NASA and IBM have teamed up to build an AI foundation model for weather and climate applications. They’re combining their respective knowledge and skills in the Earth science and AI fields, respectively, for the model, which they say should offer “significant advantages over existing technology.”

Current AI models such as GraphCast and Fourcastnet are already generating weather forecasts more quickly than traditional meteorological models. However, IBM notes those are AI emulators rather than foundation models. As the name suggests, foundation models are the base technologies that power generative AI applications. AI emulators can make weather predictions based on sets of training data, but they don’t have applications beyond that. Nor can they encode the physics at the core of weather forecasting, IBM says.

NASA and IBM have several goals for their foundational model. Compared with current models, they hope for it to have expanded accessibility, faster inference times and greater diversity of data. Another key aim is to improve forecasting accuracy for other climate applications. The expected capabilities of the model include predicting meteorological phenomena, inferring high-res information based on low-res data and "identifying conditions conducive to everything from airplane turbulence to wildfires."

This follows another foundational model that NASA and IBM deployed in May. It harnesses data from NASA satellites for geospatial intelligence, and it's the largest geospatial model on open-source AI platform Hugging Face, according to IBM. So far, this model has been used to track and visualize tree planting and growing activities in water tower areas (forest landscapes that retain water) in Kenya. The aim is to plant more trees and tackle water scarcity issues. The model is also being used to analyze urban heat islands in the United Arab Emirates.

Following Sam Altman's rollercoaster of a return as OpenAI's CEO, the company announced that it will now include Microsoft as a non-voting observer on its board. The question remains as to why the firm's largest investor wasn't on its board in the first place, but this seems to be somewhat addressed for now, at least. Altman is joined by co-founder Greg Brockman who resumes his role as President, whereas Mira Murati, who very briefly served as interim CEO throughout the drama, will return to her role as CTO.

The announcement also confirms a new board consisting of former Salesforce CEO Bret Taylor (chair), former Clinton Treasury Secretary Larry Summers, and original member Adam D'Angelo, who is also Quora's co-founder and CEO. It was earlier rumored that Altman's exit was partly influenced by D'Angelo's seeming conflict of interest, as OpenAI was developing a potential competitor to Quora's Poe service — the latter offers OpenAI's ChatGPT and GPT-4, along with several other text-generating AI models.

D'Angelo's presence on OpenAI's new board came as a surprise, and Altman took to X to address the elephant in the room. "Quora is a large customer of OpenAI and we found it helpful to have customer representation on our Board." The exec added that D'Angelo "has always been very clear... about the potential conflict and doing whatever he needed to do," including offering to leave the board, if necessary. As to why the original board wanted Altman out, he said "it is clear that there were real misunderstandings between me and members of the board."

OpenAI co-founder and chief scientist Ilya Sutskever was a former board member who allegedly led the ouster of Altman. The exec later openly admitted that he "deeply regret my participation in the board's actions" (with Elon Musk begging for attention in his thread), and he had since voiced his support for Altman's return as CEO. In his open message, Altman says "I harbor zero ill will towards him," and that his team is figuring out a way to let Sutskever continue his work at OpenAI.

In the same official announcement, OpenAI's new Chair, Taylor, assured that the company will "enhance the governance structure," and put together "an independent committee of the Board to oversee a review of the recent events," for the sake of the organization's stability.

Elon Musk, facing the fact that an already financially-precarious X could be poised to lose another $75 million in ad revenue following his boosting of an antisemitic conspiracy theory, has a new message for advertisers pulling back from the platform: “Go fuck yourself.”

Musk repeated the sentiment multiple times during an appearance at The New York Times’ DealBook event. “Don’t advertise,” Musk said. “If somebody is going to try and blackmail me with advertising, blackmail me with money? Go fuck yourself. Is that clear? I hope it is.”

“Hey Bob,” Musk added, in an apparent reference to Disney CEO Bob Iger, who appeared at the same event earlier in the day and spoke about the company’s decision to pull ads following Musk’s tweet earlier this month. Iger said that Disney’s association with X was “not necessarily a positive one for us,” according to Variety.

While Musk again denied being antisemitic, he did express some regret for engaging with the tweet that’s resulted in another exodus of advertisers from X. “I should have not replied to that particular person… I essentially handed a loaded gun to those who hate me,” Musk said about the post, per Variety.

