I never tried the recipe. Instead, I found a human-created one that the LLM might have scraped. That’s the way these models work, of course. They repackage collective knowledge into something that feels tailored to you. This may be OK for dairy alternatives (unless you’re a vegan blogger). But on the order of the world, and truth—the focus of my role as a fact-checker at WIRED—the stakes are exponentially higher.
Over the past year or so, more and more people have looked at me with great pity. Surely a fact-checker at a magazine isn’t long for this AI-upgraded world. Call me foolish, but I’m not that worried. Very little of humanity’s collective knowledge, I’ve concluded, lives on the internet. And according to my research, AI is even more wrong than people might think.
WIRED’s fact-checking department is old-school: meticulous line-by-line annotations, primary sources whenever possible, and a broader-scale ethical and legal review. We question basic assumptions, look for new or conflicting information, call and talk to people—make sure. It’s a quick-hit peer review, functioning as best it can at the same pace as the news itself.
As far as I can tell, AI hasn’t come for this process quite yet. What it has come for is “post hoc” fact-checking, the Snopes-style analysis of something’s factuality after the fact. In the UK, an initiative called Full Fact has built out its own AI tools to help thwart the spread of misinformation. These tools, used in more than 40 countries, process huge volumes of data, from social media posts to podcast transcripts, then pinpoint specific claims that humans can investigate further. “You definitely need a human being,” says Mark Frankel, Full Fact’s head of public affairs.
The reason for that is simple: AI still gets things wrong. As a fact-checker, I’d love to be able to tell you exactly how often. But it’s not so easy. Since 2018, nearly 17,000 papers have been posted to arXiv on LLMs, many focused specifically on the question of their reliability. Still, it’s worth trying to pin down a working figure.
In any article that comes across WIRED’s fact-checking desk, there’s usually a decent amount of “b-matter”: statistics, news events, quotes, anything that helps contextualize the topic. Fact-checkers tend to Google this basic information, and that process, in the form of the search engine’s dreaded AI Overviews, constitutes my main interaction with AI. In my professional opinion, it’s unusable—wrong—about a third of the time.
This might be a generous assessment, though. A March 2025 study from the Tow Center for Digital Journalism found that more than 60 percent of responses from AI-powered search engines were inaccurate. A BBC study puts the wrongness of chatbots closer to 45 percent, the number I see cited more often. Because percentages are distancing, let me put this more plainly: AI could be wrong about half the time.
Does it matter which model? Elon Musk has said Grok is the smartest, but I haven’t seen much research that agrees. Claude led the pack in RealFactBench, a fact-checking-focused benchmark test developed by computer scientists in China and the UK last year. It scored 73 percent accuracy across all metrics. (To be fair, Grok was not assessed.) Another benchmark, SimpleQA, developed by OpenAI in October 2024, posed more than 4,000 single-answer questions to models from OpenAI and Anthropic. None of the models exceeded 50 percent accuracy. Google updated the benchmark earlier this year, winnowing the question set to 1,000. Gemini 2.5 Pro came out on top, with 55.6 percent accuracy.
Then there’s the models’ own assessments. When I asked ChatGPT how accurate the major LLMs are, it told me that most models had 90 to 96 percent accuracy on some professional-style tests. It then offered a link, confusingly, to a paper on a sleep medicine certification exam. On “general real-world questions,” it simply offered me the rate at which models like it have been shown to hallucinate: 1 to 2 percent, apparently, though when I tried to click through to that referenced source, it didn’t exist.
Some say the models are getting smarter, but this doesn’t necessarily mean fewer hallucinations. In fact, it could mean more, a kind of overcompensation rooted ineradicably in their programmed need to please users. In a 2025 report on the future of AI by the Association for the Advancement of Artificial Intelligence, 60 percent of surveyed researchers doubted that the “factuality” problem would be solved anytime soon.
Recently, I dug out that old test and gave it to (the free versions of) ChatGPT, Claude, Gemini, and Grok.
Grok came out of the ether like I was interrupting its supper: “Yes, I know exactly what fact checking is.” OK. It talked a lot about bias and put “credible” and “truth” in very loud quotation marks. It was also obsessed with data, along with gathering and analyzing more data than would ever be practicable or possible for a working fact-checker. It did, somewhat to my surprise, point out that fact-checking was historically women’s work.
Claude and Gemini did pretty well. They understood the task, laid out a reasonable approach, even flagged potential legal issues. Gemini did give me this very cringe phrase: I would look for “Paper Trails” to back up the “People Trails.”
ChatGPT seemed overeager and insecure. It spoke in buzzwords and generalizations. The approach it laid out seemed very time-consuming (including building a fact-checking grid where each sentence was broken apart and diagrammed). It offered to show me how it would “mark it up,” exactly “like a professional fact checker.” It then generated a paragraph that didn’t exist in the story. We tried that for a while, and then it offered to check a real paragraph for me. I gave it a fairly googleable selection, but it didn’t actually check any facts. None of the models did. They all gave me a plan of attack, told me exactly what they would do, and then stopped short of actually doing it.
“I don’t think it’s an option to sit AI out as some kind of fad or something that won’t dramatically impact how people find information,” says Angie Holan, head of the International Fact-Checking Network, a Poynter initiative that connects more than 170 fact-checking organizations across the world. Holan says she finds herself more comfortable with AI than some of her colleagues are. If a model leads you to authoritative sources that you are able to verify yourself, there you go, she says. Fact-checkers, journalists, librarians, archivists—all should be engaging with these models, learning how they’re put together: “That way you can understand the strengths and weaknesses of these tools,” she says.
I don’t disagree. In fact, the more time I spend with AI, the more capable I feel as a human fact-checker.
Most physical media in the world remains offline. In Lost in Time: Our Forgotten and Vanishing Knowledge, Jack Bialik points out that the technologies and knowledge bases we assumed were recent are actually in many cases millennia old (assembly lines, cataract surgery, even batteries). “Perhaps even more sobering is the realization that our storage technologies are far more likely to succumb to deterioration and useful obsolescence than hieroglyphics or ancient Sanskrit carved in a pyramid or on a temple wall,” he writes.
Years ago, during a fact-checking assignment, I talked to the sci-fi writer and history professor Ada Palmer, who told me what she often tells her students: We know less than 1 percent of what happened 500 years ago, and two-thirds of what we know is wrong. Knowledge exists on a timeline too, and the work of generations is carrying on that knowledge without little bits slipping through and getting lost. Are we really OK entrusting our legacy to a bunch of distributed servers, operated by microchips with lifespans of 5 to 10 years?
One final thing that I’ve been ignoring, which is so very human of me, is that humans make mistakes too. As Holan reminded me, abstaining from chatbots isn’t some foolproof saving grace. At least, I’m 33 to 90 percent sure that’s what she said. At the end of our interview, when I looked down at my recorder, I found I’d forgotten to turn it on.

What Say You?
Let us know what you think about this article in the comments below. Alternatively, you can submit a letter to the editor at mail@wired.com.

Can AI do fact-checking? A WIRED fact-checker fact-checks.

I’m a Professional Fact-Checker. AI Is Wrong More Often Than

large_AiFactChecking-Final.jpg

medium_AiFactChecking-Final.jpg

small_AiFactChecking-Final.jpg

thumbnail_AiFactChecking-Final.jpg

AiFactChecking-Final.jpg

I’m a Professional Fact-Checker. AI Is Wrong More Often Than You Think

## Nobody Has AI Security Fully Figured Out — Not Even Google

We are all operating in uncharted territory when it comes to artificial intelligence security. That includes the biggest technology companies on the planet.

During a recent backstage conversation at a Los Angeles event, Francis de Souza, Chief Operating Officer of Google Cloud, offered a grounded perspective on the AI security challenges facing organizations today. Speaking with the calm, deliberate clarity of an academic, de Souza acknowledged that businesses are currently in the middle of a difficult transition — but expressed confidence that things will improve. "There'll be a transition period, and then I think we get to this better place," he said.

While his remarks weren't specifically directed at Google's own situation, recent events make it clear that even Google itself is still working through significant security challenges.

