Elektrine lite

← Feed

@toxi@mastodon.thi.ng

2026-08-19 09:03 UTC

Great article about the loss of diversity induced by recommender systems (including LLMs): Temperature Zero for Culture: Why Everything Is Starting to Look the Same "A world that keeps removing the unlikely also removes the things that could have become the next likely. Ecologists have known this for a long time already and even priced it: they call it the insurance value of biodiversity, the standing reserve of rare species that cost energy and contribute nothing, until the climate shifts and one of them turns out to be what survives. A monoculture is efficient right up until the weather changes." https://laurenleek.substack.com/p/temperature-zero-for-culture-why Related: In my 2013 EYEO talk I used the Tower Of Babel as metaphor to talk about polyglot, diverse culture as key long-term survival strength (not punishment!) and to address the growing lack of diversity and rise of monocultures in creative coding fields, back then largely caused by centralization effects of popular frameworks (e.g. Processing, OpenFrameworks) and their social attractor effects restricting the practice and surrounding educational efforts to mostly a single family of (C-based) programming languages whilst completely disregarding other approaches, techniques and influences. My main argument was (and still is) that diversity of languages and programming approaches (functional, imperative, object-oriented, logic, declarative...) is an important factor for long-term cultural survival and robustness, something I've also been warning about being increasingly under threat by LLM-based coding. The latter is accelerating the monopolization of existing tools/frameworks, due to their over-representation in training sets and therefore offering more "stable", usable LLM solutions, leading to diversity collapse and making it increasingly hard for new contenders to establish themselves. Also wrote more about that previously... #Average #Diversity #Robustness #Survival #Culture #ProgLang #LLM

Replies (3)

  • @toxi@mastodon.thi.ng only when it is same enough can you see the stars

    Open ##5082425

  • Awright cuh so whachoo be sayin n sheeit is that I be makin dat deah oh-rihj-jihn-nul cawntent, like we wuz duh rijnuhl kangz n what have you if I may pro-ceed my nig-gaaaah. :fap:

    Open ##5082431

  • @petpet@social.tchncs.de 2026-08-19 10:35

    @toxi@mastodon.thi.ng The same applies to recomendation algorythms which always point to the more mainstream, so e.g. you are looking for something totally niche and start with an artist who has less than 100 streams or so, it will not recomend another artist on the same lavel of exposure but always a bigger coverage. result: its extremely difficult to be niche at all.

    Open ##5082436