News & Updates
The latest news and updates from companies in the WLTH portfolio.
Anthropic probes unauthorised access to Mythos model
Anthropic says it is investigating unauthorised access to its Mythos Artificial Intelligence model, according to French news agency AFP. Mythos is intended to identify security holes that have existed for decades but now appears to be vulnerable itself. Earlier this month, Anthropic made a restricted release of Mythos to 40 major technology firms, to give them a head start in fixing cybersecurity vulnerabilities before they could be exploited by attackers.
From defence tool to risk factor: Understanding Anthropic's Mythos AI model
Anthropic's Mythos AI model is built to identify vulnerabilities in software systems US-based artificial intelligence entity Anthropic's unreleased AI model, Mythos, is not supposed to be widely accessible. Built under the company's Project Glasswing as a tightly controlled cybersecurity system, it is still in limited testing and restricted deployment. However, some of the results that it has produced are astonishing to say the least. In one internal exercise, the model identified 271 previously unknown vulnerabilities in Mozilla's Firefox codebase, according to a report by Ars Technica. These were not routine bugs flagged by automated scans, but exploitable weaknesses that had persisted despite years of active development and security review. That kind of output changes how vulnerability discovery works in practice. It increases the volume of issues that can be surfaced, compresses the time needed to find them, and shifts the balance between what is visible and what remains hidden inside complex systems. At the same time, Mythos is beginning to surface outside its intended boundaries. Reports from TechCrunch suggest that tools linked to the model may have already been accessed beyond authorised channels, while other accounts indicate that government-linked actors are exploring similar capabilities. Within the AI industry, there is no clear agreement on how to interpret this moment. OpenAI CEO Sam Altman has dismissed some of the concern as "fear-based marketing," even as others frame models like Mythos as a step-change risk. Also Read Singularity - Learn introduces real-time academic tracking for parents Amazon to invest another $5 bn in Anthropic as part of $100 bn cloud deal Govt body plans legal framework for AI companies, tech oversightpremium 3 in 4 global leaders to prioritise AI investment despite uncertainty OpenAI's India test: Can massive AI use scale turn into a paying market? What is Mythos Anthropic has not released a full technical paper on Mythos, but its own disclosures through Project Glasswing and the Mythos preview outline the model's intended role with unusual clarity. Mythos is not a general-purpose AI model adapted for security tasks. It is built specifically for analysing software systems in adversarial ways - examining how code behaves under stress, where assumptions fail, and how those failures can be turned into real vulnerabilities. ALSO READ: Meta to track employee activity on computers to train its AI agents: Report Under Glasswing, access to models like Mythos is limited to a small group of external partners, including security researchers, enterprise teams, and organisations working on infrastructure protection. Reuters reporting on Anthropic's cyber efforts notes that such systems are being handled with caution, reflecting concerns about how easily they could be misused if broadly available. How Mythos works inside a system Anthropic's preview of Mythos describes a model that operates through interaction rather than inspection. Instead of scanning code for known patterns, the system engages with it. It executes functions, tests different inputs, and observes how the software responds. Each result feeds into the next step, allowing the model to refine its approach and move deeper into the system. This creates a continuous loop of testing and adjustment. A failed attempt does not end the process; it informs the next one. Over time, this allows the model to map out behaviour that would be difficult to capture through static analysis alone. Another key element is how it handles complexity. Modern software systems are not made up of isolated components. Vulnerabilities often emerge from how different parts interact. Mythos is designed to follow those interactions, tracing how a small issue in one part of the system might combine with another to create a more serious flaw. This is where the model's reasoning becomes important. It is not just identifying individual bugs, but evaluating how those bugs could be used in practice. The company has also suggested that Mythos can operate alongside external tools, allowing it to simulate environments and validate its own findings. That gives it a way to move beyond theoretical vulnerabilities and confirm whether they can be exploited. What Mythos has already uncovered Anthropic says Mythos has identified thousands of zero-day vulnerabilities, previously unknown flaws, in critical software systems. These include vulnerabilities in every major operating system and web browser, as well as widely used libraries and infrastructure components. Some of the examples are particularly striking. In OpenBSD, a security-focused operating system, the model uncovered a 27-year-old vulnerability that could allow remote attackers to crash systems. In FFmpeg, a widely used media processing library, a flaw was identified that had persisted for 16 years despite extensive automated testing. The model has also demonstrated the ability to chain multiple vulnerabilities together. In the Linux kernel, for instance, it combined separate weaknesses to escalate privileges from a standard user account to full system control. In another case, it autonomously developed a remote code execution exploit in FreeBSD, allowing an attacker to gain root access without authentication. Anthropic's internal evaluations suggest that Mythos significantly outperforms its previous models on these tasks, achieving a much higher success rate at turning vulnerabilities into working exploits. Why this changes cybersecurity Finding critical flaws in complex systems requires specialised knowledge, and even skilled researchers can only examine a limited portion of a codebase at any given time. As a result, many vulnerabilities remain hidden - not because they are impossible to find, but because they have not yet been explored in sufficient depth. ALSO READ: OpenAI releases ChatGPT Images 2.0 model for detailed, multi-image outputs Mythos changes that equation by reducing the cost and effort required to identify and exploit vulnerabilities. Models like Mythos expand how much of a system can be tested and how quickly weaknesses can be uncovered. Anthropic notes that AI-driven cyber capabilities are already lowering the barrier to entry, making it easier to perform tasks that previously required advanced expertise. The risks Anthropic acknowledges that the impact of these capabilities may not be evenly distributed, at least in the short term. While tools like Mythos could eventually make systems more secure, the transition period is uncertain. The same capabilities that help defenders find and fix vulnerabilities can also be used to exploit them. This creates a situation where attackers could benefit from faster, more automated ways of discovering weaknesses. Anthropic describes this phase as potentially "tumultuous," with the balance between attackers and defenders still evolving. How this plays out depends on how quickly defensive systems improve and whether safeguards can keep up with the technology Project Glasswing It is the company's attempt to manage the transition. It brings together organisations including Apple, Google, Microsoft, and Nvidia to use Mythos in controlled environments for defensive purposes. More than 40 additional organisations working on critical infrastructure have also been given access. The goal is to give defenders a head start, allowing them to identify and fix vulnerabilities before similar tools become widely available. At the same time, the model's growing visibility raises questions about