## Security Must Be Built In From the Start

The central point de Souza pressed was one that cybersecurity professionals have long been urging business leaders to take seriously — and that AI has now made impossible to ignore: security cannot be treated as an optional add-on.

"As companies embark on this AI journey, they need to take a platform approach," he said. "Security is not something you can bolt on later, and it's not something you can leave up to employees to do on their own."

One specific concern he raised was the rise of so-called **shadow AI** — employees independently turning to consumer-grade AI tools without any oversight or approval from their organizations. De Souza argued that businesses must demand security, governance, and accountability from their AI platforms right from day one. "There's no such thing as an AI strategy without a data strategy and a security strategy. They need to go hand in hand."

## The Threat Landscape Has Fundamentally Changed

Beyond internal governance, de Souza painted a stark picture of just how dramatically the external threat environment has evolved. The average time between an initial security breach and the escalation to the next phase of an attack has collapsed — dropping from eight hours down to just 22 seconds. Meanwhile, the attack surface companies must defend has expanded well beyond traditional network boundaries.

"In addition to your usual estate, you have models now. You have data pipelines used to train the models. You have agents, you have prompts. All of this needs to be protected," he noted.

### The Hidden Danger of AI Agents Inside Your Organization

One underappreciated risk de Souza highlighted involves AI agents operating within a company's internal systems. These agents can inadvertently expose long-forgotten data repositories that haven't received a security review in years.

"A lot of organizations have old SharePoint servers and access controls they haven't really updated, but it didn't matter because nobody really knew where they were," he explained. "But agents roaming your enterprise will find those data assets and will expose the data on them."

This is a particularly subtle threat — not a sophisticated external hack, but the unintended consequence of deploying powerful tools inside environments that were never fully secured to begin with.

## Fighting Machine-Speed Attacks With Machine-Speed Defenses

De Souza's prescription for dealing with these accelerated threats is to match them with equally fast, AI-powered defenses. He described the emergence of fully agentic security systems — where AI agents drive the defensive response autonomously, with human professionals overseeing rather than directly executing every action.

"Instead of having a human-led defense or even a human in the loop, you can now have humans overseeing a fully agentic defense," he said.

He also made clear that this is no longer purely a technical matter. "This is a board-level issue and an executive team issue. It's not just a security team's issue."

## The Talent Gap Is Real — And Growing

Even as AI takes on a greater share of defensive responsibilities, the human expertise needed to manage and oversee these systems remains scarce. At the same time, AI is introducing new vulnerabilities faster than security teams can realistically address them.

LinkedIn's Chief Information Security Officer, Lea Kissner, captured the scale of the challenge bluntly in comments to The New York Times, warning that the industry should brace for a "bug-pocalypse" and suggesting that a sustainable, long-term understanding of AI security is still several years away.

## Google's Own Security Gaps Come Into Focus

The gap between the advice being offered by platform providers and their own current practices is worth examining closely — and Google provides an instructive case study.

Investigative reporting by *The Register* revealed a troubling pattern affecting Google Cloud developers over recent weeks. A number of developers found themselves hit with unexpected bills in the tens of thousands of dollars following unauthorized API calls to Google's Gemini AI models — services many had never intentionally activated.

The root cause traced back to API keys that developers had originally set up for Google Maps — deployed publicly, as Google's own documentation instructs. Without clear disclosure, Google had quietly expanded the scope of those keys to include access to Gemini services.

### Real Developers, Real Financial Damage

The human cost was significant. Rod Danan, CEO of the interview preparation platform Prentus, saw his account generate a bill of over $10,000 in approximately 30 minutes after attackers exploited a compromised API key. Isuru Fonseka, a developer based in Sydney, woke to charges of roughly AUD $17,000 — despite believing a $250 spending cap was protecting him.

What neither developer had been clearly informed of was that Google's automated systems had upgraded their billing tiers based on account history, quietly raising their effective spending ceilings to as high as $100,000 without obtaining their explicit consent.

Google refunded both developers after *The Register* published its initial coverage. However, the company stated it has no intention of changing its automatic tier-upgrade policy, indicating that preventing service interruptions takes priority over enforcing users' stated budget limits.

### Deleting a Compromised Key May Not Be Enough

A separate investigation added another layer of concern. Research conducted by security firm Aikido found that even developers who act quickly to delete a compromised API key may still be at risk. According to Aikido's findings, attackers can continue using a revoked key for up to 23 minutes, because Google's revocation process propagates gradually across its infrastructure rather than taking effect immediately.

Aikido researcher Joseph Leon reported that during this window, attack success rates are highly variable — in some instances, over 90% of requests using the supposedly revoked key still authenticated successfully. During that time, bad actors can exfiltrate files and cached conversation data from Gemini.

Leon also pointed out that this delay does not appear to be an unavoidable engineering limitation. Google's newer service account API credentials revoke in approximately five seconds, while the company's newer AQ-prefixed key format takes about one minute. "Both run at Google scale," Leon wrote. "Both suggest this is technically solvable for Google API keys, too." In other words, the 23-minute revocation window reflects a matter of organizational priorities — not technical impossibility.

## The Bottom Line: Sound Advice, But Mind the Gap

De Souza's guidance on AI security strategy is both thoughtful and timely. The principles he outlined — treating security as foundational, adopting a platform-wide approach, matching AI-powered threats with AI-powered defenses, and elevating security to the boardroom — represent exactly the kind of strategic thinking organizations need right now.

But the situation at Google itself serves as a useful reminder that even the most sophisticated players in the industry are still catching up. There is currently a meaningful gap between what platform providers are recommending and how quickly they are adapting their own systems and policies. For any organization navigating the AI security landscape today, that context is just as important to keep in mind as the advice itself.

From shadow AI risks to Google's own API billing failures, discover why AI security remains a work in progress for everyone — including Big Tech.

AI Security in 2025: Even Google Is Still Figuring It Out

large_GettyImages-2266466589.jpg

medium_GettyImages-2266466589.jpg

small_GettyImages-2266466589.jpg

thumbnail_GettyImages-2266466589.jpg

GettyImages-2266466589.jpg

AI Security Is a Work in Progress — Even for Tech Giants Like Google

## The Autonomous Vehicle Revolution: Closer Than You Think, Further Than You'd Hope

Robotaxis are simultaneously here and not quite here — a paradox that perfectly defines the current state of autonomous vehicle technology. Step onto the streets of San Francisco and you might reasonably conclude that the self-driving revolution has already arrived. But presence alone doesn't equal permanence, and that distinction is keeping every player in the autonomous vehicle space on edge.

Waymo, widely regarded as the frontrunner in commercial robotaxi deployment both in terms of riders served and fleet size, is experiencing this tension firsthand. The company recently suspended freeway operations across San Francisco, Los Angeles, Phoenix, and Miami while it addresses performance challenges in active construction zones. It's a pointed reminder that going live commercially is just the beginning of a much longer journey — not the finish line.

Every new city Waymo enters and every new capability it rolls out seems to expose a fresh set of edge cases that need solving. The commercial launch is conditional, and while that condition won't last forever, it underscores how much work remains.

## SpaceX's IPO Filing Pulls Back the Curtain on Musk's Business Empire

Shifting gears from wheels to rockets — the recently filed SpaceX IPO documentation has offered an unusually detailed look at the financial entanglements running through Elon Musk's constellation of companies.

The connections between Tesla and SpaceX have never been a secret. Tesla, as a publicly traded company, has always been required to disclose financial dealings with Musk-affiliated entities. But the IPO filing goes deeper, and with the recent merger of xAI into SpaceX, all of these transactions now fall under a single corporate umbrella.

### What the Numbers Reveal

The financial flows between Musk's various ventures are substantial. In 2025 alone, SpaceX purchased $506 million worth of Tesla's Megapack commercial energy storage systems — nearly triple the figure from the prior year. The company also acquired $131 million in Cybertrucks. SpaceX paid The Boring Company, Musk's infrastructure firm, $1 million to build tunnels at its Bastrop, Texas facility. Separately, X — the social media platform acquired by xAI and now folded into SpaceX — spent another $1 million leasing space from The Boring Company.

Tesla's earlier investment in xAI has since been converted into an equity stake in SpaceX following the acquisition.

