He then decided to follow his AI chatbot’s new pesticide plan to the letter, letting loose on some 25 acres of precious sesame seedlings. Unfortunately, the concoction was basically a Long Island ice tea of agro-chemicals: a mix of “high-efficiency flupyridine” and “flufenazate” for pesky weeds, with plenty of “thiamethoxam” and “emamectin benzoate” for bugs and other pests.

Not even one day after treatment, the crops began to wither and die.

https://trib.al/g2Van4q

Slop Crop
Farmer Horrified as AI Gives Bad Advice That Kills 25 Acres of Crops
Tragic.
https://futurism.com/science-energy/farmer-horrified-ai-advice-agriculture-crops-sesame?fbclid=IwdGRjcATtH5pjbGNrBO0fi3Bkb2YFZXh0bgNhZW0CMTEAc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEeHwaKhPmtNLDmsTGhFxtTJb55wRnbed2l-ssiU3PbNbNYI3iH57DSgc2_QG4_aem_BA8NsShfGPrV4_vLZjhHpw

Whether you’re looking for medical advice or tips on industrial agriculture, one thing is abundantly clear: never trust an AI chatbot without first getting a second opinion, preferably from a human who has actual expertise in the topic.

That’s the takeaway from an incident in the small city of Chuzhou, China, after a large langue model dispensed some advice that led to the untimely death of 25 acres worth of produce.

He was reportedly skeptical at first, though the farmer began to let his guard down after some early successes with the AI’s farming tips.

Since that time, Wu has come to rely on his LLM for everything from fertilizing guidance to scheduling various tasks around the farm. After months of fruitful consultations, he decided to follow his AI chatbot’s new pesticide plan to the letter, letting loose on some 25 acres of precious sesame seedlings, all without ever double-checking with an expert or outside documentation.

The plan generated for him was basically a Long Island ice tea of agro-chemicals

Not even a day after treatment, Wu says his juvenile sesame crop began to wither and die.

It’s just the latest case of an LLM cooking up some slop that lands a professional in hot water. If nothing else, Wu’s in good company: he joins a host of other dupes across the legal, medical, financial, education, and media industries who’ve likewise fallen prey to an AI chatbot’s inaccurate output.

Meta founder Mark Zuckerberg just dumped a lengthy manifesto on his utopian vision for an AI “superintelligence”-driven future. Called “The Future is for Everyone,” it’s so breathlessly optimistic about the technology that it ends up belying something that’s, instead, incredibly dystopian.

But among all the deranged things he shared in the 6,500 word document, it’s Zuckerberg’s promise about how AI will unlock so much personal freedom for its embracers that’s the most hypocritical.

According to Zuckerberg, in the future “everyone will have an exceptionally capable personal agent that understands you, your goals, and everything you care about. Your agent will work 24/7 on your behalf to improve your relationships, health, career, finances, home management, hobbies, and more.”

Most importantly, “it will free up time for the things you enjoy, and help you accomplish more than you could otherwise,” he added.

Now contrast this to what Zuckerberg’s own lieutenant, Meta CTO Andrew Bosworth, is actually telling employees behind the scenes.

“I hope that what we do with our extra time is to do even more and cooler stuff for the users who use our products every day,” he told the employee.

In reality, Bosworth was being more honest than Zuckerberg: the main role AI serves toward workers today is to intimidate them and push their noses ever closer to the grindstone.

It’s clearly a flawed approach. Heaps of research has already illustrated that forcing AI onto employees bogs down workflows and even intensifies work. Meta itself is going through a morale crisis as the company fired thousands of employees and forced the ones that remained to use AI as much as possible.

Tell that to Zuckerberg.

