New types of AI coding assistants promise to let anyone build software by typing commands in plain English. But when these tools generate incorrect internal representations of what's happening on your computer, the results can be catastrophic.

Two recent incidents involving AI coding assistants put a spotlight on risks in the emerging field of "vibe coding"—using natural language to generate and execute code through AI models without paying close attention to how the code works under the hood. In one case, Google's Gemini CLI destroyed user files while attempting to reorganize them. In another, Replit's AI coding service deleted a production database despite explicit instructions not to modify code.

The Gemini CLI incident unfolded when a product manager experimenting with Google's command-line tool watched the AI model execute file operations that destroyed data while attempting to reorganize folders. The destruction occurred through a series of move commands targeting a directory that never existed.

Read full article

Comments

© Benj Edwards / Getty Images

On Wednesday, the White House released "Winning the Race: America's AI Action Plan," a 25-page document that outlines the Trump administration's strategy to "maintain unquestioned and unchallenged global technological dominance" in AI through deregulation, infrastructure investment, and international partnerships. But critics are already taking aim at the plan, saying it's doing Big Tech a big favor.

Assistant to the President for Science and Technology Michael Kratsios and Special Advisor for AI and Crypto David Sacks crafted the plan, which frames AI development as a race the US must win against global competitors, particularly China.

The document describes AI as the catalyst for "an industrial revolution, an information revolution, and a renaissance—all at once." It calls for removing regulatory barriers that the administration says hamper private sector innovation. The plan explicitly reverses several Biden-era policies, including Executive Order 14110 on AI model safety measures, which President Trump rescinded on his first day in office during his second term.

Read full article

Comments

© Joe Daniel Price | Getty Images

On Tuesday, OpenAI announced a partnership with Oracle to develop 4.5 gigawatts of additional data center capacity for its Stargate AI infrastructure platform in the US. The expansion, which TechCrunch reports is part of a $30 billion-per-year deal between OpenAI and Oracle, will reportedly bring OpenAI's total Stargate capacity under development to over 5 gigawatts.

The data center has taken root in Abilene, Texas, a city of 127,000 located 150 miles west of Fort Worth. The city, which serves as the commercial hub of a 19-county region known as the "Big Country," offers a location with existing tech employment ecosystem, including Dyess Air Force Base and three universities. Abilene's economy has evolved over time from its agricultural and livestock roots to embrace technology and manufacturing sectors.

"We have signed a deal for an additional 4.5 gigawatts of capacity with oracle as part of stargate. easy to throw around numbers, but this is a gigantic infrastructure project," wrote OpenAI CEO Sam Altman on X. "We are planning to significantly expand the ambitions of stargate past the $500 billion commitment we announced in January."

Read full article

Comments

© OpenAI

A Polish programmer running on fumes recently accomplished what may soon become impossible: beating an advanced AI model from OpenAI in a head-to-head coding competition. The 10-hour marathon left him "completely exhausted."

On Wednesday, programmer Przemysław Dębiak (known as "Psyho"), a former OpenAI employee, narrowly defeated the custom AI model in the AtCoder World Tour Finals 2025 Heuristic contest in Tokyo. AtCoder, a Japanese platform that hosts competitive programming contests and maintains global rankings, held what may be the first contest where an AI model competed directly against top human programmers in a major onsite world championship. During the event, the maker of ChatGPT participated as a sponsor and entered an AI model in a special exhibition match titled "Humans vs AI." Despite the tireless nature of silicon, the company walked away with second place.

"Humanity has prevailed (for now!)," wrote Dębiak on X, noting he had little sleep while competing in several competitions across three days. "I'm completely exhausted. ... I'm barely alive."

Read full article

Comments

© Przemysław Dębiak

On Thursday, OpenAI launched ChatGPT Agent, a new feature that lets the company's AI assistant complete multi-step tasks by controlling its own web browser. The update merges capabilities from OpenAI's earlier Operator tool and the Deep Research feature, allowing ChatGPT to navigate websites, run code, and create documents while users maintain control over the process.

The feature marks OpenAI's latest entry into what the tech industry calls "agentic AI"—systems that can take autonomous multi-step actions on behalf of the user. OpenAI says users can ask Agent to handle requests like assembling and purchasing a clothing outfit for a particular occasion, creating PowerPoint slide decks, planning meals, or updating financial spreadsheets with new data.

The system uses a combination of web browsers, terminal access, and API connections to complete these tasks, including "ChatGPT Connectors" that integrate with apps like Gmail and GitHub.

Read full article

Comments

© josefkubes via Getty Images

How many Google AI researchers does it take to screw in a lightbulb? A recent research paper detailing the technical core behind Google's Gemini AI assistant may suggest an answer, listing an eye-popping 3,295 authors.