Looking ahead, two major joint projects between SpaceX and Tesla are expected to dwarf these figures: Terafab, a chip manufacturing facility currently in development, and Macrohard, an AI platform designed to deploy autonomous agents that work alongside human employees.

All of this raises a compelling question: could a full SpaceX-Tesla merger be on the horizon?

## Funding Rounds and Industry Moves Worth Watching

### Startups Securing Capital

Several transportation and mobility startups have recently closed meaningful funding rounds. **Aboard**, a Southern California company building extended-range electric travel trailers, secured $13 million in a pre-Series A round led by Ondine Capital and Llama Ventures. The startup has brought on noted automotive designer Richard Kim — known for the BMW i3 and i8 — as a consultant.

**Quartermaster**, an Arlington, Virginia-based firm developing distributed ship-sensing networks, raised $43 million in a Series A round co-led by First Round Capital and Quiet Capital.

**Scapia**, an Indian travel booking platform, closed a $63 million funding round spearheaded by General Catalyst, with Peak XV Partners and Z47 also contributing.

### Strategic Partnerships Shaping Tomorrow's Roads

**May Mobility** struck a significant strategic agreement with Geely-backed automotive tech firm **Ecarx**, under which Ecarx will supply May Mobility with thousands of purpose-built robotaxi vehicles. Initial deployments through a third-party partner are slated for next year, with full commercial scaling targeted by 2028. The total estimated value of the partnership sits around $750 million over its lifespan.

**Stellantis**, the group behind Jeep and Ram, has partnered with autonomous driving startup **Wayve** to bring hands-free driving capability to its vehicles by 2028. The automaker also unveiled a $70 billion turnaround strategy featuring 11 new models for the North American market, including a revived Chrysler lineup.

**Uber** has grown its stake in German food delivery giant Delivery Hero to 19.5%, according to Bloomberg.

## Tesla FSD Expands in Europe as Lyft Clarifies Its AV Strategy

Tesla's Full Self-Driving (Supervised) driver-assistance software has received approval in Lithuania, making it the second European country to greenlight the technology. European expansion is central to both Tesla's product roadmap and CEO Elon Musk's financial interests — his $1 trillion compensation package is tied to several milestones, including reaching 10 million active FSD subscriptions by 2035.

Meanwhile, **Lyft** has published its official stance on autonomous vehicles, aligning closely with rival Uber's position: a functional ride-hailing ecosystem requires both human and robotic drivers working in tandem. The statement is as much a political calculation as a practical one, designed to reassure the company's gig worker base while acknowledging that robotaxis remain far from mainstream in most American cities.

## A Fresh Look at the 2026 Nissan Leaf Platinum+

The Nissan Leaf has come a long way. The 2026 Leaf Platinum+, priced at $42,635 including destination charges and select upgrades, represents a meaningful generational leap over its predecessors.

The third-generation Leaf now offers an EPA-estimated range of 259 miles, with certain configurations exceeding 300 miles. But the range improvement isn't what stands out most. The interior has been substantially reimagined — lighter, more modern, and noticeably more premium in feel, complete with sophisticated ambient lighting.

The top-trim version comes equipped with wireless phone charging, a panoramic dimming roof, a heads-up display, and a sweeping 14.3-inch curved central screen. Previous criticisms around the lack of standard tech have been addressed head-on: the 2026 model now includes a 360-degree camera, wireless Apple CarPlay and Android Auto compatibility, and adaptive cruise control as standard features. It's a compelling package that makes a strong case for the Leaf's continued relevance in an increasingly competitive EV segment.

## The Bigger Picture

Across robotaxis, IPO filings, startup funding, and EV hardware, the transportation sector is in a state of energized flux. Progress is real, but so are the constraints. The companies navigating this space most effectively are those treating their commercial launches not as endpoints, but as starting lines.

From Waymo's freeway pause to SpaceX's IPO revelations and the 2026 Nissan Leaf review — here's the full state of transportation tech in 2025.

Robotaxis, SpaceX IPO & EV Tech: 2025 Mobility Update

large_waymo-back-getty.jpg

medium_waymo-back-getty.jpg

small_waymo-back-getty.jpg

thumbnail_waymo-back-getty.jpg

waymo-back-getty.jpg

Robotaxis, SpaceX IPO, and the Real State of Autonomous Vehicles in 2025

## Robots Step Into the Kitchen at a San Francisco Nonprofit

In the heart of San Francisco's Tenderloin district, a four-story building hums with quiet purpose. Inside, staff members, volunteers, and — increasingly — robots work side by side to prepare and package thousands of medically customized meals for people who depend on them. The organization making it all happen is Project Open Hand, a nonprofit with nearly four decades of history and a mission that has never been more relevant.

### A Legacy Built on Compassion

Project Open Hand was established in 1985 by Ruth Brinker, a local grandmother and HIV-awareness advocate who recognized a growing crisis and decided to act. What began as a direct response to the AIDS epidemic has since grown into a comprehensive meal program serving individuals managing serious medical conditions including heart disease, diabetes, and chronic kidney disease.

The meals the organization produces are far from generic. Every box is tailored to meet specific nutritional requirements, accounting for individual allergies and health needs. It is a complex, detail-intensive process — one that demands both precision and a steady workforce.

## The Volunteer Problem

For years, Project Open Hand leaned heavily on corporate volunteer groups, teams of employees sent by their companies to spend a few hours assembling meal kits as part of organized charity initiatives. That model worked well — until it didn't.

The COVID-19 pandemic effectively wiped out that pipeline. As workers left the city and corporate volunteer programs were suspended, Open Hand found itself short-handed at exactly the wrong moment. San Francisco has since experienced a degree of recovery, energized by an influx of artificial intelligence companies and biotech startups. But the renewed economic activity has not translated into the kind of community engagement the nonprofit once relied on.

"We used to have so many corporate groups come in here," says Paul Hepfer, CEO of Project Open Hand. "There are so many new businesses — AI businesses, biopharma businesses — that aren't engaged the way they were pre-pandemic, which is really unfortunate."

With volunteers harder to recruit, the organization began exploring unconventional solutions. The answer arrived, fittingly enough, during a casual conversation on the Bay Area Rapid Transit.

## Enter Chef Robotics

Chef Robotics is a San Francisco-based company that describes its work as building physical AI for the food industry. Its robotic systems are designed specifically for the plating stage of food preparation — not cooking, not chopping, but the mechanical act of portioning and placing food into trays at scale. The company counts established food brands like Amy's Kitchen and Factor among its clients and is currently training its robots to eventually tackle more complex assembly tasks.

When a chance encounter between employees from both organizations led to a conversation about a potential partnership, Hepfer did not hesitate. Despite the subscription cost involved in renting the robots, he saw the investment as worthwhile.

"Nonprofits often operate under a scarcity mindset, and I think that's a disservice to the people we serve, because then you're not looking for innovations or quality improvements," Hepfer says.

### How the Robots Actually Work

If you visited Open Hand's facility today, you might not immediately notice the robots at all. There are just two of them, and they operate only a couple of hours each day as part of a conveyor-belt assembly line staffed by a small team of human volunteers.

The robotic arms function somewhat like oversized claw machines, reaching down into trays of portionable foods and depositing measured scoops — potato salad, corn, and other items — into the correct compartment of each meal tray. The arms can be fitted with different attachments to handle roughly 70 separate ingredients.

The system is not flawless. The robots occasionally miss their mark, leaving stray bits of food on the tray edges. A human volunteer is stationed nearby specifically to wipe down trays before they are sealed. Scattered frozen corn on the floor is a regular sight, swept up and discarded at the end of each session.

"Food is weird," acknowledges Rajat Bhageria, CEO of Chef Robotics. "It's sticky, it's malleable, it's wet. Even the best simulation doesn't completely get it."

But the imperfection is manageable — and notably, not much worse than what human hands sometimes produce.

## Real Impact on the Assembly Line

Prior to the robots' arrival, Open Hand's volunteer teams were capable of completing around 500 meal kits per hour. The robotic systems add approximately 200 more on top of that during smooth operations. Perhaps more meaningfully, the robots free up human volunteers to focus on higher-skill kitchen tasks like vegetable prep and cooking batches of plant-based protein — work that is harder to automate and arguably more engaging.

