Weekend Links

July 12, 2026

This week’s* links. AI’s resource use and bias replication, pushback, community computing, CHATONS, CO2 tracking, DRM-free books.

* happy weekend: two weeks’ worth of!

• AI Data Centers’ Hidden Water Use Alarms Korea

  According to Google’s recent “Sustainability Report,” its water usage in 2023 surged by 34% year-on-year to 10.9 billion gallons (approximately 41.26 billion liters). However, a study by the Netherlands’ Vrije Universiteit Amsterdam analyzing Google’s power consumption structure estimated that hidden indirect water usage could be at least three times higher than direct usage.

  Microsoft and Amazon also disclose only direct water usage. Meta revealed that its 2024 indirect water usage reached 19 billion gallons (approximately 71.92 billion liters), exceeding direct usage by over 20 times. However, it has yet to establish concrete measures to reduce indirect water usage, only setting an abstract goal of restoring water resources by 2030.

• also, Google’s exponential path to climate-wrecking digital bloat

  The company’s total electricity consumption jumped from 31 terawatt hours (TWh) in 2024 to 43 TWh in 2025. This is very easily the biggest increase in their electricity consumption ever, and it puts them way ahead of Microsoft. It is almost certainly a reflection of the obscene energy hunger of their ever-expanding bloated generative AI systems, and a vindication of the warnings we’ve been raising for several years now. […] Google’s power consumption rose by 7 TWh between 2023 and 2024. That was bad. But it rose by a whopping 12 TWh between 2024 and 2025, almost double last year’s increase. Google’s power consumption isn’t just growing – the rate at which it is growing is growing. We have a word for this: exponential growth.

• Godot: Changes to our Contribution Policies

  AI contributions have the added pain of being demoralizing. Reviewing PRs is already tedious work, but it is rewarding because reviewers generally feel that their efforts are contributing to educating a new contributor (who may become a future maintainer/reviewer). If your feedback on PRs is just being absorbed by a machine and not going towards mentoring a potential future maintainer, it becomes much harder to justify spending your free time on PR review. […]

  • No autonomous AI agent use or vibe coding […]
  • No use of AI to generate substantial pieces of code […]
  • No AI-generated text in human-to-human communication […]

  From the point of someone not using generative AI but having to discuss and review contributions generated by others: this is spot on. The use of generative AI undermines the knowledge sharing part of programming cand causes a regression in the understanding of the system by its authors.

• People power works: evil Thiel’s Ireland retreat cancelled after campaign

  The pressure works.

  In a sign of how hollow power can be once confronted, a Wicklow hotel has cancelled a Peter Thiel led event after a relatively brief campaign from Irish activists.

  A collection of the world’s most despicable ruling class parasites were set to infest the Powerscourt Hotel as part of a conference run by a secretive group called Dialog. Co-founded by tyranny advocate and genocide backer Thiel in 2006, Dialog were set to use the meeting just south of Dublin to discuss topics entitled Navigating WWIII, Battlefield Technologies and Build-a-Cult.

• Community Computing: The Future We Deserve?

  We can’t all self host everything.

  Near as I can tell, the only workable alternative is a grassroots effort to bring ethical services to small communities. And because support for the services should be the responsibility of that community, a cooperative model makes a good deal of sense. Like webrings, this is hardly a new idea. Co-ops for bulletin board systems have been around longer than the web we know. Also like webrings, the idea could use some updating for modern realities.

  This strategy is not without challenges.

  I have been living in France for a couple of years now and one of the most promising things here is that I see many more associations and coops than companies focused on leveraging other people’s cash to become unicorns, e.g. Framasoft or hosters at CHATONS (“Collective of Hosters Alternative, Transparent, Open, Neutral and Solidarity”/“Collectif des Hébergeurs Alternatifs, Transparents, Ouverts, Neutres et Solidaires”).

• The bottleneck might be the air in the room

  Researchers at Lawrence Berkeley National Laboratory put people in a chamber and varied only the CO2. At 1,000 ppm, performance dropped significantly on six of nine decision-making measures compared with a clean-air baseline of 600. At 2,500 ppm, seven of the nine fell substantially, some into a range they called dysfunctional. A separate study out of Harvard found cognitive scores declining as CO2 rose, with the steepest losses in exactly the domains you called the meeting for: strategy, planning, and using information under pressure.

  Covid could have been the “oh shit we need clean water and sewage management” moment for air.

• AI Hiring Tools Can Yield Racial Bias and Systemic Rejection

  We discovered that 26% of Black applicants and 15% of Asian applicants applied to positions where the AI system discriminated against their racial group. To put this in perspective: If the AI had recommended Black and Asian candidates at the same rate as it recommended the most-favored group (typically white applicants), 40,000 more of their applications would have advanced to the next stage of hiring.

  This is not surprising, Emily M. Bender and others warned about this way back in On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? but now we are going from talking about future logical consequences to evidence from already observed activity.

• Zuckerberg ‘Admits’ Meta’s Layoffs Were Ineffective

  Zuckerberg yet again plunged his company into chaos after the success of Claude Code in December 2025. Knowing the company was behind in developing AI products after observing the rise of agentic coding, Zuckerberg effectively put Alexandr Wang, the chief of Meta’s AI division, in charge of the entire company. The only thing Wang did was wrongly determine that all human programmers were a waste of time and money and that it would be better to fire them and spend the freed-up cash on talented AI engineers who would unwittingly develop their own replacements.

• The Ultimate List of DRM-free Bookshops looks interesting.

  Mostly pages of individual authors and smaller, special interest publishing houses, but there’s also kickstarter and itch.io.