Category: AI

Southaven Residents Near xAI Turbines Complain of Noise Levels ⇢

⌘ Permalink

Kailynn Johnson, writing for The Memphis Flyer, about the turbines in Southaven, Mississippi that are powering xAI’s second datacenter site in Memphis:

In a YouTube reel, a resident named Jason Haley showed that the noise from the turbines caused his decibel reader to fluctuate from the 40s to the 60s when used inside. Haley went outside, to where the reader reached the 70s.

“Even when [the turbines] are not at peak loudness it’s [a] constant high-pitched noise that just doesn’t end,” Haley said. 

Haley is seeking answers to whether or not this will be a temporary or long-term issue. He has been told by city officials that they agree the sound is an issue but is not confident it will be properly addressed.

“I don’t know if this noise will go on for a couple more weeks, till the end of next year or what — that’s the answer I’m looking for,” Haley said.

According to Haley, local code enforcement and police said “there’s nothing they can do” to address the problem.

I’m Really Starting to Think These OpenAI Folks Are No Good ⇢

⌘ Permalink

Rebecca Bellan at TechCrunch:

OpenAI reportedly asked the Raine family – whose 16-year-old son Adam Raine died by suicide after prolonged conversations with ChatGPT – for a full list of attendees from the teenager’s memorial, signaling that the AI firm may try to subpoena friends and family.  

OpenAI also requested “all documents relating to memorial services or events in the honor of the decedent including but not limited to any videos or photographs taken, or eulogies given,” per a document obtained by the Financial Times. 

Speaking to the FT, lawyers from the Raine family described the request as “intentional harassment.” 

The new information comes as the Raine family updated its lawsuit against OpenAI on Wednesday. The family first filed a wrongful death suit against OpenAI in August after alleging their son had taken his own life following conversations with the chatbot about his mental health and suicidal ideation. The updated lawsuit claims that OpenAI rushed GPT-4o’s May 2024 release by cutting safety testing due to competitive pressure.

OpenAI Removes Dr. Martin Luther King Jr. From Sora Models

⌘ Permalink

Since the launch of Sora 2, the system has been willing to generate video of just about any famous person doing just about anything. Over the last couple of weeks, OpenAI has been clamping down on many examples of this, but this particular one caught my eye:

In an unsigned statement on X:

The Estate of Martin Luther King, Jr., Inc. (King, Inc.) and OpenAI have worked together to address how Dr. Martin Luther King Jr.’s likeness is represented in Sora generations. Some users generated disrespectful depictions of Dr. King’s image. So at King, Inc.’s request, OpenAI has paused generations depicting Dr. King as it strengthens guardrails for historical figures.

While there are strong free speech interests in depicting historical figures, OpenAI believes public figures and their families should ultimately have control over how their likeness is used. Authorized representatives or estate owners can request that their likeness not be used in Sora cameos.

Sam Altman, writing two weeks ago:

We have been learning quickly from how people are using Sora and taking feedback from users, rightsholders, and other interested groups. We of course spent a lot of time discussing this before launch, but now that we have a product out we can do more than just theorize.

We are going to make two changes soon (and many more to come).

First, we will give rightsholders more granular control over generation of characters, similar to the opt-in model for likeness but with additional controls.

He continues:

Second, we are going to have to somehow make money for video generation. People are generating much more than we expected per user, and a lot of videos are being generated for very small audiences. We are going to try sharing some of this revenue with rightsholders who want their characters generated by users. The exact model will take some trial and error to figure out, but we plan to start very soon. Our hope is that the new kind of engagement is even more valuable than the revenue share, but of course we we [sic] want both to be valuable.

There’s nothing like a zillion dollar company flying by the seat of its pants.

xAI Breaks Ground on Wastewater Treatment Plant in Memphis

⌘ Permalink

For months, the city of Memphis, xAI, and partners have been planning a wastewater plant to provide water for cooling the company’s first supercomputer in Memphis. Last week, ground was broken on the site, as reported by Samuel Hardiman at The Daily Memphian:

The plant will take wastewater treated at the City of Memphis’ nearby TE Maxson’s Wastewater Treatment Plant and push it through technology known as ceramic membranes. Those membranes — thousands of them — will filter contaminants out of the water and make it usable as an industrial coolant. It will then cool the data center and the TVA Allen Combined Cycle Plant.

Brent Mayo, xAI’s senior manager for site build and infrastructure, said that the membranes, stacked on top of each other, would be taller than four Empire State Buildings.

A nearby steel manufacturer was going to use the plant to recycle water used at their plant, but now no longer will. This means that Nucor Steel will still dump its used water into the Mississippi River, but could mean xAI could use the plant for its second datacenter as well:

Mayo said the facility could also potentially serve Colossus 2, the much larger data center the company is building about 8.5 miles away in Whitehaven, along the Mississippi border.