"It's not even that they're faster," says Alma Caceres, a sous chef involved in the meal preparation process. "It's that we don't have the volunteers."

Bhageria frames the technological shift in practical terms. "Having an arm and a scooping motion turns a physics problem — like how cooked is your onion — into a software problem — like do you have the right motion path. So it's a lot more scalable."

### A Community That Keeps Giving Back

Not everyone at Open Hand came through the front door as a volunteer. Joseph Sobiesiak first walked into the organization in the early 1990s as someone in need of its services. Decades later, he now helps manage the meal assembly line.

"I didn't die," he says simply. "And so now I'm here as a way to give back."

Asked about the robots, Sobiesiak admits he was initially skeptical. Over time, however, his position has softened.

"I'm old-school," he says. "It's working better than it did at first. Things are definitely much faster than before."

## Looking Ahead

Hepfer is candid about the limits of the technology. The robots do not replace the need for volunteers — they supplement it. But he is hopeful that embracing innovation so visibly will send a signal to San Francisco's growing tech community that Open Hand is an organization worth engaging with.

"I'm hoping that maybe the gravy on top of all this — the low-salt gravy on top — might be that people from the tech world might see that we are open to innovating and using technology and AI to improve the product we're providing for people's health," he says.

In a neighborhood that has long been overlooked by the city's wealthier interests, a nonprofit is proving that cutting-edge technology and human compassion are not mutually exclusive — and that feeding people well is a problem worth solving with every tool available.

Project Open Hand is using Chef Robotics' robotic arms to fill its volunteer gap and deliver medically tailored meals to San Francisco's most vulnerable residen

Robots Help San Francisco Nonprofit Feed the Vulnerable

large_Robot-Meals-Gear-Chef+-robot-(1).jpg

medium_Robot-Meals-Gear-Chef+-robot-(1).jpg

small_Robot-Meals-Gear-Chef+-robot-(1).jpg

thumbnail_Robot-Meals-Gear-Chef+-robot-(1).jpg

Robot-Meals-Gear-Chef+-robot-(1).jpg

How Robots Are Helping a San Francisco Nonprofit Feed People Who Need It Most

## Garlic's Hidden Power Over Insect Reproduction

Researchers at Yale University have uncovered a surprising new use for one of the world's most common kitchen ingredients. In a detailed phytochemical study examining 43 different fruits and vegetables, scientists set out to identify natural compounds capable of disrupting the reproductive behavior of flying insects. Their unexpected conclusion? Garlic stands in a class of its own.

The research team selected fruit flies as their primary test subject, largely because this species is known to mate on food surfaces — making them an ideal model for studying how edible substances might influence insect behavior. The working hypothesis was straightforward: if insects regularly interact with food during mating, certain plant compounds might interfere with that process.

## What Happened When Insects Encountered Garlic

After exposing fly specimens to mashed versions of all 43 foods tested, the scientists found that the vast majority had no meaningful effect on reproductive activity. None of the other foods produced a notable aphrodisiac or inhibitory response. Garlic, however, was a dramatic exception — it completely shut down both mating behavior and egg-laying.

This striking result prompted a deeper investigation. The team wanted to understand whether the effect was driven by smell, taste, or both. They designed two separate experiments: one in which the flies could only detect the odor of garlic puree, and another in which they were allowed direct contact with it. The data clearly pointed to taste as the critical factor. Simply smelling garlic was not enough — the insects had to experience it on their sensory organs for reproduction to be suppressed.

## The Chemical Compound Behind the Effect

A subsequent chemical analysis of garlic identified the specific compound responsible for this biological interference: diallyl disulfide. This sulfur-containing molecule targets a sensory receptor found in the fly's taste organs called TrpA1 — a receptor that functions as a biological alarm system, triggering immediate avoidance responses when potentially harmful or aversive substances are detected.

When diallyl disulfide activates TrpA1, it stimulates a specific group of bitter-taste neurons, prompting not only a physical rejection response but also measurable changes at the genetic level. Published findings in the journal *Cell* reveal that garlic exposure alters the expression of multiple genes in these insects, including one strongly associated with the sensation of fullness or satiety.

### How Satiety Suppresses Reproduction

This connection to satiety is particularly significant. The researchers believe that when garlic compounds trigger feelings of fullness in female insects, it directly interferes with the biological signals that normally drive appetite, feeding, and ultimately reproduction. In other words, garlic appears to short-circuit the chain of events that leads to mating by making the insects feel as though they have already fed — even when they haven't.

## Effective Across Multiple Insect Species

To confirm that these findings were not limited to fruit flies, the research team repeated their experiments on several other flying insect species. These included two mosquito species responsible for transmitting serious illnesses such as dengue fever, yellow fever, and the Zika virus, as well as tsetse flies, which are vectors for sleeping sickness. In every case, garlic demonstrated a consistent ability to suppress reproductive behavior.

This broad effectiveness suggests that the mechanism at work is not species-specific but may be a common vulnerability across a range of flying insects with similar sensory biology.

## A Scalable, Natural Solution for Pest Control

The implications of this research extend well beyond the laboratory. The findings suggest that *Allium sativum* — common garlic — could be developed into a cost-effective, environmentally friendly tool for managing insect populations that threaten both public health and global food supply chains.

John Carlson, a Yale professor and co-author of the study, highlighted garlic's practical appeal: "It's inexpensive and grown all over the world." He also noted a curious historical footnote to the discovery. "The idea of using it to ward off hematophagous creatures was proposed in 1897 by Bram Stoker in his novel *Dracula*, and perhaps he was right."

While Stoker's intentions were purely fictional, modern science may have validated his instinct in ways he never anticipated — with garlic proving to be a genuine deterrent not just for mythical vampires, but for some of the world's most dangerous real-world pests.

Yale scientists found that garlic's key compound, diallyl disulfide, blocks mating and egg-laying in mosquitoes and other insects. Here's how it works.

Why Garlic Repels Mosquitoes and Blocks Reproduction

large_GettyImages-117581895.jpg

medium_GettyImages-117581895.jpg

small_GettyImages-117581895.jpg

thumbnail_GettyImages-117581895.jpg

GettyImages-117581895.jpg

Science Confirms Garlic Can Stop Mosquitoes From Mating and Reproducing

## Breaking Free From the Bedside Phone Trap

Sleeping without a smartphone within arm's reach sounds like a modest achievement. Yet for millions of people, it might as well be climbing Everest barefoot. That was precisely my situation — until the Dreamie alarm clock entered my life and quietly dismantled a decade-long dependency I didn't fully realize I had.

I'll be the first to admit that needing a $250 gadget to stop scrolling in bed is not exactly a heroic story. But if you've ever found yourself opening Instagram at 2 a.m. when you only meant to turn on a podcast, you already know exactly what I'm talking about.

## Why Phones and Bedrooms Are a Terrible Combination

The science is clear: phone use before and during sleep disrupts rest quality, and poor sleep cascades into problems with mental health, focus, physical wellbeing, and mood. I've known this for years. I've read the articles, nodded at the advice, and then placed my phone on my nightstand anyway — every single night for more than a decade.

Over time, I did manage to build one genuinely healthy bedtime habit: reading before sleep. It helps me decompress and signals to my brain that the day is winding down. But even with a book in hand, my phone was never far away. On the nights when my mind simply refused to quiet down, the only reliable solution was closing my eyes and listening to a podcast or audiobook until consciousness slipped away.

The problem? Accessing those podcasts meant picking up my phone. And picking up my phone meant notifications. And notifications meant being wide awake for another two hours.

A survey of 2,000 American adults found that 87% keep their phones in their bedrooms at night. I was firmly embedded in that majority — and I needed something to change.

## What Makes the Dreamie Different

The Dreamie alarm clock doesn't reinvent the wheel with elaborate biometric sensors or AI-powered sleep tracking. Its standout feature is almost disarmingly simple: **it can play podcasts directly from the internet, no phone required.**

Because podcasts are distributed through open RSS feeds, any developer can build a custom RSS-based application — and that's exactly what Dreamie does. The device connects to Wi-Fi, pulls your chosen podcast from the web, and plays it on demand. It's a quiet nod to one of the internet's most enduring open standards, one that streaming giants like Spotify have been steadily working to replace with closed, proprietary systems.

