But perhaps one food item not on the list will be the one that the Artemis II crew will talk about forever: Uncrustables.

Commander Reid Wiseman, pilot Victor Glover, and mission specialists Christina Koch and Jeremy Hansen have all been promised a lifetime supply of the peanut-butter-and-jelly treat upon their splashdown, expected later today.

“Artemis, you rang? We’ve got the crew covered on the OG uncrusted sandwiches from here on out … no joke,” the brand wrote in a now viral Instagram post showing an Uncrustable hovering over the Earth, much like photos from the far side of the Moon taken aboard the Artemis II.

When the four astronauts climbed aboard NASA’s Orion spacecraft (which they dubbed Integrity), they embarked on a 10-day journey that will finally see their splashdown tonight, April 10, at around 8:07 in the evening. They knew the food that was aboard the ship—and they made one request for when they landed: the pocket-size PB&J treat. The Navy confirmed that the recovery vessel, when they make splashdown, will have an “abundant amount” of the sandwiches.

And Smucker’s replied with an even better PR stunt than maybe what Nutella had in the first place.

The Artemis II crew just traveled 252,756 miles from Earth, breaking the distance record set by Apollo 13 in 1970, and after all that record-breaking, what they really wanted was a peanut butter and grape jelly sandwich with the crusts cut off.

This story was originally featured on Fortune.com

“It’s 13 minutes of things that have to go right,” said NASA’s Artemis II flight director Jeff Radigan on Thursday at a news briefing.

Before the Orion spacecraft, named Integrity by the crew, ever left the Kennedy Space Center launchpad in Florida on April 1, NASA knew there was a problem. During the unmanned Artemis I mission in 2022, engineers discovered more than 100 locations on the Orion heat shield that had cracked and broken off during reentry.

Here’s the issue: it’s not supposed to do that. The shield was designed to melt away, not pop off in chunks. Instead, scientists discovered the culprit was a pressure problem buried within the shield itself. As the capsule dipped into the atmosphere, internal layers became scorching hot through a process called pyrolysis, trapping gas.

When the capsule briefly climbed back out of the atmosphere during its “skip” (meaning skip entry, which is when a spacecraft returning from high speed dips into the earth’s upper atmosphere. It’s the guided maneuver it uses to skip along the layer, closely mirroring a stone “skipping” across a pond, all before it reenters for a final landing. The outer layer hardened and became impermeable. This posed a problem because the gas had nowhere to go. On the second descent, the pressure burst through, taking chunks of the heat shield with it.

Now you’re wondering, that was Artemis I, surely they would never put four people—commander Reid Wiseman, pilot Victor Glover, and mission specialists Christina Koch and Jeremy Hansen—aboard a ship with such flaws. And you’d be partially right: the Artemis II has, remarkably, an even less permeable shield than the one on Artemis I, meaning the same failure mode was even more likely to occur.

It’s all about the right angle

Rather than delay the mission by more than a year to install a redesigned heat shield (as one engineer wanted), NASA flew Artemis II with the same flawed design and simply changed how the capsule returned. The solution was counterintuitive, with NASA instructing the crew to apply more heat more consistently. This shortened the skip phase and maintained higher temperatures throughout the descent, ensuring the outer char layer never cooled sufficiently to trap gas beneath it.

So these four astronauts, who broke a 56-year-old distance record and became the furthest humans to travel from earth when the mission brought them around the moon, not only had to overcome faulty Outlook problems and smelly toilet issues, but they had to enter the earth’s atmosphere at the right angle, at the right speed, and the right time—and they did it.

The four astronauts reached speeds of over 24,000 mph, equivalent to traveling across the continental U.S. in about six minutes. The 16.5-foot-wide heat shield reached approximately 5,000 degrees Fahrenheit, about half the temperature of the sun’s visible surface. The steeper, hotter trajectory also gave the capsule less range to maneuver away from bad weather near the Pacific splashdown zone.

It paid off

Not everyone was on board with the plan. Former NASA engineer Dr. Charles Camarda had publicly warned that NASA didn’t fully understand the root cause of the cracking and that the modified trajectory amounted to “playing Russian roulette.” But NASA stood by its data. Associate administrator Amit Kshatriya pointed to Artemis I flight data, ground testing, and engineering models as justification, and Glover acknowledged the risk head-on, noting the heat shield and parachutes are systems with zero fault tolerance built in.