how tightly it can be controlled. Reports from TechCrunch indicate that tools linked to Mythos may have been accessed beyond authorised channels, raising questions about how tightly such systems can be contained once they move beyond internal testing. In parallel, other reports suggest that government-linked actors, including those associated with the US National Security Agency (NSA), are already exploring or using Anthropic's cyber-focused models despite internal disagreements within defence circles. Within the AI industry, there is still no clear consensus on how to interpret models like Mythos. In comments reported by TechCrunch, OpenAI CEO Sam Altman criticised what he described as "fear-based marketing" around cyber-focused AI systems. This suggests that some view the concerns as overstated, while others see these systems as fundamentally changing the security landscape. What happens next Anthropic has made it clear that Mythos, in its current form, is unlikely to be released as a general-purpose product. Instead, the focus is on developing safeguards that can limit the model's most dangerous capabilities while still allowing it to be used for defensive purposes. Future models may incorporate these protections before being deployed more broadly. What happens next depends less on one model, and more on how quickly the rest of the industry catches up, and how well systems can adapt to a world where finding vulnerabilities is no longer the hardest part. More From This Section Meta to track employee activity on computers to train its AI agents: Report Tesla to bring humanoid robot Optimus to India at 'appropriate' time OPPO Find X9 Ultra, Find X9s set to launch in India in May: What to expect Pixel phones face battery drain issue after April update, Google probes OpenAI releases ChatGPT Images 2.0 model for detailed, multi-image outputs
Want SpaceX Exposure? These 3 AI Titans Have Previously Invested in What Should Be Wall Street's Largest-Ever IPO
Three of SpaceX's minority investors -- one of which may be sitting on a greater than $120 billion windfall -- are artificial intelligence (AI) leaders. If you thought the dot-com era had some impressive initial public offerings (IPOs), you haven't seen anything yet. This might be the year of the mega-IPO, with artificial intelligence (AI) large language model (LLM) developers OpenAI and Anthropic both considering going public. However, the crème-de-la-crème is the upcoming IPO of Elon Musk's SpaceX. While investors seemingly can't wait to get their share of the SpaceX universe, there are three artificial intelligence titans you can buy right now -- Alphabet (NASDAQ: GOOGL)(NASDAQ: GOOG), Nvidia (NASDAQ: NVDA), and Cisco Systems (NASDAQ: CSCO) -- which have previously invested directly into one of SpaceX's numerous puzzle pieces. Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue " Image source: Getty Images. When most investors hear the name Alphabet, they likely think of its virtual monopoly in internet search traffic through Google, its streaming platform YouTube, which is the second-most-visited social site on the planet behind Google, or perhaps its cloud infrastructure services platform, Google Cloud. The integration of generative AI and LLM solutions into Google Cloud ramped up its sales growth to 48% in the December-ended quarter. But what you may not realize is that Alphabet is also an early investor in SpaceX. In 2015, Alphabet invested approximately $900 million in SpaceX when it was valued at roughly $12 billion. Although dilutive events modestly reduced Alphabet's ownership, regulatory filings at the end of 2025 show it still had a 6.11% stake in Musk's next trillion-dollar public company. If SpaceX drums up enough investor interest to command a $1.75 trillion to $2 trillion valuation, Alphabet's $900 million initial investment could be worth $107 billion to $122 billion. This has the potential to be one of the largest investment windfalls we've ever witnessed on Wall Street. Image source: Getty Images. The face of the AI revolution, Nvidia, and the largest publicly traded company during the dot-com era, Cisco Systems, are also investors in the SpaceX conglomerate -- more specifically, xAI. Earlier this year, SpaceX acquired AI start-up xAI, which later merged with social media platform X (formerly Twitter). But before these acquisitions and mergers took place, xAI announced $20 billion of Series E funding on Jan. 6. Nvidia and Cisco were two of the strategic investors listed in this late-stage round of funding. According to CNBC in early February, the merger of SpaceX and xAI valued the two businesses at $1 trillion and $250 billion, respectively. This means Nvidia's and Cisco's minority investments are in one of the smaller pieces of the SpaceX pie. Nevertheless, Nvidia and Cisco Systems can increase the value of their investments in SpaceX by simply doing what they do best. Nvidia's graphics processing units (GPUs) are the undisputed top choice by businesses in AI-accelerated data centers. In October 2024, Nvidia announced that xAI was in the process of doubling its GPU cluster, known as Colossus, to 200,000 chips. Meanwhile, Cisco is helping xAI scale its compute infrastructure. Given how incredibly pricey SpaceX is, relative to its estimated sales, these three AI titans -- Alphabet, Nvidia, and Cisco Systems -- offer a potentially smarter way to gain exposure to this record-breaking IPO. Before you buy stock in Alphabet, consider this: The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now... and Alphabet wasn't one of them. The 10 stocks that made the cut could produce monster returns in the coming years. Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $511,411!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $1,238,736!* Now, it's worth noting Stock Advisor's total average return is 986% -- a market-crushing outperformance compared to 199% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors. Sean Williams has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet, Cisco Systems, and Nvidia. The Motley Fool has a disclosure policy.
Unauthorised users hack Anthropic's 'too dangerous to release' AI
Hackers have gained access to Anthropic's Mythos, an AI the company considers too cybersecurity-sensitive to release publicly. A group of unauthorised users reportedly gained access to Anthropic's new product, which the artificial intelligence company says is too powerful to release to the public as it "poses unprecedented cybersecurity risks". Anthropic's new AI technology, Mythos, is designed for enterprise security and is being tested by a few technology and cybersecurity firms. A "private online forum" has managed to gain access to Mythos through a third-party vendor, according to Bloomberg. The company said it was investigating the Bloomberg report, an Anthropic spokesperson told TechCrunch, adding that there was so far no evidence that the reported activity had impacted Anthropic's systems. Members of the unauthorised group are part of a Discord channel that seeks out information about unreleased AI models, Bloomberg reported. Citing a person employed by a third-party contractor that works for Anthropic, Bloomberg added that the group tried several strategies to gain access to the model. The outlet also reported that the unauthorised group had been regularly using Mythos once it gained access. Euronews Next has reached out to Anthropic for comment but did not receive a reply at the time of publication. Anthropic said it would limit the release of its new AI model to a few tech and cybersecurity firms as part of its so-called Project Glasswing. The list includes Amazon, Apple and JP Morgan Chase. Goldman Sachs, Citigroup, Bank of America and Morgan Stanley are reportedly testing the Anthropic model too, according to reports. Treasury Secretary Scott Bessent convened a meeting of senior American bankers in Washington in April to discuss the Mythos model. The meeting encouraged the banking executive to use Antropic's Mythos model to detect vulnerabilities, according to Bloomberg. Goldman Sachs, Citigroup, Bank of America and Morgan Stanley are reportedly testing the Anthropic model too, according to reports.