X didn’t immediately respond to a request for comment. X CEO Linda Yaccarino had a front-row seat to the remarks, according to The Hollywood Reporter, which reports the former ad exec sat “stone-faced” during Musk’s tirade.

Elon Musk, facing the fact that an already financially-precarious X could be poised to lose another $75 million in ad revenue following his boosting of an antisemitic conspiracy theory, has a new message for advertisers pulling back from the platform: “Go fuck yourself.”

Musk repeated the sentiment multiple times during an appearance at The New York Times’ DealBook event. “Don’t advertise,” Musk said. “If somebody is going to try and blackmail me with advertising, blackmail me with money? Go fuck yourself. Is that clear? I hope it is.”

“Hey Bob,” Musk added, in an apparent reference to Disney CEO Bob Iger, who appeared at the same event earlier in the day and spoke about the company’s decision to pull ads following Musk’s tweet earlier this month. Iger said that Disney’s association with X was “not necessarily a positive one for us,” according to Variety.

While Musk again denied being antisemitic, he did express some regret for engaging with the tweet that’s resulted in another mass exodus of advertisers from X. “I should have not replied to that particular person… I essentially handed a loaded gun to those who hate me,” Musk said about the post, per Variety.

X didn’t immediately respond to a request for comment. X CEO Linda Yaccarino had a front-row seat to the remarks, according to The Hollywood Reporter, which reports the former ad exec sat “stone-faced” during Musk’s tirade.

Consumer Reports has published an extensive ranking of vehicle reliability, and the results pour cold water on the dependability of EVs and plug-in hybrids. The survey says electric vehicles suffer from 79 percent more maintenance issues than gas- or diesel-powered ones, while plug-in hybrids have 146 percent more problems. The troubles portray the industry’s growing pains with the relatively new technology as the planet sets record temperatures, and scientists warn of rapidly approaching deadlines to thwart global climate catastrophe.

The survey polled CR’s members about issues with their rides from the past year, gathering data on 330,000 vehicles. The publication’s data included models from 2000 to 2023, alongside a few (early launched) 2024 models. CR studied 20 “trouble areas,” including relatively minor issues like squeaky brakes or a broken interior trim and more problematic ones related to the transmission, engine or EV battery. The number of potential trouble areas varies by type: internal combustion engine (ICE) vehicles have 17, EVs have 12, traditional hybrids have 19 and plug-in hybrids have all 20.

The publication combined the data with its own track testing, owner satisfaction survey results and safety info. It then averaged it to assign each brand a numerical score (out of 100).

Lexus

Non-plugin hybrids scored well, with the survey indicating they suffer from 26 percent fewer issues than gas- and diesel-powered vehicles. CR highlighted the most reliable brands in that space, including the Lexus’ UX and NX Hybrid and Toyota’s Camry Hybrid, Highlander Hybrid and RAV4 Hybrid.

If only plug-in hybrids (PHEV) could enjoy those ratings. Instead, their longer list of trouble spots led to 146 percent more problems than traditional gas-powered vehicles. Lowlights include the Chrysler Pacifica, which scored an abysmal 14 out of 100, and Audi Q5. However, several PHEVs defied the category’s expectations, including “standouts” like the Toyota RAV4 Prime and Kia Sportage. Several others, including the BMW X5, Hyundai Tucson and Ford Escape, scored “average” in reliability.

Fully electric cars and SUVs, the vehicles many automakers aim to fill their dealership lots with by 2030, have mediocre average scores: 44 and 43, respectively. Electric pickups, the newest technology in the bunch, perhaps unsurprisingly scored worse with an average of 30.

Lexus came out on top among EV brands. All but one of its models scored above average or better in CR’s ratings. And the lone exception, the NX, still had an average score. Toyota also did well, including the 4Runner SUV, which CR describes as “among the most reliable models in the survey.” However, its electric Tundra pickup scored poorly. Other EVs with above-average scores include Acura’s RDX and TLX.

Photo by Roberto Baldwin / Engadget

Once practically synonymous with electric vehicles, Tesla had overall scores in the middle of the pack (alongside brands like Chevrolet, Buick, Ram, Cadillac and Dodge). CR says the Elon Musk-led company’s EV powertrains tend to fare better than those from traditional automakers. However, Ars Technica notes the company’s reliability scores struggled more with things like bodywork, paint / trim and climate systems.