The capsule splashed down safely in the Pacific, capping the first crewed lunar mission since Apollo 17 in 1972.

This story was originally featured on Fortune.com

The queries by Fed examiners are intended to assess the level of stress in the private credit industry and the potential for it to spill over to the wider financial system, said the people, requesting anonymity to discuss the work. 

Among the queries the Fed has been incorporating into its routine oversight process, the central bank has been seeking detail on the debt private credit funds have taken on from banks. In good times, that debt can juice returns and make private credit funds more enticing. In bad times, it risks exposing banks to losses.

The Treasury Department is also questioning the insurance industry about exposures to private credit, said people with knowledge of those separate discussions.

Representatives for the Fed and Treasury had no immediate comment.

The questions are one of the strongest signals yet that US regulators are working to get a handle on the scale of the strains in private credit, which has ballooned to an $1.8 trillion industry marketed first to institutional investors and increasingly now to individuals.

Private credit, which relies on investor money — rather than bank deposits — to make loans, had been on examiners’ radar for years. They stepped up focus when retail credit funds came under pressure in the recent months and investors rushed to pull cash. 

Regulatory Push

A growing chorus of international regulators have been warning about the risks of private credit. Financial Stability Board Chair Andrew Bailey said this week that private credit may be facing more stress after the shock to markets from the Iran war. The Financial Stability Oversight Council said at the end of March that it had discussed recent developments in the private credit sector.

The Fed’s questioning comes as President Donald Trump’s top financial watchdogs seek to loosen rules for Wall Street lending giants. Part of that deregulation effort is meant to both bolster banks’ ability to lend to private-credit outfits and to have traditional lenders better compete with nonbank firms in areas such as mortgage and small-business loans.

The move also shows that officials such as Fed Vice Chair for Supervision Michelle Bowman want to balance relaxed rules with more strategic queries from the industry about what they perceive as potential areas of risk, said some of the people.

Banks have sought to distance themselves from their less regulated nonbank rivals. JPMorgan Chase & Co.’s Jamie Dimon warned that the private credit industry has a lack of transparency and poor valuation standards, but that he didn’t think private credit was a systemic risk, according to his latest CEO letter.

Wall Street and their private credit peers are deeply intertwined. Credit funds rely on banks to safeguard and custody assets. They also need banks for lines of credit. If private credit portfolios sour, this puts the collateral banks are lending against at risk. 

Blackstone Private Credit Fund had a debt-to-equity ratio of 0.7 times at the end of 2025, while Blue Owl Credit Income Corp.’s was 0.8 times as of Feb. 28. KKR FS Income Trust’s was about 0.7 times at the end of February.

Insurance Firms

The Fed questioning comes on top of another initiative at the Treasury Department to question insurers about their exposure. The regulator has put together a team to handle this, according to people familiar with the matter.

The Treasury is planning to meet with state regulators, which directly oversee insurers in the US, to discuss emerging risks and outlooks for the sector, the agency said in an April statement. The Treasury also expects to discuss it with international regulators, it said.

The review is expected to continue over the coming months and some financial firms may hold their own meetings with Treasury, the people said. 

In the last decade, insurance companies have fueled the rise of nonbank lenders, handing them more influence over vast pools of cash. Private credit players have used that money to make loans to businesses and parked them in complex investment structures.

This story was originally featured on Fortune.com

Instead of righting the ship, Mazzucato and Collington argued, these consultants created just an “impression of value,” an illusion of helpfulness, and little else, all while the government and private companies burned money to hire them. 

In an era of AI that promises to save companies cash by automating white-collar jobs, the use of chatbots for guidance may be an appealing alternative for firms no longer willing or able to shell out for consultants. But emerging research shows that while you can ask AI what you would a consultant for a fraction of the price, its advice may not be worth taking either. In fact, AI assistance might just present an old problem in a new medium.

A recent study led by the Esade Business School at the Universitat Ramon Llull in Barcelona found that when various large language models (LLMs) were asked to provide guidance on a workplace issue, they gravitated toward a response that was most aligned with buzzwords, rather than providing guidance that best aligned with the scenario. Researchers dubbed the proclivity of AI to gravitate toward the same jargon to inform their judgments “trendslop.”