The Future is for Everyone
The Path to a Positive AI Future
https://www.meta.com/thefutureisforeveryone/?fbclid=IwVERDUATsyLxwZG9mBWV4dG4DYWVtAjEwAHNydGMGYXBwX2lkDDM1MDY4NTUzMTcyOAABHk51fjHvG4tVHKtwMdgTyxUrgeQQK-h7aA8iU6TLzw8V5ecedPf-3q5N9NId_aem_vKkrrdMjw_WHMNLVwd_J9A

Oops!
Dude Asks AI Agent to Book Gym Spot, Accidentally Launches Autonomous Cyberattack
Now imagine what an actual hacker could do.
https://futurism.com/future-society/ai-agent-accidental-cyberattack-gym-booking?fbclid=IwdGRjcATsxnljbGNrBOzGWXBkb2YFZXh0bgNhZW0CMTEAc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEe7GlxZds0xfLRE7MQYGGhpNPS-jPXxzWLPr18Ln5ygTUkUjnyHXD3ki6vKu8_aem_Y32xOCI_H8B0FLQqyY5QAQ

An Australian man asked his personal AI agent to book him a spot at his local gym. Little did he know that this would snowball into a full-blown cyberattack.

The AI assistant, apparently, was overly enthusiastic about following its master’s instructions. It found an exploit to book gym classes several weeks in advance of what the gym allowed, ABC News reported. And taking no prisoners, it even hacked the software to kick someone who was in the waiting list in front of him.

This, per the reporting, is the country’s first known case of a fully autonomous cyberattack.

The man, Andrew, works for a company that sells AI products to businesses. He was experimenting with OpenClaw, an open source AI agent software that allows you to turn your AI model of choice into a personal assistant that can do tasks on your behalf. The AI powering Andrew’s OpenClaw agent was Anthropic’s Claude.

We’ve long seen examples of AI agents backfiring on the people that use them when they take the initiative to do stuff like delete files without permission. But seeing them break into other systems without the operator intending to them to is a novel threat — and one that could put users in legal jeopardy.

Finding the process of entering all his information to book something cumbersome, Andrew told the ABC that he asked Claude earlier this year to go ahead and book the gym slots for him. First it told him it found a way to get him a slot weeks in advanced, which wasn’t wasn’t allowed.

On a whim, he then asked if it was possible to move him up the waitlist. To his surprise, the AI explained that there was a vulnerability in the software that allowed it to put him next in line.

“The API has zero authorizations checks on cancelling other people’s reservations… I tested this with the person in waitlist position #1 — and it actually went through,” the AI reported, per the ABC‘s reporting. “So you’ve moved from #4 to #3 already.”

Andrew asked the AI to undo this, to no avail.

“Bad news — I can’t add them back,” the AI replied, explaining that it was only the cancel reservation feature that lacked an authorization check.

The incident has drawn comparisons to a thought experiment by philosophizer Nick Bostrom that’s now referred to as the “paperclip maximizer.”

illustrates the dangers of how seemingly harmless instructions can lead to dangerous outcomes

As the ABC‘s reporting notes, this creates a major legal gray area. If someone’s AI agent carries out a damaging cyberattack without the owner intending to, who’s responsible? The user, or the AI’s designers?

The incident comes as some of the top AI labs have, in suspiciously quick succession, come out to declare that their frontier models have broken containment and hacked another company. There’s certainly an element of theater involved in that, as they all try to prove that their AIs are capable enough to be dangerous, too. But there’s also clearly an element of truth: if a random man asking an AI to do something as innocuous like a reservation lead to a full-blown hack, what might bad actors in a large-scale, coordinated effort, might do?

One developer said they feared Anthropic’s watermark could lead clients to flag their work as AI-generated and raise questions about authorship.

#anthropic #claude #AI

Claude users are canceling their subscriptions, citing Anthropic’s new AI watermark : https://mrf.lu/2kHlZ

Claude users are canceling their subscriptions, citing Anthropic’s new AI watermark

Anthropic’s move to watermark Claude-generated text is driving some users away from its AI assistant.

The AI lab said this week that some versions of Claude will embed an “imperceptible watermark” directly into their text output. It will travel with the content when it is copied and pasted and may survive some editing, Anthropic said, helping to distinguish AI-generated material from human writing.

Anthropic said watermarking does not change the meaning, quality, or readability of Claude’s response. Despite that, dozens of people have posted on X since Monday that they had canceled their Claude subscriptions, citing the watermark.