## How the Dreamie Sleep System Works

### Ambience Mode

In its most basic state, Dreamie functions as a standard bedside clock with a gentle ambient display. Nothing complicated, nothing overwhelming.

### Wind Down Mode

This is where your nightly routine begins. You can program Dreamie to ease you toward sleep with a combination of soothing sounds and warm lighting. My personal setup features the soft crackling of a fireplace paired with a dim, flickering orange glow — running for roughly 25 minutes while I read. It's a remarkably effective way to signal to your body that sleep is approaching.

### Noise Mask Mode

Once the wind-down period ends, Dreamie transitions to a continuous background soundscape. I use a thunderstorm setting, which plays through the night until the morning routine begins. If sleep arrives early, you can manually trigger this mode whenever you're ready.

### Sunrise Alarm

Rather than jolting you awake with a harsh buzzer, Dreamie gradually brightens a simulated sunrise light, easing you out of sleep naturally before the alarm tone sounds.

### Back to Sleep Mode — The Star of the Show

This is the feature that genuinely changed my nights. If you wake up at 3 a.m. with a restless mind, a simple downward swipe activates "back to sleep" mode. You can choose from a built-in breathing exercise, an ambient soundscape, or — crucially — any podcast episode you've selected in advance.

The key detail here is the pre-selection. You choose your podcast before you go to bed, so there's no bleary-eyed scrolling through menus in the dark. You swipe, the audio begins, and your brain has something to follow other than its own anxious chatter. For those sharing a bed with a partner, Dreamie is also Bluetooth-compatible with wireless headphones, keeping your midnight listening private.

## The Real-World Impact on Sleep

Before Dreamie, my middle-of-the-night routine looked like this: wake up, reach for phone, intend to open a podcast app, notice a notification, open that notification, and then spend the next two hours falling deeper into a rabbit hole I never meant to enter. My own impulse control was the problem — but Dreamie elegantly removed the temptation from the equation entirely.

The mornings improved too. Without my phone on the nightstand, I wasn't spending 30 minutes scrolling before getting out of bed. I was simply getting up, starting my day, and functioning like someone who had actually slept.

## Where Dreamie Falls Short

No device is without its limitations, and honesty demands acknowledging them.

Audiobooks remain a gap in Dreamie's capabilities. Apps like Libby or Libro.fm aren't supported, and there's currently no way to upload downloaded audio files directly to the device. For those who rely on audiobooks as their primary sleep aid, this is a meaningful limitation. Podcast support is excellent; audiobook support is essentially nonexistent.

During testing, there were nights when I bypassed Dreamie entirely and reached for my phone to listen to something specific that wasn't available as a podcast. Once the phone was in my hand, old habits crept back in. Dreamie is a powerful tool, but it works best when you commit to its ecosystem.

## How Does It Compare to the Brick?

Toward the end of the review period, I also tested a device called the Brick — a $59 gadget designed to block all apps on your smartphone except the ones you explicitly allow, such as podcast or audiobook players. At less than a quarter of Dreamie's price, it's a compelling alternative for anyone who wants to curb phone use without removing it from the bedroom entirely.

But there's a philosophical difference worth considering. With the Brick, your phone is still your phone — it's still the last thing you see before closing your eyes. With Dreamie, your phone can live in another room entirely. For some people, that physical separation is exactly what makes the difference.

## Is Dreamie Worth $250?

The price is steep for an alarm clock, no question about it. However, there are no subscription fees and no companion app required — the cost is a one-time purchase. The interface is clean and intuitive, borrowing familiar design cues from the standard smartphone clock app, so the learning curve is minimal even for less tech-savvy users.

If you are someone who genuinely struggles to keep your phone out of bed — and the data suggests most of us are — the Dreamie offers a thoughtfully designed solution that addresses the root of the problem rather than just treating the symptoms. It won't work miracles on its own, but paired with a willingness to change, it's a remarkably effective nudge in the right direction.

For the podcast-dependent, poor-sleeping, notification-addicted among us, Dreamie might be the most practical $250 you spend on your health this year.

The Dreamie alarm clock plays podcasts so you can leave your phone out of the bedroom. Here's how it works and whether it's worth the $250 price tag.

Dreamie Alarm Clock Review: Sleep Better Without Your Phone

large_Dreamie-playing-a-podcast-2.png

medium_Dreamie-playing-a-podcast-2.png

small_Dreamie-playing-a-podcast-2.png

thumbnail_Dreamie-playing-a-podcast-2.png

Dreamie-playing-a-podcast-2.png

How the Dreamie Alarm Clock Finally Broke My Bedtime Phone Addiction

## Elon Musk's Clean Energy Vision Appears to Be Fading

Elon Musk built his reputation on the promise of a cleaner, electrified world. For years, Tesla's various Master Plans served as a roadmap toward what Musk himself described as moving away from a "mine-and-burn hydrocarbon economy" and into a "solar-electric economy." Yet a closer look at his companies' recent decisions tells a very different story — one that raises serious questions about whether that vision still holds any weight.

## xAI Goes All-In on Natural Gas

Musk's artificial intelligence venture, xAI, has made headlines not for its cutting-edge technology alone, but for how it powers that technology. The company has deployed dozens of unregulated natural gas turbines to keep its data centers running, with plans already in place to invest an additional $2.8 billion in fossil fuel-powered infrastructure. For a billionaire whose entire brand was built around clean energy disruption, this is a striking contradiction.

What makes the situation even more puzzling is Musk's well-documented habit of directing his companies to do business with one another. SpaceX has purchased over 1,200 Cybertrucks worth $131 million, and xAI has spent roughly $697 million on Tesla Megapacks — the large-scale battery storage systems used to manage peak energy loads. Yet xAI has not made any meaningful purchases of Tesla's solar panels. The clean energy piece of the puzzle appears to have been quietly set aside.

## SpaceX's IPO Filing Points to Space as the Energy Solution

The recently released SpaceX IPO filing sheds light on where Musk's energy ambitions have shifted. Rather than doubling down on terrestrial solar, SpaceX is placing its bets on space-based solar power. The company argues that solar arrays deployed in orbit can generate more than five times the energy of ground-based systems, thanks to continuous, uninterrupted sunlight around the clock.

This idea has gained traction among several Silicon Valley executives who are increasingly frustrated with the limitations and opposition that Earth-bound data centers face. The concept is straightforward: put massive server infrastructure into orbit, power it with abundant space solar energy, and sidestep the regulatory headaches and community pushback that come with building on the ground.

## The Economics of Orbital Data Centers Are Far From Settled

### Cost and Complexity Remain Major Barriers

While the vision sounds compelling, the financial and technical realities are far more sobering. The cost of delivering power to satellites in orbit is currently several times higher than what a conventional terrestrial data center would pay. Protecting sensitive computing hardware from radiation and the harsh conditions of space adds another layer of expense and complexity.

There is also a fundamental unresolved question: whether AI training workloads can actually be distributed across multiple satellites in orbit. A significant portion of AI computing may simply need to remain on the ground regardless of how advanced space infrastructure becomes. SpaceX is not solving one problem here — it is attempting to solve many simultaneously.

### Musk's First-Principles Gamble

Musk appears to view xAI's current natural gas-powered facilities as temporary measures — a stopgap until SpaceX can launch gigawatts' worth of computing power into orbit. In his view, that milestone may only be a few years away. Once achieved, the ground-based infrastructure, fossil fuel turbines included, would presumably be retired.

His reasoning stems from what he calls "first principles" thinking. The SpaceX filing references "terawatt-scale annual AI compute growth," suggesting Musk believes demand for computing power will far exceed what the planet's existing energy infrastructure can support. For context, all of the world's data centers currently consume roughly 40 gigawatts combined. The scale Musk envisions is extraordinary by any measure.

## A Missed Opportunity Closer to Home

Despite the boldness of the orbital energy concept, there is a compelling counterargument that deserves attention. Transporting solar panels on a truck to a ground-based installation requires a fraction of the energy needed to launch those same panels into space. Earth-based solar technology is already proven, increasingly affordable, and nowhere near its full potential.

