Logdyco

Endgame for the Open Web - Anil Dash

The open web is something extraordinary: anybody can use whatever tools they have, to create content following publicly documented specifications, published using completely free and open platforms, and then share that work with anyone, anywhere in the world, without asking for permission from anyone. Think about how radical that is.

Whether or not they succeed is going to be decided by decisions that we all make as a community in the coming months.

J'ai du mal à croire tout ce qui se base sur une décision collective massive, surtout si c'est pas facile.

  • Publishers who still share their content openly [...] are being absolutely hammered by ill-behaved AI bots. [...] search engines had permission to crawl sites because they were going to be sending users to those sites.If they're hitting your site half a million times for every one user they send to your site, all they're giving you is higher costs.

Ouch. Les taux de clic vs. citation sont très mauvais en effet.

  • LLM-based AI platforms that have trained their AI models on this content gathered without consent typically have almost no links back to the original source content, and either bury or omit credits to the original site; as a result, publishers in categories like tech media have seen their traffic crater by over 50%, with some publishers seeing drops of over 90%.
  • As publishers see the danger from AI bots expand, they retreat to putting more and more content behind either password protection or payment walls or both, leaving the only publicly-accessible content to be AI-generated slop.
  • Open APIs, a building block of how developers build new experiences for users, and for how researchers understand people's behavior online, are rapidly being locked down due to abuse from LLMs [+ exemple de Twitter a fermé son API publique]
  • Independent media based on open formats, like podcasts, are also under siege as platforms like Apple's podcasts move to closed infrastructure which means that content creators are now required to work with Apple's approved partners.
  • Open source software projects, which power the vast majority of the internet's infrastructure, are now beleaguered by constant slop code submissions being made by automated AI code agents.
  • The most vital open content platforms, like Wikipedia, are under direct attack from bad-faith campaigns. Elon Musk has created Grokipedia to directly undermine Wikipedia with extremist hate content and conspiracist nonsense, by siphoning off traffic, revenues, and contributors from the site.
  • A simple text file called robots.txt functioned for decades to describe the way that tools like search engines ought to behave when accessing content on websites, but now it is effectively dead as Big AI companies unilaterally decided to ignore more than a generation of precedent, and do whatever they want with the entirety of the web, completely without consent.
  • Open source software licenses, which used to be a bedrock of the software community because they provide a consistent way of encoding a set of principles in the form of a legal contract, are now treated as a minor obstacle which can be trivially overcome using LLMs.

Une belle liste de méfaits bien connus, et pas seulement des spécialistes.

Alright, if it’s not hopeless, what are the concrete things we can do? The first thing is to directly support organizations in the fight.

Extraits cités issus de ce lien visité le , crédits: Anil Dash