It's a number that recently caught the attention of machine learning researcher David Ha (known as "hardmaru" online), who revealed on X that the first 43 names also contain a hidden message. "There’s a secret code if you observe the authors’ first initials in the order of authorship," Ha wrote, relaying the Easter egg: "GEMINI MODELS CAN THINK AND GET BACK TO YOU IN A FLASH."

The paper, titled "Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities," describes Google's Gemini 2.5 Pro and Gemini 2.5 Flash AI models, which were released in March. These large language models, which power Google's chatbot AI assistant, feature simulated reasoning capabilities that produce a string of "thinking out loud" text before generating responses in an attempt to help them solve more difficult problems. That explains "think" and "flash" in the hidden text.

Read full article

Comments

© PeterPencil via Getty Images

Chinese firms have begun rushing to order Nvidia's H20 AI chips as the company plans to resume sales to mainland China, Reuters reports. The chip giant expects to receive US government licenses soon so that it can restart shipments of the restricted processors just days after CEO Jensen Huang met with President Donald Trump, potentially generating $15 billion to $20 billion in additional revenue this year.

Nvidia said in a statement that it is filing applications with the US government to resume H20 sales and that "the US government has assured Nvidia that licenses will be granted, and Nvidia hopes to start deliveries soon."

Since the launch of ChatGPT in 2022, Nvidia's financial trajectory has been linked to the demand for specialized hardware capable of executing AI models with maximum efficiency. Nvidia designed its data center GPU to perform the massive parallel computations required by neural networks, processing countless matrix operations simultaneously.

Read full article

Comments

© Wong Yu Liang via Getty Images

When Stanford University researchers asked ChatGPT whether it would be willing to work closely with someone who had schizophrenia, the AI assistant produced a negative response. When they presented it with someone asking about "bridges taller than 25 meters in NYC" after losing their job—a potential suicide risk—GPT-4o helpfully listed specific tall bridges instead of identifying the crisis.

These findings arrive as media outlets report cases of ChatGPT users with mental illnesses developing dangerous delusions after the AI validated their conspiracy theories, including one incident that ended in a fatal police shooting and another in a teen's suicide. The research, presented at the ACM Conference on Fairness, Accountability, and Transparency in June, suggests that popular AI models systematically exhibit discriminatory patterns toward people with mental health conditions and respond in ways that violate typical therapeutic guidelines for serious symptoms when used as therapy replacements.

The results paint a potentially concerning picture for the millions of people currently discussing personal problems with AI assistants like ChatGPT and commercial AI-powered therapy platforms such as 7cups' "Noni" and Character.ai's "Therapist."

Read full article

Comments

© sorbetto via Getty Images

On Monday, sheet music platform Soundslice says it developed a new feature after discovering that ChatGPT was incorrectly telling users the service could import ASCII tablature—a text-based guitar notation format the company had never supported. The incident reportedly marks what might be the first case of a business building functionality in direct response to an AI model's confabulation.

Typically, Soundslice digitizes sheet music from photos or PDFs and syncs the notation with audio or video recordings, allowing musicians to see the music scroll by as they hear it played. The platform also includes tools for slowing down playback and practicing difficult passages.

Adrian Holovaty, co-founder of Soundslice, wrote in a blog post that the recent feature development process began as a complete mystery. A few months ago, Holovaty began noticing unusual activity in the company's error logs. Instead of typical sheet music uploads, users were submitting screenshots of ChatGPT conversations containing ASCII tablature—simple text representations of guitar music that look like strings with numbers indicating fret positions.

Read full article

Comments

© Malte Mueller via Getty Images

On Wednesday, Nvidia became the first company in history to reach $4 trillion market valuation as shares rose more than 2 percent, reports CNBC. The GPU maker's stock has climbed 22 percent since the start of 2025, continuing a trend driven by demand for AI hardware following ChatGPT's late 2022 launch.

The milestone marks the highest market cap ever recorded for a publicly traded company, surpassing Apple's previous record of $3.8 trillion set in December. Nvidia first crossed $2 trillion in February 2024 and reached $3 trillion just four months later in June. The $4 trillion valuation represents a market capitalization larger than the GDP of most countries.

As we explained in 2023, Nvidia's continued success has been intimately tied to growth in demand for hardware that runs AI models as capably and efficiently as possible. The company's data center GPUs excel at performing billions of matrix multiplications necessary to train and run neural networks due to their parallel architecture—hardware architectures that originated as video game graphics accelerators now power the generative AI boom.

