Today, Instagram unveiled a refresh of its wordmark, accompanied by the usual bland corporate platitudes these kinds of announcements are always packaged with.

Head of Instagram Adam Mosseri called it “leaner and more modern, with references to the original and the simplicity and craft that’s always made it Instagram.”

Sure, Adam. Simplicity and craft are definitely what everyone thinks of when Instagram comes to mind.

Read full article

Comments

Google is announcing a new Gemini model today, but it's not the long-awaited 3.5 Pro. Gemini 3.7 Flash is now rolling out to replace 3.6 Flash, which itself was released only three weeks ago . This new "workhorse" model is supposedly the product of core optimizations and developer feedback, offering improved coding and agentic performance. And Google is hoping to counter the lower cost of some competing models with a lower "introductory price" for 3.7 Flash.

According to Senior Director Tulsee Doshi, Gemini 3.7 Flash is noticeably better at coding than the previous Flash release. She cites a jump in the FrontierCode 1.1 Main test from 34.4 to 43.6 percent and DeepSWE v1.1 going from 49 to 65.3 percent. As for the vibes, Gemini 3.7 Flash's WebDev Arena score has risen to 1,588 from 1,538.

People turning to Gemini and hoping it will "know" things may also see modest improvements in Gemini 3.7 Flash. The GDP.pdf benchmark, which measures how well a model can process complex documents, has gone up to 34 percent versus 22 percent with 3.6 Flash. AutomationBench tests how well models can execute common business workflows, and Gemini 3.7 Flash rose to 30.4 percent from 3.6's 17 percent score.

Read full article

Comments

It's time for Rivian's changeover from model year 2026 to 2027, and the electric vehicle startup has decided to take a page from its recently released R2 SUV and apply it to the company's larger, more expensive R1 line. The changes aren't massive—the R1 got a refresh back in 2024 that saw the R1S SUV and R1T pickup each get some comprehensive re-engineering with new powertrain options, a new electronic architecture, and new batteries .

But among them is the No. 1 most-requested feature by Rivian R1S customers, the automaker told Ars: captain's chairs for the second row. The six-seat interior is an alternative to a middle bench, and the Rivian captain's chair looks thoughtfully designed, with a hidden cupholder that I'm not sure I've seen on any other SUV before. The chairs are also spaced so it's easy to get to the third row, or use that volume between them for skis or snowboards or surfboards.

Rivian has also changed the trims, bringing them into line with the offer structure on the recently released R2 . Instead of choosing from Dual, Tri, or R1s, Rivian is now calling them Premium and Performance; Quad stays Quad. But the base price for each stays the same: The R1T still starts at $79,990 and the R1S at $83,990.

Read full article

Comments

Anthropic investors expect the AI startup to float at a valuation of $2 trillion or more in October, a dizzying figure that would eclipse SpaceX and make the AI lab’s debut the largest-ever initial public offering.

Half a dozen of the company’s backers told the FT that Anthropic’s rapidly rising revenue would enable it to more than double its current valuation in a planned autumn float.

A listing at that level could unlock billions of dollars in gains for the five-year-old company’s early investors but would also test public markets that are growing more nervous about the AI boom.

Read full article

Comments

Solar storms, like the one in August 1972 that hit during the gap between the Apollo 16 and Apollo 17 missions, throw bursts of protons intense enough to raise an astronaut's cancer risk or even cause radiation sickness. Earth's atmosphere and magnetic field absorb this radiation, but crews heading to the Moon or Mars won't have that protection, and no spacecraft built so far has enough shielding to stop it.

A team led by Jordan Houri and Oren Milstein of StemRad, an Israeli-American startup developing personal protective equipment against radiation, proposed we could solve this by shielding the astronauts instead of shielding the spacecraft.

To test this idea, StemRad’s team flew a wearable radiation-shielding vest called AstroRad to the Moon and back aboard NASA's uncrewed Artemis I mission, then used the flight data to calculate how it would perform during an actual solar storm. It turns out the vest would perform roughly as well as the Orion’s heavily shielded onboard shelter the crew was supposed to hide in to wait out a storm.