“An LLM is not the colleague who critically evaluates current ideas, looks into the contextual specifics, stress-tests assumptions, and pushes back when everyone gets comfortable,” the study authors wrote in a Harvard Business Review post summarizing their research. “On strategy, LLMs might be more akin to a freshly minted MBA or junior consultant, parroting what’s popular rather than what’s right for a particular situation.”

Recent layoffs among the Big Four consultancies, amid a wider industry slowdown, have suggested firms may already be losing value in the view of potential clients. PwC slashed 150 business support staff in November 2025, around the same time that McKinsey shed hundreds of jobs. 

“As our firm marks its 100th year, we’re operating in a moment shaped by rapid advances in AI that are transforming business and society,” a McKinsey spokesperson told Bloomberg last year.

But the emergence of “trendslop” suggests AI is far from able to provide direction to companies seeking counsel from the technology, and this research exposes the bias LLMs struggle with.

How ‘trendslop’ manifests

In order to measure AI’s tendency to give responses aligning with trends rather than logic, researchers tested seven models, including GPT-5, Claude, Gemini, and Grok, across 15,000 simulations and scenarios. Models were asked to choose between two solutions when presented with workplace tensions, such as if a company should prioritize long-term versus short-term growth, or if a firm should use technology to automate versus augment workers’ jobs.

Researchers predicted that if LLMs were providing advice based on the situation-specific details, there would be diversity in which solution the models choose. Instead, the seven models usually clustered their answers around the same strategy, indicating a preference for “modern managerial buzzwords and cultural tropes.”

Even when researchers reworded prompts or asked for pros-and-cons analysis, the AI models, in many cases, demonstrated a strong preference toward a similar business strategy. The study authors warn relying on AI as a consultant will not result in bespoke business solutions, but rather a cookie-cutter solution it could propose to any business when prompted, regardless of the specifics of a presented challenge. 

“This reveals a real risk for leaders,” the researchers said. “An LLM can sound highly tailored to your situation while quietly steering you toward the same small cluster of modern managerial trends.”

Exposing LLM bias

The “trendslop” tendencies of LLMs are a result of biases they take on when the models are being trained, researchers noted. Because LLMs are trained on heaps of information from internet texts to social media to news, they tend to cling to the positive or negative connotations attached to certain phrases or concepts, deeming “commoditization” as outdated and negative, and “augmentation” as progressive and positive.

In other words, when prompted to provide guidance on a tricky workplace scenario, AI isn’t analyzing the situation in question, it’s regurgitating key phrases based on how often it encountered them while being trained on data. In the case of ChatGPT, the study noted, the bot sometimes rejected providing a binary choice, instead recommending both solutions. Research published in Nature last year found AI sycophancy isn’t just unproductive, it can be harmful to science, confirming the biases of those prompting it instead of presenting users with data supported from scientific literature or other reliable, more impartial sources.

The “trendslop” researchers did not completely eschew the use of LLMs in navigating tricky workplace situations. They suggested models could still be helpful in generating alternative solutions or identifying blind spots in certain scenarios. If you’re aware of AI’s biases toward concepts like augmentation or long-term strategizing, you can challenge those biases to reveal more insightful guidance, according to the study.

“Leadership is ultimately about making hard choices in conditions of uncertainty and taking responsibility for them,” the researchers said. “AI cannot and should not be a substitute.”

This story was originally featured on Fortune.com

The 62-year-old Amazon founder is worth $254 billion, less than only Tesla CEO Elon Musk and Google cofounder Larry Page, according to the Bloomberg Billionaires Index. Yet during nearly his entire 25-year tenure as CEO of the top company on the Fortune 500, Bezos has received the same salary of $81,400 from Amazon—less than the average construction worker, according to the Bureau of Labor Statistics. 

Amazon’s 2026 proxy, filed with the Securities and Exchange Commission Thursday, confirmed Bezos still makes the same wage he has made since 1998, the year he boosted his salary to $81,400 from $79,197 for the prior year. 

“Mr. Bezos requested not to receive additional compensation and has never received annual cash compensation in excess of his current amount,” the filing reads. 

