Dec 11, 2023: Guillaume Pitron, author, "The Dark Cloud"
Show Notes
• The Dark Cloud: The Hidden Costs of the Digital World, by Guillaume Pitron
• Total Trust, documentary now playing at Film Forum (209 West Houston St in Manhattan) - through this Friday, Dec 15
• A.I. Could Soon Need as Much Electricity as an Entire Country (NYT, Oct 10, 2023): “In a middle-ground scenario, by 2027 A.I. servers could use between 85 to 134 terawatt hours (Twh) annually. That’s similar to what Argentina, the Netherlands and Sweden each use in a year, and is about 0.5 percent of the world’s current electricity use.”
• Big Tech’s thirst for AI dominance may bring literal thirst for everyone else (The Hustle, Aug 4, 2023): “Globally, data centers are forecast to consume 450m gallons of water daily by 2030, up from ~205m in 2016, according to data reviewed by Bloomberg. This is particularly worrisome for drought-stricken farming regions like Spain’s Talavera de la Reina, where a $1.1B Meta-planned data facility could gulp 176m gallons annually. . . . [And] a 20-question convo with ChatGPT equates to ~500 milliliters of water use, about the size of a water bottle.”
• Critics Furious Microsoft Is Training AI by Sucking Up Water During Drought (Futurism, Sep 26, 2023): “Microsoft’s data centers in West Des Moines, Iowa guzzled massive amounts of water last year, the Associated Press reported earlier this month, to keep cool while training OpenAI’s ChatGPT-4, the Microsoft-backed company’s most advanced publicly available large language model. Critics point out a further inconvenient detail: this happened in the midst of a more than three-year drought, further taxing a stressed water system that’s been so dry this summer that nature lovers couldn’t even paddle canoes in local rivers.”
“. . . Microsoft increased worldwide water consumption by a whopping 34 percent — up to almost 1.7 billion gallons annually — last year, which outside researchers told the AP is most likely due to increased AI training. That’s dwarfed by Google, which used 5.6 billion gallons last year, a 20 percent jump that’s also likely attributable to machine learning.”
• Arizona is running out of water. Big Tech data centers are partly to blame. (Insider, 2023): “[A few years ago] Google was planning a massive data center in Mesa, [Arizona,] just east of Phoenix. The deal guaranteed Google 1 million gallons of water a day to cool the facility, and up to 4 million gallons a day if it hit project milestones. (That’s a lot of water. Arizona residents each use about 146 gallons a day).”
• Google Is Using A Flabbergasting Amount Of Water On AI (Futurism, July 28, 2023): “According to the tech giant’s 2023 Environmental Report, the company used an astronomical 5.6 billion gallons of water last year. That’s a 20 percent increase over its 2021 usage, which can likely be attributed in large part to Google’s growing AI efforts. Training these algorithms in massive data centers consumes immense amounts of energy, plus huge amounts of water for cooling. And the majority of this water isn’t even being pulled out of streams — it’s clean enough to be used as drinking water.”
• “Digital and Sustainability” is not like “Q and U” (CB Bhattacharya, Sep 21, 2023): “Many of the metals, minerals, and rare elements necessary in the manufacture of digital devices are often harvested in the world’s conflict zones (e.g., the Democratic Republic of Congo) where child labor, slave labor, and violent and inhumane working conditions make the ‘inexpensive’ final product cost possible. This is true for devices built for personal use, but also for the devices that are used in mass deployment in data centers worldwide.”
• Generating AI Images Uses as Much Energy as Charging Your Phone, Study Finds (Gizmodo, Dec 1, 2023): “Creating images with generative AI could use as much energy as charging your smartphone, according to a new study Friday that measures the environmental impact of generative AI models for the first time. Popular models like ChatGPT’s Dall-E and Midjourney may produce more carbon than driving 4 miles.”
• Every Bitcoin payment ‘uses a swimming pool of water’ (BBC, Nov 29, 2023): “Every Bitcoin transaction uses, on average, enough water to fill “a back yard swimming pool”, a new study suggests. That’s around six million times more than is used in a typical credit card swipe, Alex de Vries of Vrije Universiteit Amsterdam, calculates. . . . Up to three billion people worldwide already experience water shortages, a situation which is expected to worsen in the coming decades, the study notes.”