Sam Altman Takes Aim at Anthropic's "Mythos" Rollout, Calls It Fear Marketing
The ongoing feud between OpenAI and Anthropic got a fresh jab this week. Sam Altman, sitting down for a podcast interview, aimed squarely at Anthropic's latest cybersecurity model, Mythos, arguing his rival is dressing up its product with scare tactics rather than substance. Key Takeaways: * OpenAI CEO Sam Altman accused Anthropic of using "fear-based marketing" to promote its new Mythos cybersecurity model. * Anthropic released Mythos this month only to a small group of enterprise clients, saying broader release could help cybercriminals. * Altman suggested the framing serves to concentrate AI power among a select few, though the AI industry broadly -- including OpenAI -- has leaned on dramatic rhetoric before. Anthropic unveiled Mythos earlier this month and handed access to a narrow slice of enterprise customers. The company's explanation? The model is simply too potent to put in public hands, since bad actors could turn it into a weapon. Plenty of observers have pushed the back on that claim, calling it exaggerated. Speaking on the Core Memory podcast, Altman suggested Anthropic's approach was a convenient way to gatekeep the technology. "There are people in the world who, for a long time, have wanted to keep AI in the hands of a smaller group of people," he said. "You can justify that in a lot of different ways." He didn't stop there. "It is clearly incredible marketing to say, 'We have built a bomb, we are about to drop it on your head. We will sell you a bomb shelter for $100 million,'" Altman added. Of course, Anthropic didn't pioneer this playbook. A good chunk of the AI sector has relied on dramatic warnings and grand claims to boost the perceived power of its products. Apocalyptic talk about AI ending civilization hasn't only come from skeptics and activists -- it's come from the very companies building and selling the tech. Altman's own name appears on that list.
Sam Altman calls out Anthropic's 'fear-based marketing' around Claude Mythos
Ever since its announcement, Anthropic has dubbed its new Claude Mythos model as so powerful that it can act as a potential hacking tool. The company has currently provided the Mythos AI model to a limited batch of software providers through an initiative called Project Glasswing to ensure that these firms test and safeguard their own systems from potential cyberattacks. However, many critics have argued that the company is overblowing Mythos' capabilities, and OpenAI CEO Sam Altman has now spoken against its competitor, Anthropic. In a podcast Core Memory, Altman described such messaging as "fear-based marketing," suggesting it can be used as a justification to keep AI in the hands of a smaller group of people. "There are people in the world who, for a long time, have wanted to keep AI in the hands of a smaller group of people," he said. "You can justify that in a lot of different ways." "It is clearly incredible marketing to say, 'We have built a bomb, we are about to drop it on your head. We will sell you a bomb shelter for $100 million,'" he added, referring to keeping AI within a small group of exclusive elites. Concerns about AI being fatal and hazardous for humans are not new. In 2020, when Anthropic founder Dario Amodei was still with OpenAI, hype was created around GPT-3, which was apparently too dangerous to be released widely. Since then, AI companies -- most notably OpenAI and Anthropic -- have used alarming narratives around their AI systems to generate buzz. Bloomberg recently reported that unauthorized users have had access to Claude Mythos since the first day of its announcement. These users gained access by using various methods, including leveraging the account of a person working with a third party for Anthropic. Bloomberg claims the group is interested in playing around with new models, not wreaking havoc with them. They have not run cybersecurity-related prompts on the Mythos model. This is not the first time Sam Altman has taken a dig at Anthropic and presented OpenAI's products as cheaper and superior alternatives to Claude. Recently, when a user reported changes to Anthropic's Claude Code pricing, Altman responded to the post highlighting that Codex is available in both free and Plus plans, saying he wants people to use a lot of AI. Altman's quoted post was written by Codex chief Thibault "Tibo" Sottiaux in response to the original discussion about Anthropic's pricing changes for Claude Code. Previously, Sam Altman, without naming Anthropic, said he did not want Codex users to face reduced limits. This was seen as a likely dig at Anthropic after Claude Code users reported hitting usage limits much faster than before.
Anthropic's Mythos AI model accessed by unauthorized users
A handful of users in a private online forum gained access to Mythos on the same day that Anthropic first announced a plan to release the model to a limited number of companies for testing purposes A small group of unauthorised users has accessed Anthropic's new Mythos AI model, per a Bloomberg report reported on Tuesday, citing documentation and a person familiar with the matter. A handful of users in a private online forum gained access to Mythos on the same day that Anthropic first announced a plan to release the model to a limited number of companies for testing purposes, the report said.
Traveller chaos as police put Perth Airport into lockdown
Perth's international airport terminal is in lock down this afternoon over the discovery of an unattended item, throwing passenger travel plans into chaos. WA Police and Australian Federal Police officers locked down part of Terminal 1, the international wing of Perth Airport, and have established an exclusion zone while the item in the forecourt is investigated. Passengers returning from overseas have faced long stagnant queues since the middle of the afternoon as they wait to clear customs. One told The West Australian they had been waiting close to two hours, since their flight from Bali had landed. "The line was long but was moving," they said. WA's biggest courts and crime stories to your inbox Sign-up to our weekly newsletter for free Sign up "At perhaps 4pm the line stopped moving - zero announcements on PA, zero communication from any staff until perhaps 4.50pm. "Then about 5pm some staff member finally came and told us there was an AFP incident out front and they're not letting us out until its cleared." A Perth Airport spokesperson confirmed the eastern end of the forecourt has been locked down, but said travellers can exit using the western end. Traffic outside is being directed to use the Terminal 2 short term carpark, due to the exclusion zone making the regular area unusable. "Check in areas at Terminal 1 International are currently closed to passengers and staff, and the Terminal 1 short term car park has also been closed, with traffic redirected to the Terminal 2 short term car park," the spokesperson said in a statement. "Arriving international passengers will be escorted from the terminal via alternative doors. "Perth Airport is working closely with the AFP and WAPOL to maintain the exclusion zone while the investigation continues." The AFP has asked members of the public to "avoid the area and await further advice" A spokesperson said more information would be provided "at an appropriate time". More to come. Get the latest news from thewest.com.au in your inbox. Sign up for our emails
Unauthorized access hits Anthropic's Mythos AI model, sparking security concerns | News.az
A group of unauthorized users gained access to Anthropic's artificial intelligence model Mythos, raising concerns about the security of the system, which is capable of supporting advanced cyber-related functions, News.Az reports, citing Bloomberg. The report said a small group of users in a private online forum obtained access to Mythos on the same day Anthropic announced plans to make the model available to a limited number of companies for testing purposes. Since gaining access, the group has reportedly been using the model regularly, though not for cybersecurity activities. Bloomberg cited sources and said it was provided with screenshots and a live demonstration of the model's capabilities. Anthropic has previously stated that Mythos is capable of identifying and exploiting vulnerabilities "in every major operating system and every major web browser when directed by a user to do so." Because of these capabilities, the company has restricted access to a select group of software providers for testing and security evaluation. According to the report, the users allegedly gained entry using a combination of methods, including access through a third-party contractor and tools commonly used by cybersecurity researchers. In response, an Anthropic spokesperson said the company is investigating the claims of unauthorized access involving a third-party vendor environment. The company added that it has no evidence suggesting that the access reported by Bloomberg extended beyond the third-party vendor's environment or that any of Anthropic's internal systems have been affected. Bloomberg also reported, citing sources, that the group involved appears to be interested in experimenting with new AI models rather than causing harm, and has not used Mythos for cybersecurity-related prompts.