Pursuing perfection in orbit should not come at the expense of progress on the ground. There is still enormous room to expand clean energy infrastructure here on Earth, and xAI's data centers represent a concrete and immediate opportunity to do exactly that.

## What Happened to the Plan to Eliminate Fossil Fuels?

Just three years ago, Tesla published its Master Plan Part 3, a detailed document outlining a credible path to eliminating fossil fuels from the global economy. The irony is difficult to ignore: one of Musk's own companies is now actively expanding its dependence on those very fuels.

Musk is undeniably skilled at identifying emerging trends and betting big before the rest of the world catches on. But being visionary does not guarantee being right — and the gap between his stated mission and current corporate behavior is growing harder to explain away.

Elon Musk once promised a solar-electric future, but xAI's fossil fuel use and SpaceX's orbital energy plans suggest a dramatic change in direction.

Is Elon Musk Giving Up on Solar Power for Space Energy?

large_GettyImages-2275592313.jpg

medium_GettyImages-2275592313.jpg

small_GettyImages-2275592313.jpg

thumbnail_GettyImages-2275592313.jpg

GettyImages-2275592313.jpg

Is Elon Musk Abandoning Solar Power in Favor of Space-Based Energy?

## IBM and Ferrari Join Forces to Revolutionize Formula One Fan Engagement

Formula One has evolved far beyond a motorsport — it has become a global cultural phenomenon. And now, two powerhouse brands, IBM and Scuderia Ferrari HP, are teaming up to ensure fans get more out of the experience than ever before.

### Why IBM Set Its Sights on Formula One

Approximately two years ago, IBM identified a notable gap in its sports partnership portfolio: Formula One. The sport's explosive growth, particularly in North America — fueled in large part by Netflix's hit docuseries *Drive to Survive* — had transformed F1 drivers into household names and attracted a wave of new, passionate fans.

F1's inherently technology-driven nature has made it a magnet for major tech firms. Companies like AWS, Oracle, and Anthropic have all staked their claim in the paddock, providing sponsorship exposure alongside advanced data analytics and artificial intelligence tools that give teams a measurable competitive advantage.

For IBM, the choice of partner was straightforward. Scuderia Ferrari — the most decorated team in Formula One history — was the obvious destination.

"They're the winningest team in history," said Kameryn Stanhouse, IBM's Vice President of Sports and Entertainment Partnerships.

## Turning Race Data Into Fan Stories

At the core of this collaboration is something that increasingly drives big-brand sports partnerships: sophisticated AI solutions. Stanhouse highlighted that sports offer an ideal environment for demonstrating AI's real-world value because of the sheer volume of data generated and the relatable context in which that data is applied.

"They actually see how it serves them," she explained, referring to how sports storytelling makes AI tangible and meaningful for everyday users.

During a single race, F1 teams process millions of data points every second, capturing granular details about every movement of the car and driver. The challenge — and the opportunity — lies in converting that raw information into compelling, accessible content that resonates with fans.

To tackle this head-on, Ferrari created a brand-new role within the organization, appointing Stefano Pallard as Head of Fan Development. His mission is clear: make every fan feel personally recognized.

"That starts with taking the data we get from the track and turning it into content that is easy to follow and engaging," Pallard told TechCrunch.

## A Complete Overhaul of the Ferrari Fan App

The IBM-Ferrari partnership has centered heavily on rebuilding the Ferrari fan app from the ground up. Previously, the app functioned as little more than a race schedule reference — fans would visit briefly for event details, then move on.

The redesigned app is a far more dynamic platform, featuring:

- **AI-generated race summaries** that break down events in an engaging, digestible format
- **Interactive prediction games** where fans can compete with one another
- **An AI-powered companion** that answers fan questions in real time
- **Exclusive behind-the-scenes content** spotlighting drivers and team operations
- **Multilingual support**, including Italian — a surprisingly overdue addition for a team deeply rooted in Italian culture

Stanhouse used a compelling example to illustrate the storytelling power embedded in the app: "There are two drivers, but did you know it takes 24 people working simultaneously in two seconds to change a tire?" Facts like these, she noted, help fans feel a deeper, more personal connection to the team.

### Personalization Powered by AI

Beyond content delivery, the team is leveraging AI to analyze how fans interact with the app — tracking which articles generate the most engagement, and interpreting the sentiment behind messages fans send through the platform.

"That helps us understand what resonates most with the Tifosi and it directly informs how we shape our storytelling and how we deliver content," Pallard explained, using the beloved nickname for Ferrari's global fanbase.

The results speak for themselves. Since IBM's involvement began, engagement metrics have trended consistently upward, with a reported **62% increase in engagement over race weekends**.

## Building Loyalty Across a More Diverse Fanbase

Another critical factor shaping the app's development is the rapidly evolving demographic makeup of Ferrari's audience. Formula One is no longer a niche interest dominated by a narrow segment. According to data released by F1, **75% of new fans are women**, with a significant proportion belonging to Gen Z.

One key driver of this shift is the **F1 Academy**, an all-female racing series designed to nurture the next generation of women in motorsport. As the fanbase diversifies, so too do its expectations.

"They are asking for more data, more insight, more features, and we have to be able to deliver that," said Pallard.

Unlike sports apps IBM has developed for tournament-based events — where user engagement tends to spike for a few weeks and then fade — the Ferrari app is engineered for year-round engagement. The goal is to keep fans connected to the team whether it's race season or the off-season.

## The Road Ahead: A Five-Year Vision

Ferrari and IBM are not resting on early wins. The partnership has an ambitious long-term roadmap centered on deeper personalization and increasingly immersive fan experiences.

"With IBM, the vision for the next five years is to make every fan feel like the experience was built for them, whether they have been with us for 30 years or 30 days," Pallard said. "That is how you build loyalty that lasts."

As artificial intelligence continues to redefine how brands engage with their audiences, the IBM-Ferrari collaboration stands as a compelling model for the future of sports fandom — one where technology doesn't replace the human connection, but makes it deeper, smarter, and far more personal.

IBM and Scuderia Ferrari HP are leveraging artificial intelligence to revolutionize fan engagement through a reimagined app and personalized storytelling.

IBM and Ferrari Use AI to Transform F1 Fan Experience

large_01-bxd-ferrari-xg-photography-v2.jpg

medium_01-bxd-ferrari-xg-photography-v2.jpg

small_01-bxd-ferrari-xg-photography-v2.jpg

thumbnail_01-bxd-ferrari-xg-photography-v2.jpg

01-bxd-ferrari-xg-photography-v2.jpg

How IBM and Ferrari Are Using AI to Build the Ultimate F1 Fan Experience

## The One Bag That Does Everything — and Then Some

Not every bag earns a permanent place in your daily rotation. Most carryalls are either too precious to abuse or too flimsy to trust with anything that matters. The **L.L.Bean Zip Hunter's Tote Bag** is neither. It sits in a rare category of gear that genuinely improves your life without demanding you treat it gently — and that's exactly why it's worth talking about.

## Born for the Market, Built for the Wild

Rainbow Grocery in San Francisco is the kind of place that reminds you food comes from the earth. Operating as an old-school co-op since 1975, the store stocks entirely organic produce — unwrapped, unbagged, and gloriously messy. Chicory arrives dewy and fresh, chanterelles come dusted with forest debris, and beets leave behind a trail of rich, dark soil. It's honest food, and it demands an honest bag.

That's where the Zip Hunter's Tote earns its keep. Every haul from Rainbow goes straight into it — parsnips, turnips, daikon radishes, and whatever else happens to be caked in garden grit. When the shopping is done, a quick rinse and a hang to dry is all it takes to reset the bag completely.

## Durability That Refuses to Quit

The bag's exterior is constructed from **1,200-denier polyester** — a heavy-duty, water-resistant fabric that borders on indestructible. After being checked as airline luggage on more than a dozen flights, it shows zero signs of wear: no scuffs, no tears, no fraying at the seams. The top zipper isn't sealed, but it's built heavier than necessary and holds up perfectly in wet conditions.