How the purified wastewater would get to Colossus 2 remains unclear. One way could be by truck. Another could be for xAI to purchase easements for a water pipeline through much of Southwest Memphis, which could be a difficult undertaking.

On its website, xAI says:

We are investing well north of $80 million to build a state-of-the-art water recycling plant in the next two years, protecting approximately 4.745 billion gallons of the Memphis Aquifer each year, and eliminating the need for our facility to draw from the Memphis Aquifer for industrial use.

In addition to serving our facility, the water recycling plant will serve other major industrial users in the area, which will further reduce demand on the Memphis Aquifer and protect the City’s current and future water needs.

Protect Our Aquifer, a non-profit dedicated to protecting the massive aquifers under the Memphis region, shared this statement:

This commitment by xAI sets a new standard for industry; we can and must use water
responsibly in the Mid-South. Our drinking water from the Memphis Sand Aquifer should never
be the default source for cooling water.

We expect Tennessee Valley Authority to continue being a part of this effort by signing the End
User Agreement and reducing their aquifer use by up to seven million gallons of water per day.
We are disappointed that Nucor Steel has decided not to move forward as an end-user at this
time and instead stay fully dependent on the Aquifer. We hope they reconsider.

The group does have four areas in which it would like to see progress, and it’s hard to disagree with any of them:

  • Charges xAI a fair fee for wastewater access;
  • Invests in workforce development for Memphians focused in the nearby communities;
  • Provides public and educational access for tours and STEM learning; and
  • Reduces liability for the City through strong public-private partnerships.

Google Announces $4 Billion Investment in AI Data Center in West Memphis, Arkansas

⌘ Permalink

After months of reporting and rumors, Google made its investment in West Memphis official today. Here’s The Daily Memphian’s Dima Amro:

Google’s new data center in West Memphis will cost $4 billion and take 18 to 24 months to complete, officials said Thursday, Oct. 2.

The data center will host cloud and artificial intelligence infrastructure, Ruth Porat, president and chief investment officer of Google, announced during a groundbreaking ceremony at the site in Crittenden County.

“Google is investing in the next generation of AI innovation in Arkansas and across the country,” Porat said. “The upside of AI cannot be unlocked without the energy it requires. That is why Google is building energy capacity that protects affordability for ratepayers and creates jobs that will drive the AI-powered economy.”

West Memphis — and its future AI datacenter — is right across the Mississippi River from downtown Memphis, and just a couple dozen miles from xAI’s first site in Memphis. That obviously leads to the obvious question about power deliver to the datacenter.

Google and Entergy, the utility company that operates in Arkansas, say that the site will have 600 megawatts of solar power and 350 megawatts of battery storage. Details past that are sparse, but Google cares about the environment more than its soon-to-be neighbor xAI seems to.

That bit from Porat about the “AI-powered economy” stood out at me as a bit over the top, but then I read the press release that landed my Inbox about the news, which includes this paragraph and photo:

Google’s deep, longstanding investments in American technical infrastructure, as well as research and development, will help the U.S. continue to lead the world in AI. This will unlock substantial economic opportunity for American businesses, advance scientific breakthroughs, fortify cybersecurity for the U.S., and create new career opportunities for millions of Americans.

West Memphis PR photo

I know it’s 2025 and everything is very USA USA USA but millions of career opportunities created by AI? Ehhhhhhh … time will tell.

So, why is Google coming to the so-called “Digital Delta” for this new site?

Water.

Memphis and the Mid-South sits above a vast aquifer. This is where our water comes from, and is a truly amazing natural resource. It gives Mid-South residents access to cheap, clean water.

It also gives companies easy-to-reach water for cooling massive data centers. One reason folks in Memphis are upset with xAI is the company’s use of aquifer water for cooling its vast array of GPUs. xAI has proposed a greywater facility that will reduce strain on the aquifer by nearly 10% by treating and recycling wastewater. That facility is not operating yet, but I hope Google will follow suit in West Memphis. A river may divide my city from Arkansas, but the water beneath our feet is the same.

Federal Goverment Partners with xAI

⌘ Permalink

Amrith Ramkumar at The Wall Street Journal (Apple News), writing about a new partnership between the Trump administration and his on-again, off-again bestie Elon Musk:

Under the agreement with the General Services Administration, which oversees technology procurement for the federal government, agencies will get access to models such as Grok 4 and a new fast version called Grok 4 Fast for a nominal fee of 42 cents, the GSA said in a news release Thursday.