• Total Trust, documentary now playing at Film Forum (209 West Houston St in Manhattan) - through this Friday, Dec 15
• A.I. Could Soon Need as Much Electricity as an Entire Country (NYT, Oct 10, 2023): “In a middle-ground scenario, by 2027 A.I. servers could use between 85 to 134 terawatt hours (Twh) annually. That’s similar to what Argentina, the Netherlands and Sweden each use in a year, and is about 0.5 percent of the world’s current electricity use.”
• Big Tech’s thirst for AI dominance may bring literal thirst for everyone else (The Hustle, Aug 4, 2023): “Globally, data centers are forecast to consume 450m gallons of water daily by 2030, up from ~205m in 2016, according to data reviewed by Bloomberg. This is particularly worrisome for drought-stricken farming regions like Spain’s Talavera de la Reina, where a $1.1B Meta-planned data facility could gulp 176m gallons annually. . . . [And] a 20-question convo with ChatGPT equates to ~500 milliliters of water use, about the size of a water bottle.”
• Critics Furious Microsoft Is Training AI by Sucking Up Water During Drought (Futurism, Sep 26, 2023): “Microsoft’s data centers in West Des Moines, Iowa guzzled massive amounts of water last year, the Associated Press reported earlier this month, to keep cool while training OpenAI’s ChatGPT-4, the Microsoft-backed company’s most advanced publicly available large language model. Critics point out a further inconvenient detail: this happened in the midst of a more than three-year drought, further taxing a stressed water system that’s been so dry this summer that nature lovers couldn’t even paddle canoes in local rivers.”
“. . . Microsoft increased worldwide water consumption by a whopping 34 percent — up to almost 1.7 billion gallons annually — last year, which outside researchers told the AP is most likely due to increased AI training. That’s dwarfed by Google, which used 5.6 billion gallons last year, a 20 percent jump that’s also likely attributable to machine learning.”
• Arizona is running out of water. Big Tech data centers are partly to blame. (Insider, 2023): “[A few years ago] Google was planning a massive data center in Mesa, [Arizona,] just east of Phoenix. The deal guaranteed Google 1 million gallons of water a day to cool the facility, and up to 4 million gallons a day if it hit project milestones. (That’s a lot of water. Arizona residents each use about 146 gallons a day).”
• Google Is Using A Flabbergasting Amount Of Water On AI (Futurism, July 28, 2023): “According to the tech giant’s 2023 Environmental Report, the company used an astronomical 5.6 billion gallons of water last year. That’s a 20 percent increase over its 2021 usage, which can likely be attributed in large part to Google’s growing AI efforts. Training these algorithms in massive data centers consumes immense amounts of energy, plus huge amounts of water for cooling. And the majority of this water isn’t even being pulled out of streams — it’s clean enough to be used as drinking water.”
• “Digital and Sustainability” is not like “Q and U” (CB Bhattacharya, Sep 21, 2023): “Many of the metals, minerals, and rare elements necessary in the manufacture of digital devices are often harvested in the world’s conflict zones (e.g., the Democratic Republic of Congo) where child labor, slave labor, and violent and inhumane working conditions make the ‘inexpensive’ final product cost possible. This is true for devices built for personal use, but also for the devices that are used in mass deployment in data centers worldwide.”
• Generating AI Images Uses as Much Energy as Charging Your Phone, Study Finds (Gizmodo, Dec 1, 2023): “Creating images with generative AI could use as much energy as charging your smartphone, according to a new study Friday that measures the environmental impact of generative AI models for the first time. Popular models like ChatGPT’s Dall-E and Midjourney may produce more carbon than driving 4 miles.”
• Every Bitcoin payment ‘uses a swimming pool of water’ (BBC, Nov 29, 2023): “Every Bitcoin transaction uses, on average, enough water to fill “a back yard swimming pool”, a new study suggests. That’s around six million times more than is used in a typical credit card swipe, Alex de Vries of Vrije Universiteit Amsterdam, calculates. . . . Up to three billion people worldwide already experience water shortages, a situation which is expected to worsen in the coming decades, the study notes.”
- Guest
- Host
- Comments
- Playlist & Comments at WFMU
- Aired
- Dec 11, 2023