What genuinely separates this tote from the competition, however, is its interior. The entire inside surface is coated with a layer of **thermoplastic lining** that creates a moisture barrier in both directions. It keeps dry things dry when the outside world is wet — and it keeps wet, muddy, or messy items contained when you need to transport them without making a disaster of everything else.

## A Camping Companion That Saves Rental Cars

The thermoplastic lining has proven its worth on more than one camping trip. On one occasion, the final day of a trip brought heavy rain. A soaked rain jacket, a dripping tent, a ground cloth peppered with wet pine needles, and boots caked in thick Northern California mud all needed to make the journey home. Everything went into the Hunter's Tote, the zipper got pulled shut, and the bag landed in the back seat of a rental car — with the interior upholstery emerging completely unscathed.

Back home, the process was simple: empty the bag, flip it inside out, and blast it clean with a garden hose.

## Smart Details Worth Noting

### Handles and Hardware
The bag's **reinforced handles** are thick and sturdy enough to lift comfortably even when the tote is packed past its intended capacity — ideal for loading and unloading from a cargo bike. Two heavy-duty plastic tabs sit at each end of the zipper, designed as clip points for the included shoulder strap. While the strap itself tends to get in the way during daily use, those tabs serve double duty as perfect hanging points when the bag needs to air dry.

### Size Matters — Go XL
The Zip Hunter's Tote is available in three sizes. The **XL version holds 53 liters**, making it one of the most spacious tote bags available anywhere. Set it upright on a flat surface and the reinforced bottom gives you a generous interior footprint of **10 by 19 inches** — enough room to organize a serious haul. For maximum versatility, the XL is the clear choice.

## What You Won't Find Inside

This bag makes no apologies for its simplicity. There are **no exterior pockets, no padded laptop sleeve, and no interior organizers**. What you get is a single, cavernous, easy-to-clean compartment ready for whatever you throw at it — literally. From cases of sparkling water to soaked camping equipment to a full grocery run, it handles the load without complaint.

## The Bottom Line

The L.L.Bean Zip Hunter's Tote Bag isn't glamorous, and it doesn't try to be. It's a workhorse carryall built with serious materials, thoughtful construction, and a lining that makes cleanup effortless. Whether you're navigating a rainy bike commute, packing up a campsite in a downpour, or hauling muddy produce from your favorite co-op, this tote is the only bag you need in your corner.

The L.L.Bean Zip Hunter's Tote is a durable, water-resistant carryall built for camping, commuting, and grocery runs. Here's why it's the only bag you need.

L.L.Bean Zip Hunter's Tote Bag Review: Worth It?

large_1%20-%20Courtesy-of-L.L.-Bean.jpg

medium_1%20-%20Courtesy-of-L.L.-Bean.jpg

small_1%20-%20Courtesy-of-L.L.-Bean.jpg

thumbnail_1%20-%20Courtesy-of-L.L.-Bean.jpg

1%20-%20Courtesy-of-L.L.-Bean.jpg

Why the L.L.Bean Zip Hunter's Tote Bag Is the Last Carryall You'll Ever Buy

## Nuclear Startup Deep Fission Is Going Public — Again

If the name Deep Fission sounds familiar in the context of stock market listings, that's because this isn't the first time the underground nuclear reactor startup has attempted to enter the public markets. The company recently announced plans for a Nasdaq IPO targeting $157 million in fresh capital — but a closer look at its financials and technical progress suggests investors should tread carefully.

## A Previous 'Public' Listing That Never Really Was

Back in September of last year, Deep Fission made headlines when it completed a reverse merger with Surfside Acquisition, a Delaware-based shell company. The deal — a common shortcut that allows private companies to achieve a public listing by absorbing an already-listed entity — raised $30 million through a private placement priced at $3 per share.

However, that listing turned out to be public in name only. While the transaction technically made Deep Fission a reporting company with SEC obligations, its shares never actually traded on any exchange. The company had previously indicated it would list on the OTCQB, a marketplace designed for early-stage companies that don't qualify for major exchanges like the NYSE or Nasdaq. Yet no trace of the company appeared on that platform, and Deep Fission itself confirmed in its S-1 filing that its stock had never been publicly traded.

The company declined to address questions on the matter, citing the customary quiet period ahead of its IPO.

## Bigger Ambitions, Shakier Foundations

This time around, Deep Fission is pursuing a full Nasdaq IPO at a price range of $24 to $26 per share — a valuation that could reach up to $1.66 billion. That's a striking figure for a startup that, just one year ago, was reportedly struggling to close a $15 million funding round.

What makes the ambition even more eyebrow-raising is the picture painted in the S-1 filed on May 20, which is arguably more concerning than the December filing that preceded it. Key milestones have slipped. The company previously aimed to achieve nuclear criticality — the threshold at which a chain reaction becomes self-sustaining — by July 2026. That target has now been quietly dropped, with no replacement date offered.

### Going Concern Warning Remains

Perhaps most telling is the persistence of a "going concern" warning in the new filing — the same cautionary language that appeared in December. In plain terms, if the IPO doesn't succeed, Deep Fission could exhaust its funds within the next 12 months.

The company's financial deterioration since December adds further weight to that concern. Its cumulative deficit climbed from $56.2 million to $88.1 million as of March. In just the past six weeks, cash and cash equivalents fell by roughly $6.4 million — approximately 7% of its reserves — indicating a significant and accelerating cash burn.

## Underground Drilling: Harder Than It Sounds

On the technical side, Deep Fission has shifted its focus toward drilling operations, which some observers interpret as an acknowledgment that burrowing into the earth is considerably more complex than originally anticipated.

The company began drilling the first of three planned test wells in March, with the well designed to collect geological data down to a depth of 6,000 feet. At just eight inches in diameter, however, this test well is far smaller than what commercial operations would require.

### Scaling Up Presents Enormous Challenges

For commercial deployment, Deep Fission says it will need boreholes measuring between 30 and 50 inches in diameter and extending a full mile underground — dimensions that already exceed what's standard in the oil and gas industry. The company has yet to settle on a specific diameter, which means its reactor design also remains unfinalized. Until it can determine what size hole it can reliably drill, the entire engineering roadmap stays in flux.

## What's Changed — And What Hasn't

One notable development since December is an $80 million equity investment, which includes $20 million from Blue Owl, a data center development firm that also signed a non-binding memorandum of understanding for potential future power plants. Despite that injection of capital, the going concern warning remains in place, suggesting the funding hasn't meaningfully stabilized the company's outlook.

It's possible Deep Fission is sitting on encouraging internal data not yet disclosed in its public filings. But given the scale of what's riding on this IPO, that explanation seems optimistic at best.

## Riding the Nuclear Hype Wave?

A more straightforward explanation is that Deep Fission and its backers are looking to capitalize on a surge of investor enthusiasm surrounding fission-based energy. Just last month, fellow nuclear startup X-energy went public in an upsized IPO, signaling strong market appetite for the sector.

However, the comparison isn't entirely flattering to Deep Fission. X-energy is already generating revenue and has made considerably more progress through the Nuclear Regulatory Commission's complex licensing process. That contrast underscores an important reality in the nuclear sector: valuation and technical progress are not the same thing, and enthusiasm can outpace reality by a wide margin.

## The Bottom Line

Deep Fission is seeking a nine-figure valuation on the strength of a compelling vision — underground nuclear reactors powering AI data centers — but its own filings reveal a company that is burning through cash, missing milestones, and still figuring out the basics of how to drill a big enough hole. Whether investors will look past those red flags in a market hungry for nuclear plays remains to be seen.

Deep Fission is pursuing a $157M Nasdaq IPO at a $1.66B valuation, but mounting losses, missed milestones, and a going concern warning raise serious red flags.

Deep Fission's $157M IPO: Bold Valuation, Bleak Financials

large_image-2.jpg

medium_image-2.jpg

small_image-2.jpg

thumbnail_image-2.jpg

image-2.jpg

Deep Fission's Second Shot at Going Public Raises More Questions Than Answers

## Your Phone Is a Target — Here's How to Fight Back

Spyware attacks are no longer the stuff of spy thrillers. Today, they are a documented, recurring threat faced by journalists, human rights defenders, political activists, and dissidents around the world. In early 2025, WhatsApp alerted approximately 90 users — the majority of them journalists and civil society workers across Europe — that they had been targeted by Paragon Solutions, an Israeli spyware firm. Shortly after, Apple issued threat notifications to another group of iOS users. Forensic investigators confirmed that at least two of those individuals, both journalists, had been infected with Paragon's Graphite spyware through a zero-click exploit — meaning their devices were compromised without them ever tapping a single link.