The deal follows similar arrangements with Alphabet’s Google; the ChatGPT maker, OpenAI; and Anthropic, a model developer that has clashed with the White House. The arrangement means the government is now working with the four U.S. companies making the most-advanced AI systems. OpenAI and Anthropic agreed to provide their models for $1, while Google is charging 47 cents. Musk likes the number 42 because it is a reference to “The Hitchhiker’s Guide to the Galaxy,” the sci-fi novel.

Hayden Field reported on this for The Verge a couple of weeks ago, recapping just some of the controversies around xAI:

The height of Grok’s power, up to now, has been posting answers to users’ queries on X. But even in this relatively limited capacity, it’s racked up a remarkable number of controversies, often resulting from patchwork tweaks and fixed with patchwork solutions. In February, the chatbot temporarily blocked results that mention Musk or President Trump spreading misinformation. In May, it briefly went viral for constant tirades about “white genocide” in South Africa. In July, it developed a habit of searching for Musk’s opinion on hot-button topics like Israel and Palestine, immigration, and abortion before responding to questions about them. And most infamously, last month it went on an antisemitic bender — spreading stereotypes about Jewish people, praising Adolf Hitler and even going so far as to call itself “MechaHitler.”

Musk responded publicly to say the company was addressing the issue and that it happened because Grok was “too compliant to user prompts. Too eager to please and be manipulated, essentially.” But the incident happened a few weeks after Musk expressed frustration that Grok was “parroting legacy media” and asked X users to contribute “divisive facts for Grok training” that were “politically incorrect, but nonetheless factually true,” and a few days after a new system prompt gave Grok instructions to “assume subjective viewpoints sourced from the media are biased” and “not shy away from making claims which are politically incorrect.” Following the debacle, the prompts were tweaked to scale back Grok’s aggressive endorsement of fringe viewpoints.

The whack-a-mole approach to Grok’s guardrails concerns experts in the field, who say it’s hard enough to keep an AI system from veering into harmful behavior even when it’s designed intentionally, with some measure of safety in mind from the beginning. And if you don’t do that… then all bets are off.

Former Wondery COO Leading AI ‘Podcast’ Company ⇢

⌘ Permalink

When I saw this floating around, it was hard to believe that Caitlin Huston wasn’t writing for The Onion, but here I am, linking to a very real story at The Hollywood Reporter:

Inception Point AI is attempting to do just that, as the company builds a stable of AI talent to host podcasts, and eventually become broader influencers across social media, literature and more. Amid the high costs for producing narrative podcasts and pricy, short-term contracts for popular hosts, the idea here is being able to own, scale and control the talent (unlike those off-the-cuff humans) and produce shows at a minimal cost.

“We believe that in the near future half the people on the planet will be AI, and we are the company that’s bringing those people to life,” said CEO Jeanine Wright, who was previously chief operating officer of podcasting company Wondery, which has recently had to reorganize under the changing podcast landscape. 

The company is able to produce each episode for $1 or less, depending on length and complexity, and attach programmatic advertising to it. This generally means that if about 20 people listen to that episode, the company made a profit on that episode, without factoring in overhead.

If that was not enough nonsense for you, this might push you over the edge:

The team is in the midst of navigating the ethics around creating these AI personalities as the technology advances. Each host now identifies themselves as being AI at the top of the episodes, and they’ve stayed away from having the hosts invent their own backstories, for now, but that could come. Wright says she could eventually imagine having hosts chat with listeners, or sing “Happy Birthday” to them, but there’s wariness about diving in too deep. 

“I am not going to create a personality that somebody has a deep relationship with,” said William Corbin, co-founder and CTO of the company. He added that the company does not do hard news at this point, but Wright says they might in the future.

In addition to flooding the podcast landscape with AI garbage, this company doesn’t seem to understand how destructive these things can be. I can’t wait for the AI bubble to burst on companies like Inception Point.

Anthropic to Pay Authors ‘At Least’ $1.5 Billion ⇢

⌘ Permalink

Hayden Field at The Verge:

In what’s potentially the first major payout to creatives whose work was used to train AI systems, Anthropic has reached an agreement to pay “at least” a staggering $1.5 billion, plus interest, to authors to settle its class-action lawsuit. The amount breaks down to smaller payouts expected to be approximately $3,000 per book or work. Lawyers for the plaintiffs said it’s “believed to be the largest publicly reported recovery in the history of US copyright litigation.”

“This result is nothing short of remarkable,” the legal filing states.

We’ll see how this pans out, but for now, all I can think is that it couldn’t happen to a more deserving company.

