3 ms·
From the article: > Using the chatbot for between ten to 50 queries consumes about two litres of water, according to experts from the University of California,
by p2detar 2y ago
From the article:
> Using the chatbot for between ten to 50 queries consumes about two litres of water, according to experts from the University of California, Riverside.
From ChatGPT:
> A single BTC transaction consumes approximately 1,000–1,500 kWh of electricity, with cooling processes for fossil fuel and nuclear power plants using about 2–3 liters of water per kWh. This translates to an indirect water consumption of around 2,500–3,750 liters (660–990 gallons) per transaction, depending on the energy mix.
If accurate, I still prefer ChatGPT.
- shakow 2y agoHow is it a “consumption” (in either case) if the water is basically just running through a pipe and released a couple miles down the line/evaporated into the atmosphere without being polluted?
- MrMcCall 2y agoWe're having a global heating problem, so the water use is just a symptom of our systemic abuse of the planet's resources.
- ttyprintk 2y agoIt’s irresponsibly common for journalists to use the term for once-through cooling. But evaporative cooling consumes water just like the human body.
- MrMcCall 2y agoHere's the reality: just because a lot of people are doing something, it doesn't mean thay're not sunk-cost fools.
- i_cannot_hack 2y ago> If accurate, I still prefer ChatGPT. It looks like you are using ChatGPT too much and feeding yourself untrustworthy information for no reason. Internet search still exists, there is no need to use ChatGPT for this. Simply typing "how much water per btc transaction" into Google will give you this top response much faster than ChatGPT would: > Buying or selling bitcoin uses 16,000 litres of clean water for every single transaction. This is reported by New Scientist, who's reporting is based on actual calculations (using real data) from a PHD candidate at the Amsterdam School of Business and Economics. Whereas the number given by ChatGTP is based on nothing but the statistical probability of what would sound correct in this context, and has no guarantee of accuracy at all. In this case it was off by an order of magnitude compared to the number based on real data. The internet search probably used a lot less water as well (beside being both faster and more trustworthy).