Australia and New Zealand central banks monitoring Anthropic's Mythos release
SYDNEY, April 22 (Reuters) - The central banks of Australia and New Zealand said on Wednesday they were monitoring the release of Anthropic's advanced Mythos artificial intelligence model, joining authorities around the world in expressing concerns about the new cybersecurity risks it poses. Designed for defensive cybersecurity tasks, Mythos' vast capabilities have sparked fears about the threat to traditional software security, after Anthropic said a preview had uncovered "thousands" of major vulnerabilities in "every major operating system and web browser." Experts have also warned that the model can identify and exploit previously unknown vulnerabilities faster than companies can fix them. The Reserve Bank of Australia said in a statement it was closely monitoring the development and was "engaging with peer regulators, government and regulated entities." The Reserve Bank of New Zealand said it was also in contact with other regulators both domestically and in Australia over what it called the "developing risk" from Mythos. On Tuesday, Bundesbank President Joachim Nagel called the model a double-edged sword, saying: "it could be used not only to improve digital security systems, but also to leverage their vulnerabilities for malicious purposes." Anthropic has introduced Claude Mythos Preview through a tightly controlled program called Project Glasswing. Access has been granted to major technology companies including Amazon, Microsoft, Nvidia, and Apple. The company has also expanded access to more than 40 additional organisations that build or maintain critical software infrastructure. Experts say Mythos' advanced coding and autonomous capabilities could significantly accelerate sophisticated cyberattacks, especially in sectors like banking, where complex, interconnected, and often decades-old systems remain common. (Reporting by Stella Qiu in Sydney; Writing by Alasdair Pal; Editing by Edwina Gibbs)
SpaceX's IPO Ambitions Put A $1.75 Trillion Price Tag On Private Markets
The story isn't just about rockets: it's about how private investors are putting mega-cap style prices on a handful of tech firms before public markets get a say. Reuters reported SpaceX was valued near $1 trillion in recent private transactions, while PitchBook has pegged it higher, around $1.25 trillion - which makes a $1.75 trillion target feel less impossible, even if it's still a jump. The same dynamic is playing out in artificial intelligence: firms l.. ike OpenAI and Anthropic have been discussed as potential future listings, and any move toward the stock market would test whether everyday investors will pay the same lofty prices private backers have been willing to accept. If they don't, it could force a broader reset in late-stage startup valuations. Why should I care? For markets: Private pricing is setting up a public reality check. Private valuations come from negotiated funding rounds, where shares don't trade freely and price discovery is murkier than it is on an exchange. If headline names in space, AI, or fintech list at or near their private marks, public investors will decide whether those expectations survive life under quarterly earnings pressure and daily trading. A weak debut wouldn't just hit one stock - it could ripple through the wider IPO pipeline by making other companies and their backers accept lower prices. Zooming out: The next listing wave could rewrite the growth playbook. For years, the biggest growth companies stayed private longer because capital was plentiful and valuations kept climbing. If public markets validate today's private numbers, that model looks sustainable - and stock indexes could end up absorbing huge new "heavyweights" in a short span. But if public investors push back, startups may have to raise less money, cut spending, or delay IPO plans, shifting power from founders and venture capital back toward public-market discipline.
Polymarket traders don't see Kelp socializing losses after $292 million exploit
Implementing a system-wide loss redistribution, similar to Bitfinex's 2016 hack response or derivatives exchanges' auto-deleveraging, would be technically complex and politically fraught. A Polymarket contract on whether Kelp DAO will spread the losses from the weekend's $292 million exploit beyond those directly affected is pointing to a clear answer: probably not. Bettors are giving a 14% chance that Kelp will "socialize the losses," or implement a mechanism forcing rsETH holders on Ethereum, which wasn't hit, to share the pain of users on other chains. The attackers drained roughly 116,500 rsETH from a LayerZero-powered bridge that held the reserves backing the token across more than 20 blockchains. That left parts of the system undercollateralized, with some holders effectively owning tokens no longer fully backed by ether (ETH). "Socializing the losses" would mean Kelp redistributes the shortfall across all rsETH holders, including those on the Ethereum mainnet, rather than leaving losses concentrated among users and protocols tied to the compromised bridge. The most widely cited precedent of this approach came in 2016, when Bitfinex imposed losses on all users after a $60 million hack, effectively mutualizing the hit to avoid shutting down. More recently, derivatives exchanges have used variations of the concept through auto-deleveraging (ADL), in which profitable positions are forcibly reduced to cover losses when insurance funds are exhausted. During the October flash crash, ADL mechanisms were triggered across some venues, closing out even market-neutral positions and leaving traders exposed. These moves are rare and controversial, but they have been used as a last resort to stabilize systems under stress. Kelp's situation is more complex. The exploit drained the reserve backing rsETH across more than 20 chains, leaving losses fragmented across different user groups and platforms. Holders on affected networks face impaired backing, while others remain relatively insulated. Any attempt to equalize losses would require coordination across chains, clear accounting of liabilities, and a willingness to impose losses on users who may not see themselves as affected. That makes a clean, system-wide redistribution both technically and politically difficult, which may explain why Polymarket traders are approaching the question with skepticism.