Regionally speaking, Asian automakers enjoyed the highest average scores in the survey at a healthy 63. European companies were second with an average of 46, while US brands slumped with a somewhat disappointing score of 39.

Absurd Ventures, the new creative studio from Rockstar Games co-founder and ex-creative director Dan Houser, has announced its first projects. As it happens, neither of them are video games, at least not yet.

The first of these two new universes is called American Caper, which will debut as a graphic novel. It will focus on two normal but damaged families who are mired "in a world of corrupt business, inept politics and bungling crime." Comic book artist Simon Bisley (ABC Warriors, Lobo) is illustrating the graphic novel.

The other project is a 12-episode audio drama that's already in production. A Better Paradise is described as a near-future existential suspense thriller. Absurd Ventures is working with well-regarded audio company Q-CODE Media on the project.

Absurd Ventures plans to reveal more details about both projects in the coming months. However, referring to both American Caper and A Better Paradise as "universes" suggests the company has plans to expand them into transmedia enterprises, which may just include video games.

Meanwhile, we're just days away from getting our first proper glimpse at what has been keeping Houser's former colleagues at Rockstar busy for the last several years. The publisher will unveil the first trailer for the next Grand Theft Auto game in early December. Rumors suggest the reveal will take place at The Game Awards on December 7.

Amazon has hopped on the same bandwagon on which many major tech companies have hitched a ride this year by debuting its own image generator. AWS customers can now check out a preview of Titan Image Generator on the Bedrock console. They can either enter a text prompt to create an image from scratch or upload an image and edit it.

Amazon says the tool can produce large volumes of studio-quality, realistic images at low cost. It claims the AI can generate relevant images based on complex text prompts while ensuring object composition is accurate and that there are limited distortions. This, according to the company, helps with "reducing the generation of harmful content and mitigating the spread of misinformation."

Those looking to edit an image can isolate areas in which they want to add or remove details. They can, for instance, replace the background or swap an object in a subject's hand. The AI can also extend an image's borders by adding artificial details, much like the Generative Expand feature in Photoshop.

Amazon says Titan applies an invisible watermark to images that it generates. The company says this will "help reduce the spread of misinformation by providing a discreet mechanism to identify AI-generated images and to promote the safe, secure and transparent development of AI technology." It claims that the watermarks are resistant to modifications. According to a demo of the image generator, the AI can also generate a description of the image or relevant text to use in a social media post.

News of the image generator emerged at Amazon's AWS re:Invent conference, at which the company also showed off its latest AI chips and revealed a business-centric AI chatbot called Q. The company recently started offering advertisers a tool that lets them add AI-generated backgrounds to product images.

Content platform Substack just released a spate of new video capabilities, placing it in direct competition with YouTube and Patreon, among others. The video tools include a direct upload option, which is handy, and customizable paywalls for content creators. Before this, users were forced to upload videos to YouTube and embed a link. The upload tool is now readily accessible via the dashboard. It’ll even automatically split the audio and video for podcasters who want to court both audiences. 

As for the paywall options, you now get the same level of flexibility available to non-video users. Content creators can select a slice of the video to give away for free, locking the rest behind a paywall. The free preview segment should transition smoothly into a prompt to become a paying subscriber.

There’s also a new AI tool that generates transcripts from videos, for those who like to, gasp, read. The transcript is automatically created alongside the video upload and users can post it to the main feed. Additionally, viewers can click anywhere on the transcript to jump to that section of the video.

Video sharing has gotten a much-needed upgrade. Viewers can create their own custom clips sourced from any video. This creates a shareable link that includes branded visuals at the end featuring the creator’s logo and URL, so there will be no obvious thievery. Obviously, viewers can quickly share links to the entire clip if they want. Users can even directly download videos for publication on services like TikTok and Instagram. Again, that bumper will be there to give credit to the original creator.

The company wrote in a blog post that these new tools, taken together, make it so “the friction in starting a media business based on video has been reduced to almost zero.” To commemorate the launch, Substack is rolling out a number of exclusive video shows. There’s a food culture program with chef Nancy Silverton, a talk show starring actress Amber Tamblyn and a news program anchored by Chris Cuomo, among many others.

