo to vote for her next month. Carlos Bravo Regidor, an independent political analyst, cautions that even if Sheinbaum Pardo wins, as expected, “the truth is that scientific policy is not going to be one of her priorities,” given her crowded policy agenda.
Bravo Regidor also notes that, in launching attacks on the research community, Morena and López Obrador adopted a tactic used by populist movements in other countries—such as Hungary, Brazil, and the United States—likely because scientists can be painted as a symbol of a technocratic elite. As a result, Sheinbaum Pardo—as his successor—might feel pressured by Morena supporters to support populist policies that might clash with her scientist identity."
"While this Comment is not a critique of climate activism per se, I am foremost concerned by an increasing number of climate scientists becoming climate activists, because scholars should not have a priori interests in the outcome of their studies. Like in any academic case, the quest for objectivity must also account for all aspects of global climate change research. While I have no problem with scholars taking public positions on climate issues, I see potential conflicts when scholars use information selectively or over-attribute problems to anthropogenic warming, and thus politicise climate and environmental change. Without self-critique and a diversity of viewpoints, scientists will ultimately harm the credibility of their research and possibly cause a wider public, political and economic backlash.
Likewise, I am worried about activists who pretend to be scientists, as this can be a misleading form of instrumentalization. In fact, there is just a thin line between the use and misuse of scientific certainty and uncertainty, and there is evidence for strategic and selective communication of scientific information for climate action6. (Non-)specialist activists often adopt scientific arguments as a source of moral legitimation for their movements6, which can be radical and destructive rather than rational and constructive. Unrestricted faith in scientific knowledge is, however, problematic because science is neither entitled to absolute truth nor ethical authority7. The notion of science to be explanatory rather than exploratory is a naïve overestimation that can fuel the complex field of global climate change to become a dogmatic ersatz religion for the wider public. It is also utterly irrational if activists ask to “follow the science” if there is no single direction. Again, even a clear-cut case like anthropogenically-induced global climate change does not justify the deviation from long-lasting scientific standards, which have distinguished the academic world from socio-economic and political spheres."
"For writers with other kinds of mental illness, a writing career can pose other challenges. Repetitive habits that are part of obsessive compulsive disorder (OCD), for example, might lead a person to spend far too much time on reporting tasks such as finding sources, reading and digesting scientific papers, and annotating stories for fact-checking—a recipe for low productivity and high stress. And spending long periods of time alone, a frequent side effect of freelancing, can amplify symptoms of any mental illness, from schizophrenia to general anxiety disorder. When a person is socially isolated, destructive thoughts can easily replace conversation; with no one there to answer or rationalize them, self-doubts can grow out of control.
There’s no one-size-fits-all prescription for writing about science while managing a mental illness. The relationship between the two is different for everyone, and for each of us, it takes time to understand a condition that lies in the black box of the brain, honed by habits that are unique to every person.
“I guess the basic answer is patience—which, as generic as it seems, is a well that many people with anxiety, depression, or other issues really need to be able to draw from,” says John Wenz, a freelance science writer in Madison, Wisconsin, who has bipolar type II."
"Besides its beautiful portrayal of a declining, melancholy world of perpetual autumn, what sets aside Elden Ring is its complexly layered difficulty. Elden Ring is quite eager to kill you, with a million ways to put a player down. But it is not meanly difficult, or insurmountably difficult. Most importantly, it is not difficult as part of a profit-seeking monetization loop. Instead, the failure states that are so often leveraged to extend playtime and coerce spending in most other games are here used as friction to build atmosphere. The constant starting again is exhausting, often stressful, sometimes infuriating. It is never meaningless, however: it confidently contradicts the worries of other mediums and the too-often-true accusations of slop with its deep understanding of how to create drama within any individual moment. Participating in its loops of death and rebirth as a player is to be fully within the Lands Between. Elden Ring presents a once-flourishing kingdom literally consumed by creeping nihilism and reflexive despair, which gives sympathetic resonance to the player’s determined and confident attempts to surmount these challenges. The most powerful or villainous enemies withdraw into themselves and let the world rot, while the weakest literally cower from the player, so exhausted by the idea of another painful death. Not the player, though: they exist in deliberate dramatic contrast to these characters by virtue of their own interactive participation with the world, making them the hero as both part of the text and as a meta-textual frame for the whole story.
