I just read an article, The average AI criticism has gotten lazy, and that’s dangerous - Redeem Tomorrow, whose proposed solution to the AI situation is for everyone to have access to their own pattern synthesizer (their neologism for “AI”), instead of just a handful of companies controlling them. In some sense that is good, it would avoid enshittification I think. They argue this because there isn’t a way to remove the synthesizers from existence, since they only need commodity computing hardware to be produced. The process is well documented and broadly available. Limiting the availability of the knowledge and/or hardware isn’t something we’re willing/should be willing to do.
I can’t help but feel like this is a bit weak. They mention bullshit generation and information warfare as one of the real problems with pattern synthesizers, but their solution does not at all address that. And what kind of energy usage will this drive if everyone is supposed to have one? The article argues that this is a non-issue because operators are incentivized to drive down operating costs. This has already happened in the computing industry at large before, with processing power in general — which suggests that it will happen again with AI. Further, if we’re going to go with this argument anyways, why the focus on this specific part of ICT (Information and Communications Technology) and not ICT at large? The problem with this argument is that while the operating costs per computation has shrunk dramatically, the overall costs have risen massively as demand has increased (Santarius et al. 2022). This is called a rebound effect. What is to say that the gains from efficiencies would be greater than the losses from global usage of pattern synthesizers?
Like, in a world where ICTs emits somewhere between 1.8% and 3.8% of global greenhouse gas emissions (Freitag et al. 2021) I don’t think we need even more energy hungry ICT.(sidenote: In a sense then I agree with the article in that I think we shouldn’t just focus on pattern synthesizers, but I suppose I approach this as a microcosm for ICT at large. I don’t think we need more “smart” devices, faster computers, or whatever.1 ) If we’re giving up on containing/reducing just one small part of ICT, how are we going to do anything at all with the wider ICT sector, let alone the global economy?
I don’t know how else to respond to how “AI” companies are breaking things but to use it as a rallying point for a Butlerian Jihad of sorts. To use this as an opportunity to advocate for some form of digital degrowth, minimalism or reevaluation where we use advancements in hardware to do less computation with less energy, where the focus is to make hardware last longer by making software that doesn’t require faster hardware and designing computers that are repairable.
This is my response. They have fucked around and we are finding out. Away with AI. Away with needlessly slow software that obsoletes functioning hardware. Away with Silicon Valley.
References 🔗︎
Articles from blogs I follow around the net
Pluralistic: Big Tech disrupted disruption (08 Feb 2024)
Today's links Big Tech disrupted disruption: We don't have to care, we're the phone company. Hey look at this: Delights to delectate. This day in history: 2004, 2009, 2014, 2019, 2023 Colophon: Recent publications, upcoming/recent appearances,…
via Pluralistic: Daily links from Cory Doctorow February 8, 2024The Language of Power
Rosemary Kirstein In the fourth and as yet final book of the Steerswoman series, Rowan and Bel return to Donner, where they …
via A Working Library February 8, 2024Wishful bio weapons
Currently when talking about very big large language models even people who want to be taken seriously talk a lot about bio or chemical weapons: Will "AI" systems make creating bio weapons too easy? But is that a real danger? Will ChatGPT give ter…
via english Archives - Smashing Frames February 5, 2024Audio Newsletter: 1,000 Hours Outside
Facts: How much can carbon farming help? 🧑🌾 🌾| Feelings: Finding purpose between "What should we do?" + "What can I do?" 🎯 🌱 | Action: 1,000 hours outside! ☀️❄️
via We Can Fix It February 2, 2024Sobre ‘relacionar-se’: consigo mesmo, com velhos amigos e com comunidades inteiras
Estou aqui a pensar como os ‘ciclos’ pautam a minha existência – toda, tanto pessoal como profissional. E se eu sinto em mim ciclos claros, curtos, como os das estações do ano do lugar do mundo onde eu cresci e como os ciclos anuais, há outros mais difíce…
via News – Transition Network January 16, 2024Generated by openring
-
In a sense then I agree with the article in that I think we shouldn’t just focus on pattern synthesizers, but I suppose I approach this as a microcosm for ICT at large. I don’t think we need more “smart” devices, faster computers or whatever. ↩︎