Read full article

Comments

Anthropic has revealed that it will soon watermark content that is processed ( not just generated! ) by any of its models. In a support article , Anthropic explained that it was rolling out machine-readable watermarks to comply with the European Union’s AI Act, which requires all AI system providers to watermark AI-generated or manipulated audio, image, text, and video outputs. The law applies to any AI model released after August 2 and provides a grace period until December 2026 for providers to update previously released models.

Anthropic confirmed that moving forward, all new models offered globally—not just in the EU—will mark AI-generated content “from day one.” Text outputs will “carry embedded watermarks,” invisible to the user, and other “generated files will include digitally signed provenance metadata where supported,” Anthropic said.

Notably, Anthropic is deploying a " nuke it from orbit " approach, applying the watermarks to all processed content where supported, even though the EU does not require it for cases where an AI system performs "an assistive function for standard editing" (the guidance's own example is grammar correction), or where it doesn't "substantially alter" the user's text or its meaning. A watermark applied at the model level can't tell wholesale generation from a comma fix, so Claude may end up stamping exactly the content the law was written to leave alone. How thoroughly it truly watermarks will not be known until Anthropic releases a detection tool that can be tested. The company said that it plans to eventually share details on how to detect marks in order to offer technical support that the EU’s law requires.

Read full article

Comments

A healthy toddler in Washington state fell ill abruptly. It began with vomiting, lethargy, and weakness on his right side. Then it progressed to a headache, fever, and trouble swallowing. At one point, he briefly became unresponsive, and he had episodes where he stopped breathing. Doctors admitted him to the hospital, but his condition only deteriorated.

The boy was infected with an ultra-rare amoeba known for ravaging the brain. The opportunistic pathogen was first identified in 1986 from the brain of a San Diego Zoo monkey, which died from the infection. Since then, only about 200 human cases have been recorded worldwide. About 90 percent of them have been fatal.

The initial symptoms can be vague and, given its rarity, the infection is notorious for being undiagnosed or misdiagnosed. Still, the scant medical literature does provide some telltale signs, clues of what it can look like on tests and scans. But for the toddler, the infection looked different, sending doctors down the wrong path.

Read full article

Comments

AI companies' penchant for scraping through large swathes of the public web in search of valuable training data has already led to lawsuits and technical fixes aimed at stopping the practice. Now, a pair of designers are hoping to stymie these scrapers with a new font designed to offer people a perfectly readable webpage while serving scrapers a subtly edited, nonsensical version in the underlying HTML.

ShieldFont, as designers Isaque Seneda and Gabriel Abrucio write in a recent white paper , was made to offer web publishers "a practical opt-out from unauthorized AI training and [to] disrupt what is collected when that choice is ignored."

When is a horse a potato?

The font is based around ligatures , a long-standing feature of many fonts that is usually used to replace certain letter pairs with a more readable version when they're smushed up next to each other. With ShieldFont, though, those ligatures are instead used to replace entire words with others in an attempt to destroy the text's value to scrapers. This substitution only happens when the font engine draws the page on screen, meaning scrapers that simply download plaintext source code get an altered version that end users never see.

Read full article

Comments

Terabytes worth of credentials, many belonging to the world’s biggest and most sensitive organizations, have been exposed in a supply-chain attack on LiteLLM, an open source tool that streamlines AI-driven software development. Microsoft, Amazon, Cisco, Samsung, and Salesforce are only a handful of the entities whose access secrets were exposed.

The revelation was posted on Tuesday and Wednesday by security firms CloudSEK and Hudson Rock. CloudSEK said it found cloud keys, repository tokens, SSH keys, Kubernetes secrets, package publishing credentials, environment variables, and AI provider keys that could allow attackers to gain access to more than 2,500 organizations.

40 minutes is all it takes

The credentials were extracted during a 40-minute window in March while the victims used compromised versions of LiteLLM downloaded from the package’s official location in the Python Package Index repository. Hudson Rock said it made the discovery after analyzing a 195TB file that it obtained. Neither firm identified the source of the information.

Read full article

Comments