Last week, the Environmental Protection Agency made lots of headlines by rejecting the document that establishes its ability to regulate the greenhouse gases that are warming our climate. While the legal assault on regulations grabbed most of the attention, it was paired with two other actions that targeted other aspects of climate change: the science underlying our current understanding of the dramatic warming the Earth is experiencing, and the renewable energy that represents our best chance of limiting this warming.

Collectively, these actions illuminate the administration's strategy for dealing with a problem that it would prefer to believe doesn't exist, despite our extensive documentation of its reality. They also show how the administration is tailoring its approach to different audiences, including the audience of one who is demanding inaction.

When in doubt, make something up

The simplest thing to understand is an action by the Department of the Interior, which handles permitting for energy projects on federal land—including wind and solar, both onshore and off. That has placed the Interior in an awkward position. Wind and solar are now generally the cheapest ways to generate electricity and are currently in the process of a spectacular boom, with solar now accounting for over 80 percent of the newly installed capacity in the US.

Read full article

Comments

Backlash over offensive Grok outputs continues, just a couple weeks after the social platform X scrambled to stop its AI tool from dubbing itself "MechaHitler" during an antisemitic meltdown.

Now, The Verge has found that the newest video feature of Elon Musk's AI model will generate nude images of Taylor Swift without being prompted.

Shortly after the "Grok Imagine" was released Tuesday, The Verge's Jess Weatherbed was shocked to discover the video generator spat out topless images of Swift "the very first time" she used it.

Read full article

Comments

At a time when subscription video-on-demand (SVOD) streaming services keep getting more expensive and their ads more extensive , Roku’s new streaming service, Howdy, is worth a closer look.

Howdy launched today in the US for $3 per month. It is currently only accessible via a web browser or the Howdy app on Roku OS. Roku said in its announcement today that Howdy will roll out to “mobile and additional platforms in the near future."

Howdy has, per Roku, “nearly 10,000 hours” of content from Warner Bros. Discovery, Lionsgate, and Filmrise. That’s a much smaller number of content distribution partners than what you’ll find on Roku’s flagship streaming service, The Roku Channel, which has shows and movies from companies like AMC Networks, MGM, NBCUniversal, and Paramount. (Roku also makes the Roku OS operating systems for smart TVs and streaming devices and Roku-brand TVs, and runs the Frndly TV streaming service.)

Read full article

Comments

Cisco said that one of its representatives fell victim to a voice phishing attack that allowed threat actors to download profile information belonging to users of a third-party customer relationship management system.

“Our investigation has determined that the exported data primarily consisted of basic account profile information of individuals who registered for a user account on Cisco.com,” the company disclosed . Information included names, organization names, addresses, Cisco assigned user IDs, email addresses, phone numbers, and account-related metadata such as creation date.

Et tu, Cisco?

Cisco said that the breach didn’t expose customers’ confidential or proprietary information, password data, or other sensitive information. The company went on to say that investigators found no evidence that other CRM instances were compromised or that any of its products or services were affected.

Read full article

Comments

A federal jury found on Friday that Meta violated the California Invasion of Privacy Act, the state's wiretap law, by collecting data from a period-tracker app without user consent.

Plaintiffs in a class-action case proved by a preponderance of evidence that Meta intentionally eavesdropped on and/or recorded conversations using an electronic device, said a verdict form released yesterday in US District Court for the Northern District of California. Plaintiffs also proved that they had a reasonable expectation of privacy and that Meta did not have consent from all parties to eavesdrop on and/or record the conversations, the jury found.

The lawsuit was filed in 2021 against Flo Health, maker of an app for tracking periods, ovulation, and pregnancy. Facebook owner Meta, Google, and app analytics company Flurry were added as defendants later. The plaintiffs settled with Flo Health , Google , and Flurry before the trial, leaving Meta as the only remaining defendant.

Read full article

Comments

OpenAI is preparing to raise what could be its final defense to stop The New York Times from digging through a spectacularly broad range of ChatGPT logs to hunt for any copyright-infringing outputs that could become the most damning evidence in the hotly watched case.