By persisting in a world that trends downward, your own failures take on a defiant quality. The failure loop of the game incentivizes the player to loop again. This is where Elden Ring’s difficulty is particularly clever: because a player pays no consequence besides dropping experience points on the ground where they died, there is a hard limit on what the game can take away from them. There is an interactive choice and freedom even within these fail states, as you can abandon them or return again, fighting through all you had before; this in turn creates an incredible carrot-and-stick effect that, should you gamble on reclaiming your hard-won gains, doubles the stakes. While it is repeating the same content on the surface, there is a tangible and meaningful sense of cumulative progress and tactical variation on every death."
"On May 10, Oklo Inc., a nuclear energy startup, began trading on the New York Stock Exchange. The chairman of the company is none other than Sam Altman, the CEO of the artificial intelligence leader, OpenAI, that launched the generative AI revolution with the release of ChatGPT late in 2022. The staggering language, image, and video processing capabilities of ChatGPT and other similar chatbots depend on large language models (LLMs) trained through computations of data at unprecedented scale. Computational power has thus become, in Sam Altman’s words, “the currency of the future” and “the most precious commodity in the world.”
The training and use of LLMs require huge amounts of electrical power and cascades of advanced microchips. Altman’s nuclear investment reflects his belief that Oklo’s microreactors can satisfy the future power requirements of AI models. In some ways, the compatibility is intuitive. Large data centers, especially those set up in remote areas with greater land availability, require power sources that avoid both the intermittency of renewables and the fuel delivery requirements of traditional thermal power plants. Such is Oklo’s narrative as it touts recent power purchase agreements with data center operators such as Wyoming Hyperscale.
But the intuitive compatibility between AI and nuclear power does not exempt the latter from traditional concerns about economics and safety, despite efforts by Oklo and other vendors of “advanced” reactors to downplay those concerns."
"For our societies, computation is a means of world ordering. From the earliest marks of symbolic notation, computation was a foundation of what would become complex culture. The signifiers on clay in Sumerian cuneiform are known as a first form of writing; in fact they are indexes of transactions, an inscriptive technique that would become pictograms and over time alphanumeric writing, including base 10 mathematics and formal binary notation. There and then, the first writing is “accounting”: a kind of database representing and ordering real world communication. This artifact of computation already prefigures the expressive semiotics, even literary writing, that ensues in centuries to come.
Over recent centuries, and accelerating during the mid-20th century, technologies for the artificialization of computation have become more powerful, more efficient, more microscopic and more globally pervasive, changing the world in their image. “Artificialization” in this context doesn’t mean fake or unnatural, but rather that the intricate complexity of modern computing chips, hardware and software did not evolve blindly; it is the result of deliberate conceptual prefiguration and composition, even if by accident.
Given the scale and complexity of computational systems, what is the future of their deliberation and composition?"
"In this paper, we study the highly competitive arena of baby naming. Through making severalExtremely Reasonable Assumptions (namely, that parents are myopic, perfectly knowledgeable agentswho pick a name based solely on its “uniquness”), we create a model which is not only tractable andclean, but also perfectly captures the real world. We then extend our investigation with numericalexperiments, as well as analysis of large language model tools. We conclude by discussing avenues forfuture research."
" The Hydrogen Ladder is my attempt to synthesise all the information known to me about all the factors driving technology uptake across all sectors of the economy in all countries of the world. Not ambitious at all!
What the Hydrogen Ladder is designed to do is to show how likely it is that any proposed use case ends up being a significant user of hydrogen (perhaps via one of its derivatives) in a decade or so, say 2035. That doesn't mean it's game over, the transition has happened, it just means it is absolutely clear by then that hydrogen is either the answer, or a major answer, to decarbonizing that use case.
In other words, it looks forward to a time after the current firehose of subsidies has subsided to affordable proportions, after there has been enough time for a bit more tweaking of technologies, after the emergence of supply chains, after a bit of familiarity has grown in the project finance sector, and so on. "