Substack has certainly been busy adding new features. The platform recently unveiled a Twitter-esque feature called Notes and last year launched a chat feature to make it more of a social space.

A team of researchers was able to make ChatGPT reveal some of the bits of data it has been trained on by using a simple prompt: asking the chatbot to repeat random words forever. In response, ChatGPT churned out people’s private information including email addresses and phone numbers, snippets from research papers and news articles, Wikipedia pages, and more.

The researchers, who work at Google DeepMind, the University of Washington, Cornell, Carnegie Mellon University, the University of California Berkeley, and ETH Zurich, urged AI companies to seek out internal and external testing before releasing large language models, the foundational tech that powers modern AI services like chatbots and image-generators. “It’s wild to us that our attack works and should’ve, would’ve, could’ve been found earlier,” they wrote, and published their findings in a paper on Tuesday that 404 Media first reported on.

Chatbots like ChatGPT and prompt-based image generators like DALL-E are powered by large language models, deep learning algorithms that are trained on enormous amounts of data that critics say is often scraped off the public internet without consent. But until now, it wasn’t clear what data OpenAI’s chatbot was trained on since the large language models that power it are closed-source.

When the researchers asked ChatGPT to “repeat the word ‘poem’ forever”, the chatbot initially compiled, but then revealed an email address and a cellphone number for a real founder and CEO”, the paper revealed. When asked to repeat the word “company”, the chatbot eventually spat out the email address and phone number of a random law firm in the US. “In total, 16.9 percent of the generations we tested contained memorized [personally identifiable information]” the researchers wrote.

Using similar prompts, the researchers were also able to make ChatGPT reveal chunks of poetry, Bitcoin addresses, fax numbers, names, birthdays, social media handles, explicit content from dating websites, snippets from copyrighted research papers and verbatim text from news websites like CNN. Overall, they spent $200 to generate 10,000 examples of personally identifiable information and other data cribbed straight from the web totalling “several megabytes”. But a more serious adversary, they noted, could potentially get a lot more by spending more money. “The actual attack”, they wrote, “is kind of silly.”

OpenAI patched the vulnerability on August 30, the researchers say. But in our own tests, Engadget was able to replicate some of the paper’s findings. When we asked ChatGPT to repeat the word “reply” forever, for instance, the chatbot did so, before eventually revealing someone’s name and Skype ID. OpenAI did not respond to Engadget’s request for comment.

Netflix continues to bet big on gaming. The streamer just announced that Rockstar’s Grand Theft Auto: The Trilogy – The Definitive Edition will hit its platform on December 14. The game will be free to play for Netflix subscribers, like the rest of its library. Each of the three titles included in the collection have been updated for mobile. These games include the iconic Grand Theft Auto III, Grand Theft Auto: Vice City, and Grand Theft Auto: San Andreas. They'll be playable via the Netflix app, but also available as standalone apps on Android and iOS.

The gameplay’s a bit dated by modern standards but Grand Theft Auto III basically invented the open-world genre, and the sequels further refined the model. Many players actually point to San Andreas as being a high watermark for the series, even outpacing more recent titles. In any event, this trilogy can hold you over until the eventual release of Grand Theft Auto VI. To that end, the forthcoming Rockstar sequel should be getting a trailer any day now.

Netflix has been snagging all kinds of games for its ever-growing library, from the massively popular roguelike Hades to the equally popular, uh, roguelike Dead Cells. The platform’s library also includes Braid, Death’s Door, Katana Zero and other hit indies.

The company’s also developing its own games, thanks to snatching up developer Night School, the team behind Oxenfree and Afterparty. This acquisition has already borne fruit, as Oxenfree II: Lost Signals released earlier this year to wide acclaim. Incidentally, the sequel isn’t stuck on Netflix, as it’s also available on Nintendo Switch, PS4, PS5 and PC via Steam. So the streamer might not be using games as just a way to lure in new subscribers. Netflix could be in it for the long haul.

Beyond Oxenfree II, other original games to hit the platform include Slayaway Camp 2: Netflix & Kill and a whole lot of titles based on pre-existing TV shows. There’s a game based on the recently-canceled fantasy epic Shadow and Bone and forthcoming titles inspired by Money Heist and The Dragon Prince, among others.