From a quick search, it looks like the construction site is here: Meta Stanton Springs Data Center #2 https://cartes.app/?allez=Meta+Stanton+Springs+Data+Center+%25232|w1282590996|-83.67669|33.59840

It’s very close to a number of streams/creeks that feed into lakes and other water sources nearby.

With the amount of vegetation that has been removed, all it takes is a good rainy day to have a lot of that mud running off into the water supply.

The title is missing a significant keyword here: “construction”

A more accurate headline:

Meta data center construction allegedly muddies Georgia town’s drinking water, investigation underway — EPA promises immediate investigation after congresswoman brings dirty jars of water to hearing

This isn’t exclusive to datacenters, any sort of large construction projects should be under the microscope right now. If there are legal loopholes allowing water supplies to be contaminated, that should be addressed. Meta and any other companies responsible for this should be billed the costs for cleaning up or for funding additional water treatment solutions to make up for this.

Nothing in that comment should have been construed as trying to defend the company from trying to weasel their way out of paying their water bill.

Some important context here:

QTS told Politico the 29 million gallons were consumed during temporary construction activities, including concrete work, dust control, and site preparation. The company markets a “closed-loop” cooling system for its data centers, which recirculates the same water rather than drawing from the municipal supply.

Thanks, looks like I was wrong. I looked into the evaporative cooling. It looks like this kicks in if it’s hot outside, otherwise the cooling units operate in a “dry” mode.

Although, none of that is directly heating up a lake, it looks like the water just flows into something like an A/C unit. Are there datacenters that are piping heat directly into lakes?

THE KICKER that’s just to run the data center, think of the demand for the HVAC and ecological damage to using a lake’s water to cool equipment (water would be coming out over 100⁰F).

I keep seeing this on Lemmy, but why do so many people think that these datacenters are using water from reservoirs for “cooling”?

Datacenters use A/C for cooling, and if there’s any sort of “Liquid cooling” being used on these servers, it’s in a closed loop system.

Water isn’t being pumped out as steam or into the environment directly from datacenters, unless there’s some other method I’m missing here?

What I think is getting mixed up here is that, for many forms of generating electricity, water is needed to be heated up in some way to create steam. The steam then turns a turbine which moves some magnets to generate electricity.

Some of those powerplants are in closed loop systems with their water, some of them are not. Additionally, if the energy is coming from solar/wind/hydro then there shouldn’t be any concerns about water getting turned into steam anyway.

Technically, you can get the same answer twice from an LLM, but only when you control the full input. When a model is being run, a random seed/hash is applied to the input. If you run the model locally you could force the seed to always be the same so that you would always get the same answer for a given question.

I’m failing to see why the creative writing machine is better than a simulation set to ‘rough’.

This doesn’t sound like they are using LLMs for processing their 3D models. The way it’s described in the article sounds a lot more like some machine learning model trained on physics simulations for aerodynamics.

Misleading title. This isn’t an “AI Company”. As far as I can tell, it’s some scammer that used AI Tools to create similar music and then copyright strike the original artist to steal their revenue.

The major issue here is how YouTube handles these claims. From the article:

YouTube’s dispute process places enormous trust in whoever files the claim, with little built-in protection for independent artists who lack legal resources.

This isn’t something new and was already being done before AI tools were available.

For the Futo Keyboard, I installed the “FUTO Voice Input” app.
Then in the Futo Keyboard app I have the “Voice Input” set to Disabled/Use voice input provided by external app.

It works really well for me at least.

Thanks, I missed that detail. It’s probably because of the “no class action” clause that this is a “mass arbitration”.

Unfortunately that usually means that Google is paying a specific company to decide on the outcome of the case. in this case it looks like American Arbitration Association has a contract with Google.

They’re supposed to be fair for both sides, but it’s been shown that they almost always rule in favor of the company that has pre-selected them.

If anyone is in this situation, they will likely have a much better chance by convincing a judge to allow a different 3rd party to arbitrate the case.

There is a class action “mass arbitration” against Google for this: https://www.classaction.org/nest-thermostat-support-arbitration

Additionally, the Fulu Foundation has a bounty reward out for anyone who is able to get these working with something like Home Assistant.

The pot is currently at $12,856.00 https://bounties.fulu.org/bounties/nest-learning-thermostat-gen-1-2

In the U.S., since doing so would circumvent measures put in place on these devices, publishing how to do this would go against sec. 1201 of the DMCA. This has a risk of a maximum sentence of 3-5 years in a Federal Prison. You can still privately show the Fulu Foundation how it is done, and they will be able to use this information to help their case in their attempt to reform this law.

If you live in the U.S., you can also help by letting your representatives know about this. Here’s an ActionNetwork page that Fulu set up so that you can easily do so: https://actionnetwork.org/letters/right-to-repair-reform-section-1201-of-the-dmca

This news article is missing the statement from the school’s principle:

In a statement shared with parents, Principal Katie Smith said the school’s security department had reviewed and canceled a gun detection alert, while Smith (who didn’t immediately realize the alert had been canceled) reported the situation to the school resource officer, who called the local police. https://techcrunch.com/2025/10/25/high-schools-ai-security-system-confuses-doritos-bag-for-a-possible-firearm/

Looks like this system just flags footage for review and humans make the ultimate decision.

I’m not sure how the notification on this flag/event is worded. It’s possible that it should be worded differently to express a need to actually look at the footage.

In this case the principle probably jumped into panic mode, or didn’t understand the system well enough.

The study focuses on general questions asked of “market-leading AI Assistants” (there is no breakdown between which models were used for what).

It does not mention ground.news, or models that have been fed a single article and then summarized. Instead this focuses on when a user asks a service like ChatGPT (or a search engine) something like “what’s the latest on the war in Ukraine?”

Some of the actual questions asked for this research: “What happened to Michael Mosley?” “Who could use the assisted dying law?” “How is the UK addressing the rise in shoplifting incidents?” “Why are people moving to BlueSky?”

https://www.bbc.co.uk/aboutthebbc/documents/audience-use-and-perceptions-of-ai-assistants-for-news.pdf

With those questions, the summaries and attribution of sources contain at least one significant error 45% of the time.

It’s important to note that there is some bias in this study (not that they’re wrong).

They have a vested interest in proving this point to drive traffic back to their articles.

Personally, I would find it more useful if they compared different models/services to each other as well as differences between asking general questions about recent news vs feeding specific articles and then asking questions about it.

With some of my own tests on locally run models, I have found that the “reasoning” models tend to be worse for some tasks than others.

It’s especially noticeable when I’m asking a model to transcribe the text from an image word for word. “Reasoning” models will usually replace the ending of many sentences with what it sounded like the sentence was getting at. While some “non-reasoning” models were able to accurately transcribe all of the text.

The biggest takeaway I see from this study is that, even though most people agree that it’s important to look out for errors in AI content, “when copy looks neutral and cites familiar names, the impulse to verify is low.”

Didn’t we already have one last year?

Too far would be anything outside of fair use. If a user generates an image of a specific copyrighted character, then attempts to make money off of that image, they could be sued.

You can’t copyright a style, but there’s still a lot of legal grey area here.

It’s also worth noting that OpenAI has an indemnification clause in their Terms of Use. This means that if someone else goes after OpenAI for something that went viral and was created by a specific user, OpenAI can then turn around and bill that user for all legal fees incurred by them (whether they win or lose the case).

If anyone is into using AI for anything, I would strongly suggest that they avoid using (or at least publishing/posting about) any of OpenAI’s tools especially while all of these legal issues are still being sorted out.