Anthropic's Unreleased Claude Mythos Found 271 Firefox Bugs
Coinbase is reportedly pursuing Mythos access as exchanges brace for AI threats. Mozilla shipped Firefox 150 this week with patches for 271 security vulnerabilities discovered by Anthropic's Claude Mythos Preview in an initial evaluation. The scan forms part of Project Glasswing, Anthropic's coordinated defense effort that grants limited Mythos access to critical infrastructure partners. Mozilla Patches 271 Vulnerabilities After Claude Mythos Evaluation In a recent blog post, Firefox CTO Bobby Holley explained that browser security has traditionally followed an offense-heavy model. Under this approach, vendors acknowledged that fully eliminating exploits was unrealistic and instead focused on making attacks so costly or complex that they would not be widely abused. "As these capabilities reach the hands of more defenders, many other teams are now experiencing the same vertigo we did when the findings first came into focus. For a hardened target, just one such bug would have been red alert in 2025, and so many at once makes you stop to wonder whether it's even possible to keep up," Holley said. Follow us on X to get the latest news as it happens The executive stated that since February, the Firefox team has been working intensively with advanced AI tools to identify and remediate "latent security vulnerabilities in the browser." Earlier collaboration with Anthropic, using its Opus 4.6 model, led to fixes for 22 security-sensitive issues in Firefox 148. The latest update represents a sharp escalation in scale, roughly a twelvefold increase, highlighting how AI-driven audits are reshaping modern cybersecurity practices. "Encouragingly, we also haven't seen any bugs that couldn't have been found by an elite human researcher,' he added. Why the Firefox Result Matters for Crypto The Firefox evaluation lands as exchanges weigh their own exposure to AI-assisted attacks. Anthropic says Mythos can "identify and then exploit zero-day vulnerabilities in every major operating system and every major web browser when directed by a user to do so." This marks the same surface that hot wallets and decentralized applications depend on. While private keys are generally protected within wallet environments, attackers can still gain control over on-chain assets by tricking users into approving harmful transactions or exploiting compromised extensions. Interest in such capabilities is already expanding. Coinbase has reportedly explored access to Anthropic's Mythos. This builds on its existing use of Claude models for customer support across more than 100 regions.
Anthropic's most dangerous AI model just fell into the wrong hands
Anthropic's Mythos AI model, a powerful cybersecurity tool that the company said could be dangerous in the wrong hands, has been accessed by a "small group of unauthorized users," Bloomberg reports. An unnamed member of the group, identified only as "a third-party contractor for Anthropic," told the publication that members of a private online forum got into Mythos via a mix of tactics, utilizing the contractor's access and "commonly used internet sleuthing tools."
Fortune Tech: SpaceX-Cursor deal, Meta tracking software, Sullivan & Cromwell AI hallucinations | Fortune
Good morning. Fascinating indeed to absorb the world's reactions to yesterday's news that John Ternus will become the next CEO of Apple. One that stuck with me? From none other than Tony "The iPod-father" Fadell, who offered as clear a north star as any for the company's next chapter: "People don't choose products," he wrote. "They choose experiences they can trust. Apple has always been built on that trust with users, with teams, and in products. This transition is continuing that trust." The actual deal is a $10 billion collaboration to develop "coding and knowledge work" artificial intelligence, pairing Cursor's AI expertise with SpaceX's Colossus supercomputer. But if all goes well, SpaceX retains the right to acquire Cursor "later this year for $60 billion." No pressure, Cursor CEO Michael Truell! Cursor's ascendance has been swift, with a big splashy feature in Fortune magazine and an almost $30 billion valuation at the end of last year. The San Francisco company has reportedly been in talks to raise $2 billion at a valuation of $50 billion; with that kind of trajectory, SpaceX was wise to strike a deal now. Still, there was an air of inevitability to SpaceX -- which acquired xAI in February -- taking Cursor off the table. As Fortune's Allie Garfinkle wrote last month: "Investors disagree about Cursor's likely fate. Some say the only natural end point is an IPO, which would require the company to get its unit economics majorly in order. Others say it could be a mega-acquisition from OpenAI, which reportedly looked into acquiring Cursor early last year." Surely Elon would never let that happen, eh? -- AN Meta is installing tracking software on U.S. employees' work computers that will capture mouse movements, clicks, and keystrokes, along with some screenshots to feed the data into its AI training pipeline, according to Reuters. The tool, disclosed in a memo to staff this week in a channel belonging to the Meta Superintelligence Labs team, which Reuters saw, will run on a designated list of work apps and websites. Meta's memo framed the effort as a way for rank-and-file employees to improve company models in areas where they struggle to emulate basic computer-use behaviors, such as navigating dropdown menus and using keyboard shortcuts. The memo told Meta staffers that they can do their part to help by just doing their daily work. The broader goal seems to be to build AI agents capable of performing white-collar tasks on their own, the exact software Meta is racing to ship out amid competition from OpenAI and Anthropic. Those agents have a lot of data, but little footage of how to actually use it. "If we're building agents to help people complete everyday tasks using computers," a Meta spokesperson wrote in an email to Fortune, "our models need real examples of how people actually use them." -- Eva Roytburg A white-shoe U.S. law firm told a federal bankruptcy court that a key filing it made in a high-profile case contained multiple hallucinations made by artificial intelligence. The case in question pits government-appointed liquidators in the British Virgin Islands against Prince Group and its owner Chen Zhi, who was charged last year by U.S. federal prosecutors with wire fraud and money laundering as part of an alleged cyberfraud ring in Cambodia. Sullivan & Cromwell deeply regrets the error, according to a letter written by Andrew Dietderich, one of the country's top business restructuring lawyers. The firm didn't name names in terms of who or what allowed the errors into the filing, but it reportedly has an enterprise license for OpenAI's ChatGPT. The good news? The hallucinations violated S&C's policies on the use of AI. The bad? It still happened, sending the firm back to the drawing board on its governance. As the Financial Times points out, the slip-up is just "the latest example of a professional services firm grappling with the use of cutting-edge technology to speed up laborious research and cut down on staffing while also trying to maintain quality standards." And, of course, justifying their $2,000 per hour billing rates. -- AN -- Microsoft cuts Xbox Game Pass prices. Ultimate drops from $30 to $23 per month, PC slides $2.50 to $14 -- but no future Call of Duty titles. -- Did ChatGPT help plan a mass shooting? Florida's attorney general issues criminal subpoenas to OpenAI to investigate. -- Mozilla's new Firefox has 271 fixes thanks to Anthropic's Mythos AI model. -- Unauthorized users have reportedly had their hands on the limited-access Mythos since launch. -- New York State sues Coinbase and Gemini, alleging that their prediction markets violate state laws against illegal gambling. -- Kalshi may offer crypto trading in the U.S., pitting it against Coinbase.