His low salary has also not kept up with inflation. In 1998, Bezos’ $81,400 yearly salary was more than double the median salary of $31,096 for men at the time. That same salary was only 16% higher than the median salary of $68,952 for men as of last year.

Part of the reason Bezos has kept his salary low and has never received stock compensation from Amazon is because he owned a large chunk of the company anyway.

“I already owned a significant amount of the company, and I just didn’t feel good about taking more,” Bezos previously said during a New York Times DealBook Summit interview. “I just felt, ‘How could I possibly need more incentive?’ I just would have felt icky about it.”

Bezos, who was replaced as CEO by Andy Jassy in 2021 and now serves as executive chairman, owns about 8% of the company and is its largest shareholder, according to Amazon’s most recent proxy statement. At Friday’s stock price, Bezos’ stake in Amazon was worth approximately $225 billion, a majority of his net worth. Because most of Bezos’s wealth is tied up in his stock and other assets, he has previously been able to pay less than his true tax rate, reported ProPublica in 2021. 

All of Amazon’s executive salaries focus more on stock compensation than cash on purpose. The company in its proxy statement said these salaries, which are meant to be “significantly less than those paid to senior leadership at similarly situated companies” is “to closely tie total compensation to long-term shareholder value.” 

Jassy, the company’s current CEO, earned $365,000 last year, which is the “base salary” for all of Amazon’s named executive officers, including the CEO of Amazon Web Services, Matt Garman, and the CEO of Worldwide Amazon Stores, Doug Herrington, according to the proxy statement. At the same time, Jassy’s stock-based compensation increased by nearly $473,000 between 2024 and 2025, according to the filing. 

To be sure, Bezos also receives some perks thanks to his role as executive chairman. The company last year paid $1.6 million for Bezos’s security expenses and business travel. 

“We believe that all Company-incurred security costs are reasonable and necessary and for the Company’s benefit, and that the amount of the reported security expenses for Mr. Bezos is especially reasonable in light of his low salary and the fact that he has never received any stock-based compensation,” the company wrote in the proxy statement.

This story was originally featured on Fortune.com

The drop coincides with news that Corey Caplan, CTO of World Liberty Financial, which was founded in 2024 by a team including multiple members of the Trump family, used the project’s reserve of WLFI tokens to make loans on a third-party crypto platform he cofounded.

Blockchain data shows wallets controlled by World Liberty Financial lent WLFI tokens worth hundreds of millions of dollars to a decentralized platform named Dolomite, borrowing a mixture of stablecoins in return. Dolomite is just the 13th-largest crypto lending platform, according to DefiLlama data, making it perhaps an unlikely place for World Liberty Financial to lend its tokens.

After CoinDesk first reported on the Dolomite loans Thursday, WLFI’s price tumbled nearly 15%. 

The size of World Liberty Financial’s loans on Dolomite pose a risk to the token’s price, Nicolas Vaiman, CEO at crypto analytics firm Bubblemaps, told Fortune. Roughly 5% of WLFI’s supply is now collateral on Dolomite, so if WLFI declines significantly in value, the collateral could be liquidated, Vaiman said. This would likely force World Liberty to sell WLFI tokens to repay the loan, exerting additional downward pressure on the token’s price. 

World Liberty Financial acknowledged its lending activities in a post to its X account but tried to quell investor worries, asserting that its loan positions are “nowhere near liquidation.” World Liberty called itself the “anchor borrower” for WLFI, arguing that it is “generating the yield that makes WLFI markets compelling for everyone else.”

Hedge funds and foundations exposed to WLFI

Multiple crypto foundations may have purchased WLFI tokens following token swap agreements cut by World Liberty Financial. Nasdaq-listed Alt5 Sigma also raised $1.5 billion to purchase WLFI tokens last summer, drawing investment from institutional buyers including Point72 and ExodusPoint. 

Some investor worries also stem from the amount of World Liberty Financial’s USD1 stablecoin that the team borrowed against its WLFI. World Liberty Financial borrowed so much USD1 from Dolomite that there is little left to borrow, meaning users who previously deposited the stablecoin on Dolomite may have trouble withdrawing, Vaiman said. 

Still, World Liberty Financial projected confidence on social media Thursday. 

Even “if markets moved dramatically against us, we’d simply supply more collateral,” World Liberty Financial wrote. “That’s not a risk. That’s how this works.”