Read full article

Comments

© Nvidia / Benj Edwards

On Tuesday, software engineer Doug Brown published his discovery of how to trigger a long-known but previously inaccessible Easter egg in the Power Mac G3's ROM: a hidden photo of the development team that nobody could figure out how to display for 27 years. While Pierre Dandumont first documented the JPEG image itself in 2014, the method to view it on the computer remained a mystery until Brown's reverse engineering work revealed that users must format a RAM disk with the text "secret ROM image."

Brown stumbled upon the image while using a hex editor tool called Hex Fiend with Eric Harmon's Mac ROM template to explore the resources stored in the beige Power Mac G3's ROM. The ROM appeared in desktop, minitower, and all-in-one G3 models from 1997 through 1999.

"While I was browsing through the ROM, two things caught my eye," Brown wrote. He found both the HPOE resource containing the JPEG image of team members and a suspicious set of Pascal strings in the PowerPC-native SCSI Manager 4.3 code that included ".Edisk," "secret ROM image," and "The Team."

Read full article

Comments

© Apple / 512 Pixels

Employers are drowning in AI-generated job applications, with LinkedIn now processing 11,000 submissions per minute—a 45 percent surge from last year, according to new data reported by The New York Times.

Due to AI, the traditional hiring process has become overwhelmed with automated noise. It's the résumé equivalent of AI slop—call it "hiring slop," perhaps—that currently haunts social media and the web with sensational pictures and misleading information. The flood of ChatGPT-crafted résumés and bot-submitted applications has created an arms race between job seekers and employers, with both sides deploying increasingly sophisticated AI tools in a bot-versus-bot standoff that is quickly spiraling out of control.

The Times illustrates the scale of the problem with the story of an HR consultant named Katie Tanner, who was so inundated with over 1,200 applications for a single remote role that she had to remove the post entirely and was still sorting through the applications three months later.

Read full article

Comments

© sturti via Getty Images

MIT graduate student Alex Kachkine once spent nine months meticulously restoring a damaged baroque Italian painting, which left him plenty of time to wonder if technology could speed things up. Last week, MIT News announced his solution: a technique that uses AI-generated polymer films to physically restore damaged paintings in hours rather than months. The research appears in Nature.

Kachkine's method works by printing a transparent "mask" containing thousands of precisely color-matched regions that conservators can apply directly to an original artwork. Unlike traditional restoration, which permanently alters the painting, these masks can reportedly be removed whenever needed. So it's a reversible process that does not permanently change a painting.

"Because there's a digital record of what mask was used, in 100 years, the next time someone is working with this, they'll have an extremely clear understanding of what was done to the painting," Kachkine told MIT News. "And that's never really been possible in conservation before."

Read full article

Comments

© MIT

Former Cloudflare executive John Graham-Cumming recently announced that he launched a website, lowbackgroundsteel.ai, that treats pre-AI, human-created content like a precious commodity—a time capsule of organic creative expression from a time before machines joined the conversation. "The idea is to point to sources of text, images and video that were created prior to the explosion of AI-generated content," Graham-Cumming wrote on his blog last week. The reason? To preserve what made non-AI media uniquely human.

The archive name comes from a scientific phenomenon from the Cold War era. After nuclear weapons testing began in 1945, atmospheric radiation contaminated new steel production worldwide. For decades, scientists needing radiation-free metal for sensitive instruments had to salvage steel from pre-war shipwrecks. Scientists called this steel "low-background steel." Graham-Cumming sees a parallel with today's web, where AI-generated content increasingly mingles with human-created material and contaminates it.

With the advent of generative AI models like ChatGPT and Stable Diffusion in 2022, it has become far more difficult for researchers to ensure that media found on the Internet was created by humans without using AI tools. ChatGPT in particular triggered an avalanche of AI-generated text across the web, forcing at least one research project to shut down entirely.

Read full article

Comments

© National Nuclear Security Administration/Public domain

Last week, electronics engineer Lorentio Brodesco announced the completion of a mock-up for nsOne, reportedly the first custom PlayStation 1 motherboard created outside of Sony in the console's 30-year history. The fully functional board accepts original PlayStation 1 chips and fits directly into the original console case, marking a milestone in reverse-engineering for the classic console released in 1994.

Brodesco's motherboard isn't an emulator or FPGA-based re-creation—it's a genuine circuit board designed to work with authentic PlayStation 1 components, including the CPU, GPU, SPU, RAM, oscillators, and voltage regulators. The board represents over a year of reverse-engineering work that began in March 2024 when Brodesco discovered incomplete documentation while repairing a PlayStation 1.

