Instead of wiring into a home’s breaker box, the Backup by BioLite home backup power solution relies on thin battery panels that can fit behind appliances to keep them running for days at a time. It’s built around two 1.5kWh lithium iron phosphate (LiFePO4, or LFP) batteries, the Backup Core and Backup Extend. Like any uninterruptible power supply, the Backup Core keeps itself perpetually charged from a wall outlet while power is available, then switches to keep whatever’s plugged into it running during a power outage.

If you need more backup power, up to five Backup Extend units can be connected to a Backup Core unit to expand the total capacity to 9kWh, and you don’t need a contractor or electrician to install any of it.

BioLite’s system was designed to be a cheaper and easier alternative to whole-home backup power solutions that rely on a central battery or gas-powered generator to keep an entire house running during a power outage. A natural gas generator alone can cost well over $5,000, while a Tesla Powerwall installation can set you back over $10,000.

The Backup Core battery panel will sell for $1,999, but BioLite will also offer a $2,999 Backup Complete solution, pairing a Core with a single Extend battery, that’s eligible for a 30 percent home energy tax credit. The company says the 3kWh Backup Complete has enough capacity to keep an 18 cubic foot fridge running for up to 60 hours or a larger 26 cubic foot fridge powered for up to 30 hours while also allowing for the occasional use of lights or other smaller appliances.

BioLite is bringing its backup power solution to consumers through a Kickstarter crowdfunding campaign that launches today, with discounts for early backers or those who opt to put down a deposit and pay in full through an installment plan later. Although the company has been around since 2006, making products like a power-generating camp stove, this will be its most expensive offering to date, and the usual caveats and risks with crowdfunded products apply here.

The Backup by BioLite batteries measure 29 inches tall and weigh between 35 and 40 lbs. They’re each just 2.8 inches thick, however, allowing them to be mounted out of sight behind appliances or furniture using hardware that takes about 30 minutes to install, claims the company. But a permanent installation isn’t necessary, as BioLite says the batteries will work just as well placed atop a fridge or slid under a bed.

AI startup Runway has made a name for itself building generative models seemingly trained on unlicensed content from around the internet. Now, the company has signed a deal with Lionsgate that will give it access to the studio’s massive portfolio of films and TV shows.

Today, Lionsgate — the studio behind films like the John Wick and Hunger Games franchises — announced that it is partnering with Runway to create a new customized video generation model intended to help “filmmakers, directors and other creative talent augment their work.”

In a statement about the deal, Lionsgate vice chair Michael Burns described it as a path toward creating “capital-efficient content creation opportunities” for the studio, which sees the technology as “a great tool for augmenting, enhancing and supplementing our current operations.” Burns also insisted that “several of our filmmakers are already excited about its potential applications to their pre-production and post-production process.”

Runway cofounder and CEO Cristóbal Valenzuela echoed Burns’ sentiment about the new model’s usefulness as an augmentation tool and said that the company’s goal is to give filmmakers “new ways of bringing their stories to life.”

Specific details about the deal — like whether creative teams will be compensated if / when their projects are used as training material for the model — are currently scant. But as The Hollywood Reporter notes, the prospect of being able to keep production costs down could have been one of the big selling points for Lionsgate, a studio known for sticking to smaller budgets compared to other entertainment outfits.

News of Lionsgate’s deal with Runway comes at a time when studios have increasingly begun implementing AI into their projects, despite many filmmakers’ concerns about how the technology’s unfettered use could threaten their jobs. Studios insistent on being able to create and use AI replicas of background performers was one of the major points of contention that ultimately led to the SAG-AFTRA strike last year.

Those concerns were part of what led to California Governor Gavin Newsom’s signing of two SAG-AFTRA-backed bills earlier this week that will grant performers and their estates more control over how and when their digitally created likenesses can be used by studios. And later this month, Newsom could very well end up signing into law SB 1047, another piece of hotly contested legislation that would make AI developers liable for the “critical harms” caused by their products.

(We reached out to SAG-AFTRA for comment about the partnership between Runway and Lionsgate but did not hear back in time for publishing.)

The word of the day in US v. Google was “parking.” As in: did Google buy some of its most ascendant and dangerous competitors in the online advertising business, all the while planning on parking them off in some far-flung corner of the company so that no one could possibly upset Google’s dominance? That is a central question of the government’s entire case against Google, and it came up over and over on Monday morning.