This story was originally featured on Fortune.com

The prospect of tolls has inflamed President Donald Trump, but it’s still unclear whether the Iranian government is charging fees across the board. Meanwhile, ship traffic through the strait is still at a fraction of its volume since the beginning of the war in late February.

Here’s what we know about the tolls, what we don’t know, and why Iran has likely chosen cryptocurrency as its preferred payment method:

What has Iran said about crypto tolls?

Representatives for the Islamic Republic have stated publicly that they plan to have ships in the Strait of Hormuz pay tolls in digital assets. In early April, the parliament released a plan that said as much, according to Iranian state media. 

Separately, Hamid Hosseini, spokesperson for Iran’s Oil, Gas, and Petrochemical Products Exporters’ Union, has told multiple outlets that the Iranian military plans to charge oil tankers for passage in crypto. 

What cryptocurrencies are Iran collecting?

The comments from Iranian officials come after reports of a more informal system of fees. In early April, Bloomberg said that members of the Islamic Revolutionary Guard Corps, an arm of the state’s military, were charging oil tankers $1 a barrel to pass through the Strait of Hormuz.  

Crypto analytics firm TRM Labs stated in a Wednesday report that Iran’s military has charged ships passing through the strategic choke point up to $2 million since mid-March, accepting payment in a variety of fiat and digital currencies: Chinese yuan, Bitcoin, and potentially the stablecoin USDT. (Stablecoins are cryptocurrencies pegged to real-world assets like the U.S. dollar.)

Still, it’s unclear if Iran is using crypto extensively for toll payments. “This is just an incredibly fast moving situation, really, in the midst of a war,” Ari Redbord, global head of policy at TRM Labs, told Fortune. “And we are not seeing on-chain evidence today that indicates that toll payments are being made at scale.”

Why would Iran want to use cryptocurrency?

For decades, the U.S. and its allies have imposed wide-ranging sanctions on Iran that largely exclude the government from the world’s financial system. That’s led the Islamic Republic to avoid currencies or banks linked to the U.S. 

That includes cryptocurrencies, which are built on financial rails not controlled by any one government. And while the public nature of blockchains makes it possible to trace the movements of Bitcoin and stablecoins, those funds can’t be easily seized. At the same time, Iran has become adept at using chains of digital wallets to obscure crypto funds tied to the regime.

“What we’ve seen from Iran over the last really couple of years is them looking for any way to circumvent the U.S. financial system,” said Redbord. “That means accepting payments for things in Chinese yuan. That means starting to really look to cryptocurrency to accept payments.”

How has Iran used crypto in the past?

Iran’s crypto ecosystem grew to $7.8 billion in 2025, according to a report from crypto analytics firm Chainalysis. And, in the fourth quarter of that year, the Islamic Revolutionary Guard Corps accounted for about half of Iran’s total crypto ecosystem.

That’s led the U.S. government to crack down on the Islamic Republic’s attempts to circumvent sanctions. In January, the U.S. Treasury sanctioned the crypto exchanges Zedcex and Zedxion for facilitating transactions for the Iranian military. And the Justice Department is probing Iran’s use of crypto exchange Binance to sidestep U.S. sanctions, according to the Wall Street Journal.

This story was originally featured on Fortune.com

The photos were released as part of a cyberattack on Patel’s personal email that revealed more than 300 messages dated between 2010 and 2019, as well as a work résumé and travel documents. The Iran-linked and pro-Palestinian hacker group Handala Hack Team claimed responsibility for the attack. 

Targeting high-profile figures like Patel is part of Iran’s larger war strategy to sow disruption in the U.S. and Israel, according to experts. 

Handala’s attack against Stryker on March 11 put the medical technology company’s 56,000 employees operating in 61 countries at a standstill, while order processing, manufacturing, and shipping were halted. The company was not fully operational for three weeks following the attack, which it reported had a material impact on first-quarter earnings. 

Earlier this week, the FBI, the National Security Agency, the Cybersecurity and Infrastructure Security Agency, and the Department of Energy issued a joint advisory, warning Iran-backed hackers were targeting critical infrastructure, including water and power plants. 

The agencies did not name specific targets but said that the hacks aimed to “cause disruptive effects” and had already led to “operational disruption and financial loss.”