This is not an anomaly. Security researchers have spent the past 15 years cataloguing case after case in which state-sponsored hackers successfully infiltrated the phones and computers of reporters, critics, and political opponents.

## What Spyware Actually Does

The tools used in these attacks are sophisticated, expensive, and engineered to be invisible. Once spyware is installed on a device — and smartphones are the primary target because they contain nearly every detail of a person's life — the attacker gains sweeping access. They can listen to phone calls, read private messages, browse through photos, and silently activate the device's camera and microphone to record conversations. Real-time location tracking is also standard.

Fortunately, the biggest names in tech have responded. Apple, Google, and Meta each offer opt-in security features specifically built to counter these kinds of targeted attacks.

## Why You Should Use These Features

These tools do involve some trade-offs. Certain standard device functions may be restricted or disabled when enhanced security modes are active. However, security researchers, tech companies, and cybersecurity journalists widely recommend enabling them — especially if your profession or public profile could make you a person of interest to government surveillance operations.

Even if you don't consider yourself a high-risk target, these features add a meaningful layer of protection that keeps your personal data safer from falling into the wrong hands.

"These features are free, easy to enable, and the best defense we have today against sophisticated spyware," said Runa Sandvik, a security researcher with over a decade of experience protecting journalists and vulnerable communities. "If the features get in the way of something you need to do, you can easily turn them off again — meaning it costs very little to turn them on and try them out."

No security solution is foolproof. Spyware developers constantly search for new vulnerabilities, and software makers continuously patch and adapt in response. But these features have a proven track record.

## Apple Lockdown Mode

### What It Does

Available across all Apple devices — including iPhones, iPads, and Macs — Lockdown Mode deliberately restricts certain standard functions in exchange for dramatically stronger defenses against advanced threats. Apple acknowledges upfront that the device will not operate as it normally does when this mode is active.

The effectiveness of Lockdown Mode is well-documented. Researchers at Citizen Lab found it successfully blocked a Pegasus spyware attack developed by NSO Group. As of early 2025, Apple has reported no confirmed successful attack on any device running Lockdown Mode.

### How to Enable It

Navigate to **Settings → Privacy & Security → Lockdown Mode**. Toggle it on and your device will restart to apply the changes. You can also whitelist specific websites or apps to operate without Lockdown Mode restrictions, without disabling the feature entirely.

## Google Advanced Protection Program

### What It Does

Launched in 2017, Google's Advanced Protection Program fortifies your Google account against sophisticated hacking attempts. It introduces stricter login verification, limits which third-party apps can access your account data, and provides enhanced protections against phishing.

### How to Enable It

Visit Google's official Advanced Protection page and click **Get Started**. You will be prompted to sign into your Google account and set up a physical security key or software passkey as a secondary authentication factor. You will also need to provide a recovery phone number, recovery email, or backup passkey to complete enrollment.

## Android Advanced Protection Mode

### What It Does

Building on the concept pioneered by Apple's Lockdown Mode, Android's Advanced Protection Mode delivers comparable security hardening for Google's mobile operating system. It tightens app permissions, restricts sideloading of unverified applications, and strengthens defenses against malicious exploits.

### How to Enable It

Go to **Settings → Security and Privacy → Other Settings → Advanced Protection → Device Protection** and activate the feature.

## WhatsApp Strict Account Settings

### What It Does

With more than 3 billion active users worldwide, WhatsApp is an attractive target for intelligence agencies and surveillance firms. In 2019, NSO Group's hacking tools were used against roughly 1,200 WhatsApp users. A similar operation in early 2024 compromised around 90 accounts across Europe.

In response, WhatsApp introduced **Strict Account Settings**, an opt-in feature that activates enhanced privacy and security controls based on your operating system. These include tighter message link previews, restricted access from unknown devices, and additional account verification steps.

### How to Enable It

Open WhatsApp on your primary device and go to **Settings → Privacy → Advanced**, then toggle on Strict Account Settings.

## The Bottom Line

No single security measure offers absolute protection, and the battle between spyware developers and software defenders is ongoing. But the features described above are free, accessible, and proven to be effective. If you work in journalism, activism, law, or any field that could draw government scrutiny, enabling these tools is one of the smartest steps you can take to protect yourself. And if you're simply someone who values privacy, they're worth switching on anyway.

Apple, Google, and Meta offer free opt-in security modes to defend against spyware. Learn what each feature does and how to enable them on your devices.

Best Built-In Security Features to Block Phone Spyware

large_dark-phone-person-spyware.jpg

medium_dark-phone-person-spyware.jpg

small_dark-phone-person-spyware.jpg

thumbnail_dark-phone-person-spyware.jpg

dark-phone-person-spyware.jpg

How to Shield Your Phone From Spyware Using Built-In Security Features

## Meta's New App 'Forum' Takes Aim at Reddit

Meta has stealthily introduced a brand-new standalone application called **Forum**, signaling yet another push into community-driven social networking. Positioned as a direct rival to Reddit, the app is described by Meta as a "dedicated space built for deeper discussions, real answers and communities you care about."

The app's existence was first uncovered by well-known social media consultant Matt Navarra, who spotted it before any official announcement from the company.

## How Forum Actually Works

Users can log into Forum using their existing Facebook credentials. Once signed in, the app automatically pulls in your Facebook Groups, profile information, and activity history. Like the standard Facebook experience, Forum allows users to post content under a nickname.

It's worth noting that your groups remain active on Facebook itself — any content you publish through Forum will simultaneously appear within your corresponding Facebook Groups, keeping both platforms in sync.

### A Feed Built Around Real Conversations

One of Forum's defining features is its conversation-first feed. Rather than surfacing trending or algorithmically boosted content, Meta says Forum prioritizes what real people within your groups are actually discussing. The goal is to help users seamlessly continue conversations from where they last left off.

### AI-Powered Features at the Core

Forum comes equipped with an intelligent **"Ask" tab**, which allows users to pose questions and receive compiled answers drawn from discussions happening across multiple groups. This AI-driven tool is designed to surface relevant community knowledge quickly and efficiently.

Administrators also benefit from a dedicated **AI assistant** built into the platform, helping them manage group activity and moderate content with greater ease.

## Meta Has Tried This Before

This isn't Meta's first attempt at a standalone Groups experience. Back in 2014, the company launched a dedicated Groups app aimed at simplifying content sharing across communities. However, that project was discontinued in 2017 after failing to gain significant traction.

Whether Forum will fare better remains to be seen.

## Part of a Broader App Expansion Strategy

Forum is just one of two new apps Meta has introduced in recent weeks. The company also recently launched **Instants**, an app that enables users to share disappearing photos with their Instagram connections — drawing clear comparisons to BeReal and Snapchat.

These launches are part of a sweeping new strategy at Meta. According to a report by *The Wall Street Journal*, CEO **Mark Zuckerberg** told employees that AI-driven efficiencies are enabling the company to develop more apps at a faster pace than ever before.

Zuckerberg was quoted as saying, "So Chris and I have been talking about 'all right, well can we build 50 new apps?' Like, yeah probably. But we probably should start by doing a few before we just, like, ramp up trying to do 50 all at once" — referring to Meta's chief product officer, Chris Cox.

## Is Meta Just Copying the Competition?

Despite the ambitious rollout, critics point out that Meta's newest apps heavily borrow from existing platforms. Instants mirrors concepts popularized by BeReal and Snapchat, while **Meta Edits** — launched last year — is widely regarded as a clone of ByteDance's CapCut video editing app.

While Meta may believe users are hungry for more applications, consumer appetite for redundant copycat apps is questionable at best. The company has not yet responded to media requests for comment regarding Forum.

Meta quietly launched Forum, a standalone Reddit-like app for Facebook Groups featuring AI tools, community feeds, and deeper discussion features.

Technology