To kick off the second week of the landmark antitrust trial over Google’s control of online advertising, the Department of Justice called Neal Mohan, the CEO of YouTube and a longtime Google advertising executive. Mohan came to Google in 2008 through Google’s acquisition of DoubleClick, which formed the basis of Google’s now-unstoppable advertising engine. Mohan also helped advocate for the acquisition of Admeld, another company at the center of the suit. He argued throughout his testimony that Google was never attempting to buy up and neuter its competitors; it was simply trying to compete.

The Justice Department grilled Mohan on one of the core tenets of its case: that Google has built an impenetrable ad empire by owning all three major parts of the adtech stack, including the system publishers use to offer ad inventory on their pages, the system advertisers use to buy and place ads around the web, and the exchange in the middle where all the buying and selling actually takes place. This empire, lawyers allege, allows no real competition and ultimately makes things worse for all parties involved, Google excepted. And whenever a possible challenger did arise, Google simply bought and shelved — or, perhaps, parked — them.

The “parking” concept came up during Mohan’s two-plus hours of testimony, when Justice Department lawyer Aaron Teitelbaum showed him an email exchange about whether Google should buy Admeld. Admeld used a technology called yield management and was making inroads into the online ad market by letting publishers assess demand from multiple ad exchanges at once.

In those emails, another Google executive wrote that “one way to make sure we don’t get further behind in the market is picking up the [company] with the most traction and parking it somewhere.” Acquiring the company in that way “would let us solve the problems from a position of strength.” In the government’s view, this seemed to be clear evidence that Google was trying to take a threat off the market.

In court, Mohan argued that’s not what “parking” means at all. He acknowledged that Google was interested in Admeld because Admeld was further ahead in development but said Google had no intention of shelving or shuttering the product. “That’s absolutely not what was going on,” he said.

Parking, he explained, refers to Google’s acquiring a company and then letting it operate more or less as before while it also begins to rebuild and integrate into Google’s technology stack. This process takes time — often years — and Mohan said that leaving the products running actually indicates their importance to Google as products and not vanquished enemies.

Mohan argued over and over, occasionally seeming frustrated to have to repeat himself, that Google was simply doing what it had to do to keep up. He told Teitelbaum that the goal was always “to build the best advertising stack for publishers, as well as tools for advertisers.”

In Mohan’s telling, the advertising business has always been fiercely competitive, and companies like Facebook, Microsoft, and Yahoo even attempted to build similarly all-encompassing strategies. Controlling all three parts of the process, he said, is crucial to ensuring that only good ads are placed on only good websites, that everything happens quickly, and that no nefarious actors can cause trouble.

When Jeannie Rhee, one of the attorneys representing Google, began to cross-examine Mohan, she had him reiterate the parking point in several ways. She noted an annual update email Mohan had written to his team in 2008, after the DoubleClick acquisition, in which he compared the integration to “changing the engines on a plane while continuing to fly it.” Rhee had Mohan go through some of the DoubleClick team’s most impressive post-acquisition accomplishments, too, seemingly to show the product was still being actively developed.

Mohan’s testimony offered a fairly straightforward version of the arguments on both sides of this all-important trial. In the government’s eyes, Google has an insurmountable advantage in the ads business, built on the back of illegally tying various products to each other and by buying up any company that even looked like competition. According to Google, though, deep integration is the only way to build great ad products, and its acquisitions have only ever been in service of building better products in a competitive space.

The government has repeatedly presented evidence that it’s nearly impossible to leave Google’s platforms. Switching platforms for any reason is hard, and the prospect of leaving behind Google’s advertiser demand and access to platforms like Search and YouTube makes it untenable. Publishers have also argued that Google’s advertising products aren’t at all impressive. They say they feel stuck. And as the government sees it, Google is happy to spend hundreds of millions of dollars on startups to keep it that way.

In 2011, Google did acquire Admeld, for a reported price above $400 million. (A number, by the way, that the Justice Department argues is far above Google’s actual valuation of the company — allegedly a signal of Google’s willingness to overspend in the name of crushing threats.) The Justice Department briefly investigated the deal at the time but ultimately let it close. Now, the company’s technology is part of Google’s dominant ad exchange, known commonly as AdX. All that’s left of Admeld itself is a Google support page telling publishers why AdX is so great. 

Is that the good kind of parking or the bad and possibly illegal kind? That’s up to Judge Leonie Brinkema. She didn’t have much to say during Monday’s testimony, but everyone in the room acknowledged she’s the only one who matters.