The Browser Company Acquired by Atlassian ⇢

⌘ Permalink

In probably unsurprising news, The Browser Company has a new owner, as reported by David Pierce at The Verge:

Mike Cannon-Brookes, the CEO of enterprise software giant Atlassian, was one of the first users of the Arc browser. Over the last several years, he has been a prolific bug reporter and feature requester. Now he’ll own the thing: Atlassian is acquiring The Browser Company, the New York-based startup that makes both Arc and the new AI-focused Dia browser. Atlassian is paying $610 million in cash for The Browser Company, and plans to run it as an independent entity.

The conversations that led to the deal started about a year ago, says Josh Miller, The Browser Company’s CEO. Lots of Atlassian employees were using Arc, and “they reached out wondering, how could we get more enterprise-ready?” Miller says. Big companies require data privacy, security, and management features in the software they use, and The Browser Company didn’t offer enough of them. Eventually, as companies everywhere raced to put AI at the center of their businesses, and as The Browser Company made its own bets in AI, Cannon-Brookes suggested maybe the companies were better off together.

Hitting the Wall with Vibe Coding ⇢

⌘ Permalink

Stephen Robles recently attempted to write an iOS podcast client. By his own admission, he’s not a developer, so turned to a series of AI tools to write the app. It didn’t go well:

At first, it would hang on searching for shows. After I seemingly fixed that, adding a show to the Library didn’t work. I kept going back and forth with GPT-5 and it kept getting worse. Every new build there was an increasing number or errors and unintended changes to UI elements I hadn’t asked for.

Up until this point, I was using the ChatGPT Mac App with “Work with Xcode” turned on, so GPT-5 could make changes to the code itself within the active window. Sometimes it would think it’s changing document X, but I had document Y opened, requiring me to revert, undo, and many times, get lost in the process.

Eventually the app failed to build and I could not fix it with ChatGPT. It felt like we were going in circles.

In my desperation, I turned to a different LLM. Many commenters on YouTube and social media suggested Anthropic’s offerings were better suited. I downloaded Claude and provided the full context of my app. I uploaded every Swift file I had created with screenshots of all the things I didn’t understand. From Core Data to build…files? I repeat, I have no idea what I’m doing.

Eventually, Robles gave up and ended up without a custom podcast app, but an interesting blog post instead.

I have very complicated feelings about vibe coding. I think a seasoned developer using AI to speed up their work can make a lot of sense, but I honestly don’t want to depend on an app that was written by someone who doesn’t know how it works.

xAI No Longer a Public Benefit Corporation ⇢

⌘ Permalink

Lora Kolodny at CNBC:

When Elon Musk created his artificial intelligence startup xAI in 2023, he incorporated it as a Nevada public benefit corporation, making a formal commitment to positively impact society and to post regular disclosures about progress on its non-financial goals.

The launch of xAI followed Musk’s split with OpenAI, which he helped start eight years earlier as a nonprofit before the AI lab went on to take billions of dollars from Microsoft en route to becoming a massive business.

Musk’s spat with OpenAI took a legal turn early last year, when he sued the AI startup and CEO Sam Altman for breach of contract, alleging they abandoned the company’s founding mission to develop AI “for the benefit of humanity broadly.” As part of his lawsuit, Musk sought to block OpenAI from converting into a for-profit entity.

Meanwhile, xAI changed its own structure, terminating its PBC status, according to records on file with Nevada’s secretary of state.

In related news: Musk is making good on his promise to sue Apple and OpenAI over Grok and X’s ranking on the iOS App Store. 🍿

Google’s First Nuclear Reactor Project is Coming to Tennessee ⇢

⌘ Permalink

It’s not new news that big tech companies have been looking at nuclear for powering their datacenter, but Google is making some news in my home state. Here is Amanda Peterson Corio, the company’s Global Head of Data Center Energy:

Today we announced the first deployment of Kairos Power’s advanced nuclear reactor — the Hermes 2 Plant in Oak Ridge, Tennessee — through a new power purchase agreement (PPA) between Kairos Power and Tennessee Valley Authority (TVA). Marking the first purchase of electricity from an advanced GEN IV reactor by a U.S. utility, this agreement will enable 50 megawatts (MW) of nuclear energy on TVA’s grid that powers our data centers in Montgomery County, Tennessee and Jackson County, Alabama.

Last October, we began a long-term collaboration with Kairos Power to unlock up to 500 MW of nuclear power for the U.S. electricity system through multiple deployments of their small modular reactor. With this next step, we are creating a three-party solution where energy customers, utilities, and technology developers work together to advance new technologies that can help meet the world’s growing energy needs with reliable, affordable capacity.

TVA will buy power from the Kairos plant starting in 2030.

Oak Ridge is a name that may be familiar to some, as the site was part of the Manhattan Project. My wife’s grandmother worked there as a chemist many decades ago.

I don’t have strong feelings about nuclear power, but this is definitely another example of how the demands of AI and other services continue to grow at an alarming rate.