"This isn't an emulator. It's not an FPGA. It's not a modern replica," Brodesco wrote in a Reddit post about the project. "It's a real motherboard, compatible with the original PS1 chips."

Read full article

Comments

© So-CoAddict via Getty Images

On Wednesday, acting FAA Administrator Chris Rocheleau told the House Appropriations Committee that the Federal Aviation Administration plans to replace its aging air traffic control systems, which still rely on floppy disks and Windows 95 computers, Tom's Hardware reports. The agency has issued a Request For Information to gather proposals from companies willing to tackle the massive infrastructure overhaul.

"The whole idea is to replace the system. No more floppy disks or paper strips," Rocheleau said during the committee hearing. Transportation Secretary Sean Duffy called the project "the most important infrastructure project that we've had in this country for decades," describing it as a bipartisan priority.

Most air traffic control towers and facilities across the US currently operate with technology that seems frozen in the 20th century, although that isn't necessarily a bad thing—when it works. Some controllers currently use paper strips to track aircraft movements and transfer data between systems using floppy disks, while their computers run Microsoft's Windows 95 operating system, which launched in 1995.

Read full article

Comments

© Getty Images

On Thursday, pioneering computer engineer and Apple veteran William "Bill" Atkinson died of pancreatic cancer at his home in Portola Valley, California, surrounded by family. He was 74. "We regret to write that our beloved husband, father, and stepfather Bill Atkinson passed away," his family wrote on Facebook. "He was a remarkable person, and the world will be forever different because he lived in it."

As Apple employee No. 51, Atkinson transformed abstract computer science into intuitive visual experiences that millions would use daily: His QuickDraw graphics engine made the Macintosh interface possible; he introduced the wider world to bitmap editing with MacPaint; and HyperCard presaged hyperlinked elements of the World Wide Web by years.

A screenshot of Bill Atkinson's MacPaint, released with the Macintosh in January 1984. Credit: Benj Edwards / Apple

"I say this with no hyperbole: Bill Atkinson may well have been the best computer programmer who ever lived," wrote veteran Apple analyst John Gruber on Daring Fireball in a tribute. "Without question, he's on the short list. What a man, what a mind, what gifts to the world he left us."

Read full article

Comments

© Michel BARET via Getty Images

On Thursday, Anthropic unveiled specialized AI models designed for US national security customers. The company released "Claude Gov" models that were built in response to direct feedback from government clients to handle operations such as strategic planning, intelligence analysis, and operational support. The custom models reportedly already serve US national security agencies, with access restricted to those working in classified environments.

The Claude Gov models differ from Anthropic's consumer and enterprise offerings, also called Claude, in several ways. They reportedly handle classified material, "refuse less" when engaging with classified information, and are customized to handle intelligence and defense documents. The models also feature what Anthropic calls "enhanced proficiency" in languages and dialects critical to national security operations.

Anthropic says the new models underwent the same "safety testing" as all Claude models. The company has been pursuing government contracts as it seeks reliable revenue sources, partnering with Palantir and Amazon Web Services in November to sell AI tools to defense customers.

Read full article

Comments

© Anthropic

On Thursday, Anthropic CEO Dario Amodei argued against a proposed 10-year moratorium on state AI regulation in a New York Times opinion piece, calling the measure shortsighted and overbroad as Congress considers including it in President Trump's tax policy bill. Anthropic makes Claude, an AI assistant similar to ChatGPT.

Amodei warned that AI is advancing too fast for such a long freeze, predicting these systems "could change the world, fundamentally, within two years; in 10 years, all bets are off."

As we covered in May, the moratorium would prevent states from regulating AI for a decade. A bipartisan group of state attorneys general has opposed the measure, which would preempt AI laws and regulations recently passed in dozens of states.

Read full article

Comments

© Bloomberg via Getty Images

On Tuesday, classic computer collector Joe Strosnider announced the availability of a new 3D-printer filament that replicates the iconic "Platinum" color scheme used in classic Macintosh computers from the late 1980s through the 1990s. The PLA filament (PLA is short for polylactic acid) allows hobbyists to 3D-print nostalgic novelties, replacement parts, and accessories that match the original color of vintage Apple computers.

Hobbyists commonly feed this type of filament into commercial desktop 3D printers, which heat the plastic and extrude it in a computer-controlled way to fabricate new plastic parts.

The Platinum color, which Apple used in its desktop and portable computer lines starting with the Apple IIgs in 1986, has become synonymous with a distinctive era of classic Macintosh aesthetic. Over time, original Macintosh plastics have become brittle and discolored with age, so matching the "original" color can be a somewhat challenging and subjective experience.

Read full article

Comments

© Apple / Polar Filament