The warning is a signal to the private sector in particular to take this threat seriously, as it operates most of U.S. critical infrastructure, said Nikita Shah, a senior fellow at the Center for Strategic and International Studies, who worked as a national security official in the U.K. government for 10 years. 

In addition to the water and energy sectors, disrupting the tourism industry, by defacing an airline’s website for example, is another likely target, she added. 

Instead of providing a military advantage for Iran, such low-level attacks on citizens and organizations are meant to cause friction and inflict costs in the hope that they will put pressure on governments to rethink any participation in the war, Shah told Fortune.

“What they’re trying to do is go after low-hanging fruit, so things that will seem very sophisticated on the outside, but from a technical perspective, when you look into it, actually, they’re not especially sophisticated,” she said.

How Iran-backed hackers find their targets 

In March, Iran’s Islamic Revolutionary Guard Corps published a list of potential office and infrastructure targets in the Middle East run by U.S. companies, including Google, Microsoft, Palantir, IBM, Nvidia, and Oracle.

But cyberattacks could hit much closer to home, said Robert Olsen, chief operating officer and managing director of cybersecurity firm Hilco Global Cyber Advisors. 

“If the ultimate goal, in this case of Iranian-sponsored threat actors, is to instill terror and uncertainty in the American population, there’s no better way to do that than through critical infrastructure attacks because it truly touches everyone’s lives in some way, shape, or form,” he told Fortune. “It becomes very personal when the local water system goes down.” 

Iranian hackers are not running highly complex attacks, he said, but rather, taking advantage of companies’ vulnerabilities. In the case of one attack that exposed nearly 3,900 U.S. devices, the hackers took advantage of an open port on a physical piece of equipment, which Olsen said is akin to using an open window to get into somebody’s house. 

“The challenge is organizations have to be pretty much perfect when it comes to all of the different aspects of building an effective security program,” he said. “The threat actors only have to be lucky once.” 

Cyberattacks have also become much easier in recent years, Olsen pointed out. A hack that would have required a PhD level of knowledge years ago can be easily executed owing to developers simplifying their technology. Now, AI is accelerating the access and scale of cyberattacks, he said.  

The Iranian strategy: Projecting power

In addition to cyberattacks, Iran is engaging in “information warfare,” by posting fake videos on social media as a means to project power in place of traditional military capabilities that have been decimated, Shah said.

Gen. Dan Caine, chairman of the Joint Chiefs of Staff, said this week that the U.S. military has hit more than 13,000 targets and has destroyed 80% of Iran’s air defense systems. 

Shah said while the cyberattacks may have little effect on military outcomes, more attacks are likely coming. 

“[It] very much depends on [Iranian] internet capacity, but we should definitely expect to see more targeting of companies or organizations that belong to countries participating in this conflict, because in many ways, the collateral damage is the point,” she said.

This story was originally featured on Fortune.com

Bessent and Powell assembled the group of high-powered execs at the Treasury’s headquarters to ensure banks were aware of the cyber risks presented by Anthropic’s new model, Mythos, and similar future models, reported Bloomberg and the Financial Times. Sources who spoke to Bloomberg said those in attendance included Citigroup CEO Jane Fraser, Morgan Stanley CEO Ted Pick, Bank of America CEO Brian Moynihan, Wells Fargo CEO Charlie Scharf, and Goldman Sachs CEO David Solomon. JPMorgan CEO Jamie Dimon was also invited but was unable to attend, the sources said.

The Federal Reserve declined to comment to Fortune. The Treasury didn’t immediately respond to Fortune’s request for comment.

The meeting comes just weeks after Fortune exclusively reported Anthropic was developing an unreleased model described by the company as “by far the most powerful AI model” it had ever developed, the existence of which Anthropic inadvertently made public last month through its content management system. Later, the company acknowledged that model was Claude Mythos. 

Anthropic on Tuesday released a report titled “Assessing Claude Mythos Preview’s cybersecurity capabilities,” noting how the model was able to find many 10- and 20-year-old vulnerabilities, as well as a 27-year-old vulnerability in OpenBSD, an operating system that has a reputation for being one of the most secure. 