Hyderabad RTC Strike Halts 2,850 Buses, Metro Crowded As 22 Lakh Commuters Face Travel Chaos
Bus services across Hyderabad came to a halt from midnight on Tuesday after talks between the government and Road Transport Corporation (RTC) unions failed, triggering a major strike that has left daily life in the city severely disrupted. With nearly 2,850 buses staying inside depots across Greater Hyderabad, around 22 lakh passengers are expected to face significant inconvenience, as commuters scramble for alternatives and Metro stations witness an unusual surge in footfall. Over 20,000 RTC Staff Stay Away, Major Bus Hubs Hit Across Hyderabad The strike has impacted public transport on a massive scale, with over 20,000 RTC staff, including drivers, conductors and mechanics from 25 depots in Greater Hyderabad, staying away from duties. As a result, bus services from major hubs such as MGBS, JBS, Dilsukhnagar, Kukatpally, LB Nagar and Uppal to districts across Telangana and parts of Andhra Pradesh remained suspended. The disruption hit passengers who had planned early morning travel on Wednesday, especially those arriving at key railway stations including Cherlapally, Secunderabad, Nampally and Kacheguda. In the absence of city buses, many were forced to depend on autos and Metro services to continue their journeys. First Major RTC Agitation In Seven Years Raises Fears Of Repeat Chaos This is the first major RTC strike in Hyderabad in seven years, reviving memories of the 2019 agitation when employees stayed off work for 52 days and public transport systems across the state were thrown into disarray. During that strike, private transport operators were accused of charging inflated fares, while Metro Rail and MMTS services struggled under heavy passenger loads. Authorities now fear similar conditions may return if the current strike continues, with concerns also emerging over possible increases in the prices of vegetables and other essential commodities due to transport disruptions. Metro Rail Becomes Lifeline As Auto Fares Rise Across The City With RTC buses off the roads, commuters quickly shifted to Hyderabad Metro Rail, leading to heavy crowding at stations across the network from the first morning of the strike. Even though trains were running at short intervals of two to three minutes, stations continued to report high passenger movement throughout the day. Many commuters used the Metro to get as close as possible to their destinations before relying on share autos for the last mile. However, that too became expensive, as auto fares reportedly rose sharply after the strike began, adding further pressure on passengers already struggling with limited transport options. The situation was worsened by seasonal travel demand, with summer holidays and ongoing wedding functions increasing movement across the city. RTC buses typically serve a wide cross-section of Hyderabad's population, from daily wage earners to IT professionals, and women in particular rely heavily on the service because of the free travel scheme. On the first day of the strike, that dependency became especially visible as alternate systems failed to absorb the full load. Government Plans Alternate Services, Commuters Seek More Metro Trains Officials said efforts are underway to reduce the impact of the strike by increasing Metro Rail and MMTS services. Authorities are also preparing to deploy private buses, and school buses may be used if the disruption continues and demand rises further. Transport officials have also warned that strict action will be taken against private operators found charging excess fares during the strike period. Meanwhile, commuters have urged the government to increase Metro frequency further and extend service timings to manage the rush. The spike in Metro users has also led to new problems, including parking shortages at several stations, as more passengers choose to drive to stations before boarding trains. If the strike continues, Hyderabad could see growing pressure on every available transport mode, with Metro Rail remaining the city's primary lifeline for now.
SpaceX secures $60bn option on Cursor as both sides found an enemy of their enemy - Cryptopolitan
Adding Cursor to its AI portfolio, SpaceX strengthens the story it needs to tell investors. SpaceX has struck a deal giving it the right to buy AI coding startup Cursor for $60 billion later this year, or settle for a $10 billion working partnership, as Elon Musk's company tries to close the gap with rivals in one of the fastest-moving corners of the technology industry. The announcement, made Tuesday in a post on X, puts one of Silicon Valley's most talked-about startups squarely inside Musk's expanding orbit, just months before SpaceX is expected to go public in what could be the largest stock market debut in history. Cursor, owned by parent company Anysphere and co-founded in early 2022 by four MIT students, Michael Truell, Aman Sanger, Sualeh Asif, and Arvid Lunnemark, builds tools that use artificial intelligence to help software developers write code faster. The company released its first product in March 2023, and within months, it had spread rapidly through the developer community. By November 2023, it had cataloged 150,000 codebases. In June 2024, it raised a $60 million in Series A funding led by Andreessen Horowitz. From zero to $2 billion in three years What followed was a funding streak rarely seen in enterprise software. Through 2025, Cursor raised three additional rounds totalling $3.3 billion. Its valuation opened 2025 at $2.5 billion and closed the year at $29.3 billion after a $2.3 billion Series D in November. Before that came a $900 million round in June 2025 when it was valued at $9.9 billion. The company is now in talks to raise another $2 billion at a valuation above $50 billion, with Andreessen Horowitz and Thrive Capital expected to co-lead, joined by Nvidia and Battery Ventures. "If you subtract out the dollars invested, it's the fastest-growing company we've ever seen," said Martin Casado, Andreessen Horowitz general partner and Cursor board member. Revenue has grown at a similar pace. Annualized revenue hit $500 million in May 2025, doubled to $1 billion by October, and crossed $2 billion in February 2026. Cursor says its tools are now used by 67% of the Fortune 500, including Uber and Adobe, and generate 150 million lines of enterprise code every day. Jensen Huang, CEO of Nvidia, an investor and partner, told CNBC in October: "My favorite enterprise AI service is Cursor. Every one of our engineers, 100 percent, is now assisted by AI coders, and our productivity has gone up incredibly." A fast rise now under pressure Yet the company's quick growth has landed it in a difficult position. Anthropic launched Claude Code as a research preview in February 2025, and it caught on fast. By early 2026, Claude Code had a $2.5 billion annual run rate and more than 300,000 business customers. The difference between the two products is significant: Cursor helps developers write code faster, while Claude Code writes entire chunks of code on its own. "We invented agentic coding as a thing," said Boris Cherny, Anthropic's head of Claude Code. Social media has begun buzzing with the idea that Cursor is in trouble. One startup, Valon, publicly said in February it was moving off Cursor, setting off a wave of "Cursor is dead" commentary online. Some investors have noticed clients pulling back. Two of Cursor's own engineers, Andrew Milich and Jason Ginsberg, left in March to join SpaceX and xAI. There is also a pricing problem. Cursor pays open-market rates to access AI models from Anthropic and OpenAI, the same companies competing directly against it. "Anthropic is trying to drown out Cursor," one venture capitalist told Fortune. To reduce that reliance, Cursor has been developing its own model, called Composer, since 2025. Composer has outperformed Anthropic's Opus 4.6 on some benchmarks, though Composer 2 came in behind OpenAI's GPT 5.4. A Cursor blog post on Tuesday said model training had been "bottlenecked by compute" and that the SpaceX deal would let it "dramatically scale up" its models using xAI's Colossus supercomputer cluster in Memphis. SpaceX, for its part, has its own reasons to move fast. The company filed IPO paperwork with the SEC in early April and plans a roadshow in early June. It merged with xAI in February in a deal valued at $1.25 trillion and is now seeking a $1.75 trillion valuation, which would make it the biggest IPO ever. It ended 2025 with $24.7 billion in cash. Cursor CEO Michael Truell, 25, said the deal was "a meaningful step on our path to build the best place to code with AI." Whether SpaceX eventually buys the company or not, Truell has said he wants to build something that lasts. In an industry where everything changes every six months, that is a harder task than it sounds.