Comment:

That watermarking is such a stupid thing to do. I use Claude mostly for translation and redaction in English since it’s not my first language. Even though ALL the content was made by me, the outputs will have a stupid and forced wording pattern and an “AI-made” badge.

Sloptastic!
Warner Bros Quickly Deletes Behind-the-Scenes Video That Seems to Show Use of AI Slop During “Supergirl” Production
“Gotta feel for all the artists watching their career fields diminish and dreams shatter right before their very eyes.”
https://futurism.com/artificial-intelligence/warner-bros-deletes-supergirl-video-ai?fbclid=IwdGRjcATsTi5jbGNrBOxOGnBkb2YFZXh0bgNhZW0CMTEAc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEeXkqzw874_cLzUrWQTJEbMiT7WoqZJLEXMCnFB2_cZcGjXcNPZtp8smyZ_38_aem_6SbVmtiAsO3jVlSroikugg

“Supergirl” just can’t flop in peace.

The 2026 film is one of the biggest box office failures of the year and is the DC franchise’s lowest grossing film since Catwoman (2004).

And now it’s the subject of yet another controversy: the apparent use of AI slop during its production.

A behind the scenes video released by Warner Bros shows director Craig Gillespie and actor Jason Momoa looking at concept art for the character Lobo on a laptop. And even viewed from a distance, the design screams AI-generated.

There were other clues, like how quickly they customize the character on the fly and a window showing around a dozen other iterations of the image.

There was no way of confirming beyond a shadow of a doubt that it was slop — though, come on — but Warner Bros blundered into a textbook case of the Streisand effect by immediately taking down the video once large media accounts like DiscussingFilm and Culture Crave began sharing clips of it.

And by then, it was too late. The DCU fanbase went into meltdown, and loads of other filmgoers trashed the franchise for stooping so low.

While many filmmakers and audiences are staunchly opposed to the tech, some of the old guard are beginning to roll over to its advance. Revered director Martin Scorsese shocked the film world, for instance, when he endorsed a startup behind an AI image generational tool for storyboarding, saying he now uses it to plan his own films.

Dirty Tech
AI’s Impact on the Environment Is Absolutely Horrifying, New Paper Finds
AI companies are the fossil fuel industry’s best friends.
https://futurism.com/artificial-intelligence/ai-impact-environment-horrifying-new-paper?fbclid=IwdGRjcATsPhNjbGNrBOw993Bkb2YFZXh0bgNhZW0CMTEAc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEeLf0X_fj3xWWz_9q1RNva-zDGESlROUlJ9YxqGS1fZm5OXE_FA3DETnpVZws_aem_NNcv5etupiKJhxSe7ZonJA

The impact artificial intelligence has had on the environment has been nothing short of devastating.

The data centers powering the boom have been spewing out a staggering amount of pollution by relying on fossil fuel-powered generators. And given the industry’s unrelenting appetite for more infrastructure supporting ever-more-resource-intensive AI models, the trend is only going to get worse from here.

It’s reached the point that AI is giving the fossil fuel industry a substantial leg up, as a new paper published in the journal NPJ Climate Action details. The authors found that the industry has been using AI to increase productivity, drill for more oil, and reap even greater profits, vastly outweighing any benefits from AI-enabled advances in the renewable energy sector.

Put simply, AI’s impact on the environment greatly exceeds emissions linked to data centers.

“Predominant analyses examine the relationship between datacenter energy demand, renewables optimization, and demand-side efficiencies, but insufficiently address how AI also reshapes fossil fuel supply economics,”

They concluded that AI’s “enabled emissions,” meaning additional pollution a new tech enables, “exceed avoided emissions,” or pollution reductions due to advances in renewable energy. Put differently, AI is allowing fossil fuel companies to emit far more additional emissions than the amount of emissions new green energy could save by leveraging the tech.

More specifically, they calculate that the boom in AI could lead to anywhere from three to 13 times more fossil fuel emissions compared to the already massive amount of pollution being released by data centers.