Anthropic briefed senior U.S. government officials and industry stakeholders on Mythos Preview’s capabilities ahead of its release, someone with knowledge of the matter told Fortune. In a blog post, the company said it is willing to work with officials at all levels of government to ensure national security is a priority when rolling out new AI models, and that the U.S. maintains a lead in AI technologies.

Anthropic told Fortune that partnering with the government was the company’s plan from the start (the company is currently in a legal battle with the Pentagon after the Defense Department blacklisted it for placing restrictions on use of its AI technologies).

Anthropic’s new AI model could ‘reshape cybersecurity,’ so it launched Project Glasswing 

In partnership with JPMorgan Chase, Amazon, and Google, along with other key tech companies, Anthropic launched Project Glasswing this week, an initiative aimed to secure critical software amid AI advancements. As part of the partnership, Anthropic said in a blog post, it would share what it learns with the tech and financial services industries. 

Aside from its partners, the company has extended access to 40 additional organizations who build or deploy critical software infrastructure, and committed up to $100 million in Mythos Preview usage credits, the most basic AI usage unit. Anthropic noted the project was launched in response to the capabilities the company has observed in its new frontier model, Mythos, that it believes could “reshape cybersecurity.” 

“Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely,” the post warned. “The fallout—for economies, public safety, and national security—could be severe.” 

Anthropic said in the post Tuesday it would release the new model initially to a limited group of industry partners to “enable defenders to begin securing the most important systems before models with similar capabilities become broadly available.”

Other business leaders have sounded the alarm on the cybersecurity risks powerful AI models pose. Dimon outlined in his annual letter to shareholders the cybersecurity risks major industries and corporations face from AI, saying cybersecurity “remains one of our biggest risks.” He added, “AI will almost surely make this risk worse.”

As part of Project Glasswing, JPMorgan aims to reduce cyber risks stemming from AI’s fast-evolving capabilities. “Promoting the cybersecurity and resiliency of the financial system is central to JPMorgan Chase’s mission, and we believe the industry is strongest when leading institutions work together on shared challenges,” Pat Opet, JPMorgan chief information security officer, said in a statement. 

“Project Glasswing provides a unique, early stage opportunity to evaluate next-generation AI tools for defensive cybersecurity across critical infrastructure both on our own terms and alongside respected technology leaders,” he added.

This story was originally featured on Fortune.com

Inflation surged nearly a full 1% in a month, according to the latest consumer price index released today by the Bureau of Labor Statistics. Most of that spike comes from costs associated with the war in Iran, such as gas: Energy prices alone climbed 10.9% for the month. Gasoline led all categories by a wide margin, surging 21.2%, the steepest monthly jump since 1967.

On a yearly basis, inflation hit 3.3% in March, up nearly a full percentage point from 2.4% in February. That’s highest annual rate since April 2024, and the biggest one-month increase since 2022.

Economists were comforted by the fact that the energy spike didn’t roll over more into other categories. Core inflation, which strips out volatile food and energy, rose only 0.2% month over month, with a rise of 2.6% year over year, slightly below expectations. “The trajectory is encouraging here and should not be overlooked,” Jeffrey Roach, chief economist for LPL Financial, wrote in a note.

Still, the energy spike has started to bleed through. Coffee, already an inflation sore spot, jumped again in March: The average retail price of ground roast coffee spiked 30.5% year over year to $9.46 per pound, driven by a 40% tariff on Brazilian imports and freight costs skyrocketing as the Strait of Hormuz remained largely inaccessible to shipping.

Nonalcoholic beverages broadly, which is a CPI category that captures coffee, tea, and juice, were up 5.6% year over year. Airfare, apparel, household furnishings, and new vehicles all climbed as well, while tobacco (8%) and hospital services (7.1%) were already running hot.

However, overall, prices for food were more or less flat, while prices for medical care, personal care, and used cars actually fell during the month.

But several economists cautioned against getting too bullish. As John Kerschner, global head of securitized products and portfolio manager at Janus Henderson Investors, wrote in a note, “Given the increase in diesel prices, it is only a matter of time before they bleed through to effect downstream components like food.”

Jamie Cox, managing partner for Harris Financial Group, wrote that core inflation’s “effect on real wage growth will bear the full brunt in April.”

“While I’m glad to see the effects to be less than expected in March, the effects in April are now more likely to be worse,” he added.

This story was originally featured on Fortune.com