Want SpaceX Exposure? These 3 AI Titans Have Previously Invested in What Should Be Wall Street's Largest-Ever IPO | The Motley Fool
Three of SpaceX's minority investors -- one of which may be sitting on a greater than $120 billion windfall -- are artificial intelligence (AI) leaders. If you thought the dot-com era had some impressive initial public offerings (IPOs), you haven't seen anything yet. This might be the year of the mega-IPO, with artificial intelligence (AI) large language model (LLM) developers OpenAI and Anthropic both considering going public. However, the crème-de-la-crème is the upcoming IPO of Elon Musk's SpaceX. While investors seemingly can't wait to get their share of the SpaceX universe, there are three artificial intelligence titans you can buy right now -- Alphabet (GOOGL 1.54%)(GOOG 1.47%), Nvidia (NVDA 1.07%), and Cisco Systems (CSCO +2.27%) -- which have previously invested directly into one of SpaceX's numerous puzzle pieces. When most investors hear the name Alphabet, they likely think of its virtual monopoly in internet search traffic through Google, its streaming platform YouTube, which is the second-most-visited social site on the planet behind Google, or perhaps its cloud infrastructure services platform, Google Cloud. The integration of generative AI and LLM solutions into Google Cloud ramped up its sales growth to 48% in the December-ended quarter. But what you may not realize is that Alphabet is also an early investor in SpaceX. In 2015, Alphabet invested approximately $900 million in SpaceX when it was valued at roughly $12 billion. Although dilutive events modestly reduced Alphabet's ownership, regulatory filings at the end of 2025 show it still had a 6.11% stake in Musk's next trillion-dollar public company. If SpaceX drums up enough investor interest to command a $1.75 trillion to $2 trillion valuation, Alphabet's $900 million initial investment could be worth $107 billion to $122 billion. This has the potential to be one of the largest investment windfalls we've ever witnessed on Wall Street. The face of the AI revolution, Nvidia, and the largest publicly traded company during the dot-com era, Cisco Systems, are also investors in the SpaceX conglomerate -- more specifically, xAI. Earlier this year, SpaceX acquired AI start-up xAI, which later merged with social media platform X (formerly Twitter). But before these acquisitions and mergers took place, xAI announced $20 billion of Series E funding on Jan. 6. Nvidia and Cisco were two of the strategic investors listed in this late-stage round of funding. According to CNBC in early February, the merger of SpaceX and xAI valued the two businesses at $1 trillion and $250 billion, respectively. This means Nvidia's and Cisco's minority investments are in one of the smaller pieces of the SpaceX pie. Nevertheless, Nvidia and Cisco Systems can increase the value of their investments in SpaceX by simply doing what they do best. Nvidia's graphics processing units (GPUs) are the undisputed top choice by businesses in AI-accelerated data centers. In October 2024, Nvidia announced that xAI was in the process of doubling its GPU cluster, known as Colossus, to 200,000 chips. Meanwhile, Cisco is helping xAI scale its compute infrastructure. Given how incredibly pricey SpaceX is, relative to its estimated sales, these three AI titans -- Alphabet, Nvidia, and Cisco Systems -- offer a potentially smarter way to gain exposure to this record-breaking IPO.
'Well Done': Vishal Dadlani Supports Mumbai Woman Who Slammed BJP MLA Girish Mahajan Amid Traffic Chaos During Protest
Reacting to the viral clip, Vishal Dadlani praised the woman's courage in standing up against the disruption. Sharing the video on Instagram stories on Wednesday, he wrote, "Well done, ma'am. Every citizen needs to stand up to all politicians like this" Singer-music composer Vishal Dadlani and filmmaker Ashoke Pandit have voiced their support for a Mumbai woman whose confrontation with BJP MLA Girish Mahajan over traffic disruption has gone viral. The incident, which took place on April 21, sparked discussion online after a video showed the woman stepping out of her car and questioning the road blockade caused by a protest march near Worli Dome. The demonstration, organised over the Women's Reservation Bill not being passed in the Lok Sabha, led to severe traffic congestion across key parts of the city. Reacting to the viral clip, Vishal praised the woman's courage in standing up against the disruption. Sharing the video on Instagram stories on Wednesday, he wrote, "Well done, ma'am. Every citizen needs to stand up to all politicians like this." Ashoke Pandit also backed the woman, highlighting the everyday struggles faced by commuters stuck in such situations. He wrote, "It's high time neta's and their political parties understand that there are old & young on the streets struggling to reach their destinations. Some of them suffer from various diseases. More power to this lady." In the video, the woman can be seen confronting Mahajan and urging protestors to move to an open ground instead of blocking busy roads. "Say what you want to say in an empty ground, get out of here, don't stop the traffic. There are hundreds of people waiting, and there is an empty ground right there," she is heard saying. Her reaction struck a chord with many on social media, as commuters were left stranded for long periods due to the protest. The visuals gained traction, with users echoing similar concerns about disruptions caused by political rallies. The protest followed the Lok Sabha's rejection of the Constitution (131st Amendment) Bill on April 11. While 298 members voted in favour and 230 against, the proposal failed to secure the required two-thirds majority of 326 votes. The bill aimed to expand the House's strength from 543 to 850 seats and enable delimitation based on post-2026 Census data. Narendra Modi was present during the voting. The viral moment has now reignited conversations around balancing the right to protest with minimizing inconvenience to the public.