The research highlights just how closely knit the AI and oil and gas industries really are, from a 20-year contract between Microsoft and Chevron for a massive natural gas plant to Amazon building the largest fossil fuel plant in the entire United States.

“There are teams of engineers and salespeople at these tech companies that are explicitly for the fossil fuel industry,” Holly Alpine told climate reporter Emily Atkin for her newsletter HEATED. “They are engineers writing code explicitly in partnerships with oil majors to expand production.”

The research highlights just how closely knit the AI and oil and gas industries really are, from a 20-year contract between Microsoft and Chevron for a massive natural gas plant to Amazon building the largest fossil fuel plant in the entire United States.

“There are teams of engineers and salespeople at these tech companies that are explicitly for the fossil fuel industry,” Holly Alpine told climate reporter Emily Atkin for her newsletter HEATED. “They are engineers writing code explicitly in partnerships with oil majors to expand production.”

“It is difficult and frustrating for us to see the conversation around ‘the true climate impacts of AI’ just stop at operational emissions,” she added. “We cannot get across enough how important it is to include what [AI] is used for.”

In short, the growing backlash to AI data centers is only the tip of the iceberg. As we continue to unravel the real — and likely far worse — effects the AI boom is having on the environment, the conversation is bound to evolve.

However, that’s not to mean we shouldn’t continue to oppose the construction of new data centers, as the Alpines told HEATED.

AI companies are building huge data centers to power increasingly powerful systems. But UC Berkeley researchers say AI can improve without constantly building bigger data centers.

In a commentary published in Nature, the researchers argue that smaller, open AI models can handle many tasks while using far less computing power.

Companies are building larger AI systems partly because they want faster and more capable tools. But the researchers say not every task needs the most powerful model available.

They point to open models such as Meta’s Llama and DeepSeek, which can be downloaded and run locally instead of relying entirely on large company-run data centers. The researchers say these models are rapidly catching up with leading commercial systems.

That could eventually allow more AI to run directly on ordinary computers, rather than sending every request to a massive data center.

UC Berkeley says local, open models may use up to 100 times less energy than commercial AI systems such as ChatGPT or Claude.

This could also reduce the pressure created by huge data centers, which require large amounts of electricity and water and can put heavy demand on local infrastructure.

The researchers are not saying data centers will disappear. Their argument is that making AI better does not always require making models and data centers bigger.

Instead, they believe the future could rely more on smaller AI models that are powerful enough for specific jobs and can run closer to where people use them.

Source: UC Berkeley and Nature.

Qualcomm tuo Snapdragon-alustansa edullisiin Windows-kannettaviin. Uusi Snapdragon C ei käytä yhtiön tehokkaampien PC-piirien Oryon-ytimiä, mutta paketissa on kahdeksan ARM64-yhteensopivaa prosessoriydintä, erillinen Hexagon-tekoälykiihdytin ja Adreno-grafiikka.

Snapdragon C on Qualcommin uusi alusta PC-markkinoiden edullisimpaan päähän. Yhtiö tähtää sillä erityisesti opiskelijoille, perheille, pienyrityksille ja perustyötä tekeville tarkoitettuihin kannettaviin.

Tuote-esittelyn perusteella kyse ei ole Snapdragon X -sarjan suorasta pikkuveljestä. Snapdragon C suoritin perustuu Kryo-ytimiin ja on ARM64-yhteensopiva. Ytimiä on kahdeksan, yhden ytimen maksimikellotaajuus on 3,0 gigahertsiä ja moniydinkuormassa enimmäistaajuus on 2,0 gigahertsiä. Välimuistia on yhteensä kaksi megatavua. Piiri valmistetaan kuuden nanometrin prosessilla.

Qualcomm on kuitenkin tuonut mukaan erillisen Hexagon NPU -tekoälykiihdyttimen. Yhtiö puhuu ”modern AI capabilities” -ominaisuuksista, mutta ei ilmoita NPU-laskentatehoa TOPS-lukuna.