In a joint letter Thursday, both sides requested to hold a confidential settlement conference on August 7. Ars confirmed with the NYT's legal team that the conference is not about settling the case but instead was scheduled to settle one of the most disputed aspects of the case: news plaintiffs searching through millions of ChatGPT logs .

That means it's possible that this week, ChatGPT users will have a much clearer understanding of whether their private chats might be accessed in the lawsuit. In the meantime, OpenAI has broken down the "highly complex" process required to make deleted chats searchable in order to block the NYT's request for broader access.

Read full article

Comments

OpenAI is releasing new generative AI models today, and no, GPT-5 is not one of them. Depending on how you feel about generative AI, these new models may be even more interesting, though. The company is rolling out gpt-oss-120b and gpt-oss-20b, its first open-weight models since the release of GPT-2 in 2019. You can download and run these models on your own hardware, with support for simulated reasoning, tool use, and deep customization.

When you access the company's proprietary models in the cloud, they're running on powerful server infrastructure that cannot be replicated easily, even in enterprise. The new OpenAI models come in two variants (120b and 20b) to run on less powerful hardware configurations. Both are transformers with a configurable chain of thought (CoT), supporting low, medium, and high settings. The lower settings are faster and use fewer compute resources, but the outputs are better with the highest setting. You can set the CoT level with a single line in the system prompt.

The smaller gpt-oss-20b has a total of 21 billion parameters, utilizing mixture-of-experts (MoE) to reduce that to 3.6 billion parameters per token. As for gpt-oss-120b, its 117 billion parameters come down to 5.1 billion per token with MoE. The company says the smaller model can run on a consumer-level machine with 16GB or more of memory. To run gpt-oss-120b, you need 80GB of memory, which is more than you're likely to find in the average consumer machine. It should fit on a single AI accelerator GPU like the Nvidia H100, though. Both models have a context window of 128,000 tokens.

Read full article

Comments

Intel has a lot riding on "18A," its next-generation manufacturing process for silicon chips that the company claims will help it catch up to the lead that competitors like TSMC have built up over the last few years. With 18A, Intel would return to manufacturing its own processor designs in its own factories, including the upcoming Series 3 Core Ultra chips for laptops (codenamed Panther Lake), after manufacturing parts of all other Core Ultra chips with TSMC. Intel is also offering 18A manufacturing capacity to external chipmakers, a major milestone in former CEO Pat Gelsinger's plan to make Intel a competitive, cutting-edge (and primarily US-based) chip manufacturer for the rest of the industry.

But a Reuters report claims that Intel is struggling to make usable chips on 18A, according to "people who were briefed on the company's test data since late last year." As of this summer, these sources say that just 10 percent of the chips being manufactured on 18A are "up to [Intel's] specifications."

Intel disputed the numbers cited in the report. "Yields are better than that," Intel CFO David Zinsner told Reuters, though neither Zinsner nor Intel provided an alternate figure.

Read full article

Comments

Lift up either side of the new Morgan Plus Four’s engine decklid and no carburetors or velocity stacks materialize on the engine of this otherwise eminently classic-looking car. The latest generation of Plus Four returns to the American market this year for the first time since 2005. But to make it happen, Morgan spent five years homologating the little roadster with the National Highway Traffic Safety Administration, technically as a replica.

That's why, underneath the left-hand hood, a hilarious list of Federal Motor Vehicle Safety Standards exemptions lays out every way the Plus Four skips out on the usual fare, such as side impact protection, ejection mitigation, a hood latch system, and tire pressure monitoring. Apparently, that’s what it takes to produce a truly lightweight sports car in the modern era.

More importantly, this latest generation of the Plus Four, which debuted in 2020 in other regions, stays shockingly light by now using a bonded aluminum chassis to help with handling and rigidity while harnessing the power output of a modern BMW drivetrain. But Morgan still builds the body on an ash wood frame, proving that the old ways that allowed a tiny British coachbuilder to remain on the market as a low-volume manufacturer at a reasonable price may never truly die.

Read full article

Comments