Mars or the Moon or A.I.? Elon Musk's Changing Goals for SpaceX.
For years, SpaceX's mission was clear: Get humans to Mars. "The most powerful thing we could do is establish a second, self-sustaining civilization outside of Earth," Elon Musk, SpaceX's chief executive, told Forbes in 2003, a year after founding the company. "And the only place that's really feasible is Mars." As a reminder of that goal, SpaceX has a mural in a cafe at its Hawthorne, Calif., campus featuring the progression of human settlement on the Red Planet. The company also sells "Occupy Mars" T-shirts, which Mr. Musk has regularly worn in public. But over the last six months, Mr. Musk has shifted SpaceX's priorities. Though the tech mogul once forecast that humans would take off for Mars as early as 2024, he has de-emphasized reaching the planet. Instead, SpaceX on Tuesday said it had struck a deal with the artificial intelligence start-up Cursor that could result in its acquiring the young company for $60 billion. And Mr. Musk, 54, has proposed other moonshots that could drive more attention and investment to SpaceX as it prepares for one of the largest-ever initial public offerings. Among his pronouncements are A.I. data centers that could orbit Earth, moon-based factories and an A.I. chip manufacturing plant, all of which will contribute to a utopian future where humans never have to work, he has said. This week, some investors and fund managers are expected to get a closer view of those plans when they visit SpaceX's facilities in Texas and Tennessee before the I.P.O., one person who was invited said. Some investors were also scheduled to visit SpaceX's Hawthorne campus next week, the person said. The changing goals have caused whiplash. "It's a hallucinogenic business plan," said Ross Gerber, the chief executive of Gerber Kawasaki, an investment firm that owns SpaceX shares. He added that Mr. Musk "has lost his mind" as he tries to drum up excitement for the public offering. Shifting aims before an I.P.O. would be unthinkable for most corporate leaders, who tend to focus on their core businesses and try to project steadiness to potential investors. Mr. Musk's new goals for SpaceX raise questions about how much shareholders can rely on his word, corporate governance experts said. Yet the billionaire has an uncanny ability to bring investors along for the ride, they said. "In most other corporations where the C.E.O. makes promises that do not prove out, investors tend to react in an adverse way, and they usually do not last long," said Brian Quinn, a law professor at Boston College. But with Mr. Musk, he said, "people believe him or want to believe him." In online posts, Mr. Musk has acknowledged SpaceX's "priority shift." But he has said the new goals do not take away from the Mars plan and are steppingstones to making humans a multiplanetary species. "The capabilities we unlock by making space-based data centers a reality will fund and enable self-growing bases on the moon, an entire civilization on Mars and ultimately expansion to the universe," Mr. Musk wrote in a February letter to SpaceX employees. Mr. Musk has a history of making bold predictions that do not materialize. But while his timelines can be imprecise, his long-term visions have delivered huge opportunities, his supporters said. "Elon is always directionally correct," said Peter Diamandis, a SpaceX investor and the founder of the XPrize Foundation, a nonprofit that supports technological development. "His time frames may be off, but he'll eventually get there." Mr. Musk and a SpaceX spokesman did not respond to requests for comment. Over the years, Mr. Musk has acknowledged his lack of business plans and his reliance on gut instinct. Eight former SpaceX executives and employees, speaking on the condition of anonymity because they feared retribution, told The New York Times that during their times at the company, they had become accustomed to Mr. Musk's whipsaw directives and his use of social media to make announcements or product changes. In 2014, Mr. Musk announced on Twitter, now known as X, that SpaceX would hold an event to unveil the second version of its Dragon capsule, a spacecraft meant to ferry passengers and cargo from orbit, two former employees said. The vehicle was not near completion, so his team scrambled to pull together a full design and event, the former employees said. "We want to take a big step in technology and really create something that was a step change in spacecraft technology," Mr. Musk said at the event, where he unveiled a vehicle that could land anywhere on Earth using jet propulsion. (SpaceX later scrapped the idea in favor of parachute-based landing after Mr. Musk determined that Dragon's jet propulsion wasn't practical, three of the people told The Times.) That same year, Mr. Musk became interested in satellite-based internet and began meeting with Greg Wyler, the founder of OneWeb, a satellite start-up, said two people familiar with the discussions, who requested anonymity out of fear of retribution. The relationship never came to fruition, and Mr. Musk set out on his own, opening a SpaceX engineering office in Redmond, Wash., in 2015 to develop internet satellites. The resulting service, Starlink, underwent layoffs as SpaceX invested in research and development. But the bet paid off: Starlink now has 10 million subscribers and generated $8 billion in sales in 2024, according to documents obtained by The Times. Now Mr. Musk appears to be trying to replicate the Starlink playbook, but with data centers in space. SpaceX had not previously focused on A.I., much less on orbital data centers, three of the former SpaceX executives said. But after Google and others began discussing orbital data centers last year, Mr. Musk declared in October that "SpaceX will be doing this." In January, SpaceX filed paperwork with the Federal Communications Commission to potentially launch one million satellites for an "orbital data center system." A week later, it announced a merger with xAI, Mr. Musk's A.I. start-up. "In 36 months, but probably closer to 30 months, the most economically compelling place to put A.I. will be in space," Mr. Musk said in a recent podcast appearance. This year, more than 20 engineers and researchers have left xAI, whose products have lagged behind those of OpenAI, Anthropic and Google in use. Mr. Musk appears eager to push SpaceX further into A.I. In the deal with Cursor announced Tuesday, SpaceX said the combination with the young A.I. company, which makes code-writing software, would "allow us to build the world's most useful" A.I. models. Another new goal is the moon. While two of the former SpaceX executives said Mr. Musk had previously dismissed landing on the moon because it was not a new achievement, he said in February that the company had "shifted focus to building a self-growing city on the moon." With the success of NASA's recent Artemis II mission and the agency's commitment to further moon exploration, Mr. Musk may see an immediate financial opportunity, the former SpaceX executives said. SpaceX will "strive to build a Mars city and begin doing so in about 5 to 7 years, but the overriding priority is securing the future of civilization and the moon is faster," Mr. Musk posted on Feb. 8. That month, he also spoke to some SpaceX employees about building lunar A.I. satellite factories and launching those satellites into orbit using a space catapult, according to a recording of the employee meeting obtained by The Times. Ryan Mac is a Times reporter who covers corporate accountability across the global technology industry.
