News & Updates
The latest news and updates from companies in the WLTH portfolio.
Exclusive | Silicon Valley banker wants to swap his $8M estate for Anthropic stock: 'Not your typical deal'
A Silicon Valley dealmaker is looking to swap his 14-acre estate for shares of the AI giant Anthropic - the latest sign of the lengths investors will go to get a piece of the red-hot artificial intelligence sector, The Post has learned. Storm Duncan, founder of the tech-focused investment bank Ignatious, detailed in an exclusive interview how the unusual transaction could work - including Duncan covering all the closing costs for any Anthropic equity holder who takes him up on the offer. Duncan said real-estate advisors have pegged the value of his spread - a 4,400-square-foot, four-bedroom mansion perched in the hills of Mill Valley just north of the Golden Gate Bridge - at roughly $8 million. The lavish compound and its infinity pool boast sweeping views of San Francisco Bay and nearby Mount Tamalpais. "There's definitely been some interest in it," Duncan told The Post. "But it's not your typical deal." The 56-year-old financier's listing - which sparked chatter in tech circles after he posted it on Linkedin this week - shows that even well-connected tech insiders are scrambling to snag shares of privately held AI startups before they go public. Some social media users flagged Duncan's pitch as the latest indicator the AI market is overheating. "Peak late-cycle vibes. Please frame this and put it on the wall," one X user commented. Duncan said he already has Anthropic shares which he acquired in a 2024 funding round. He has gotten to know Anthropic's management team through prior work and said he believes they will succeed in the AI race long term. "It'll be a conversation," Duncan said. "But I don't take my offer as a sign that the market has reached the top." San Francisco-based Anthropic has lately fielded offers from venture capitalists that would value it at $800 billion or more, according to a Bloomberg report this month. Duncan said his own analysis of Anthropic's financials yields a similar valuation. Acquiring shares directly from another individual versus buying them through so-called secondary markets offers more transparency, Duncan said, adding that determining the legitimacy of still-private startup shares can be difficult and risky. There are also tax advantages to doing a swap for his house, he added. A swap would have to occur at a current valuation for Anthropic and he would allow any takers to retain 20% of the upside value of the exchanged shares before they could be legally transferred to him, Duncan said. Duncan has invested in numerous AI startups including Gecko Robotics which has been valued at $1.25 billion; he has also backed enterprise networking company Upscale AI, which inked a $200 million Series A financing in January. The Mill Valley estate at 114 Inez Place has long served as a party destination. In the years before Covid, Duncan said he threw shindigs for as many as 700 guests, with models, real-estate bigwigs and influential technologists in attendance. "Suffice to say that if the house blew up during a party, it would have been a problem for the tech world," Duncan said. "My goal with those parties was to get people together from different walks of life." Duncan said he now splits his time between homes in Miami and Wyoming and that a "very famous" venture capitalist currently lives at his Mill Valley estate, declining to name the renter.
Peak hour breakdown on major Perth freeway causing chaos
A breakdown on Mitchell Freeway northbound in Osborne Park is causing traffic headaches for motorists heading home on Thursday evening. The breakdown occurred on Mitchell Freeway just after Hutton Street and was reported by Main Roads at 5.43pm. The right lane has been closed by Incident Response, and traffic is heavy leading up to the site. Main Roads has warned motorists to "exercise extreme caution" if driving in the area.
SpaceX's S-1 excerpts list "manufacturing our own GPUs" among the "substantial capital expenditures" it is undertaking, with the size of the expenditure TBD
New York Times: SpaceX says it's working with Cursor to build "the world's most useful models" and it has the right to acquire Cursor for $60B or pay $10B for the partnership SpaceXAI and @cursor_ai are now working closely together to create the world's best coding and knowledge work AI. The combination of Cursor's leading product and distribution to expert software engineers with SpaceX's million H100 equivalent Colossus training supercomputer will allow us to build the world's most useful models. Cursor has also given SpaceX the right to acquire Cursor later this year for $60 billion or pay $10 billion for our work together.
Pakistani co-founder AI startup Cursor eyes $60bn SpaceX deal
AI coding startup Cursor, co-founded by Pakistani entrepreneur Sualeh Asif, is set for a potential $60 billion acquisition deal with SpaceX, marking a major milestone for Pakistan's growing presence in the global tech industry. According to multiple reports, SpaceX has secured the right to acquire Cursor later this year. If the deal does not go through, the aerospace company will reportedly pay $10 billion for collaborative work between the two firms. The partnership is expected to combine Cursor's AI-driven coding tools with SpaceX's advanced computing infrastructure. Cursor has emerged as a leading player in AI-powered software development, offering tools that automate coding tasks and improve developer productivity. The company has gained traction alongside major AI firms such as OpenAI and Anthropic, reflecting rapid growth in demand for generative AI solutions. Originally from Karachi, Sualeh Asif studied at the Massachusetts Institute of Technology, where he co-founded the startup with fellow students. His journey from local education to global recognition has been widely praised as an example of innovation and entrepreneurship. The company reportedly generates over $1 billion in annualised revenue and serves millions of developers across thousands of enterprises, including major global firms. Its rapid rise has positioned it among the fastest-growing startups in the AI sector. Former IT minister Umar Saif lauded Asif as a role model for young Pakistanis, highlighting his achievements as a self-made entrepreneur. If finalised, the deal would represent one of the largest acquisitions in the AI industry and a significant moment for Pakistani representation in global technology leadership.
Elon Musk's Tesla Invested $2 Billion Into SpaceX, Q1 2026 Earnings Show - Tesla (NASDAQ:TSLA)
$2 Billion Investment In its quarterly earnings report released on Wednesday, the EV giant shared its investment in SpaceX. The EV giant's report showed a $2 billion charge due to its "SpaceX equity investment." The company also expanded on why its collaboration with the company mattered. "We are expanding our scope of manufacturing to include semiconductor fabrication, an important step to ensure sufficient and resilient chip supply," Tesla said, adding that the SpaceX partnership would "build the largest chip fab ever: vertically integrating logic, memory and advanced packaging." The EV maker also said that it expects "greater chip demand," which would exceed existing industry capacity. Tesla's Earnings CallElon Musk's Optimus Update According to Benzinga Edge Rankings, Tesla offers satisfactory Momentum, but poor Value. It provides a favorable price trend in the Long term. Price Action: TSLA declined 2.17% to $377.95 during pre-market trading on Thursday. Check out more of Benzinga's Future Of Mobility coverage by following this link. Photo courtesy: Frederic Legrand - COMEO from Shutterstock Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.
Vercel Finds More Compromised Accounts in Context.ai-Linked Breach - IT Security News
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user. The technical storage or access that is used exclusively for statistical purposes. The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
SpaceX Secures $60B Option to Acquire Cursor Amid $2B Funding Round
Cursor gains a $10B safety net while SpaceX gains AI credibility. The story behind the $60B option deal. Just days before it was expected to close a $2 billion funding round at a $50 billion valuation, Cursor found itself in a very different kind of negotiation. The fast-growing AI coding platform, known for building its own AI-native development environment for programmers, was reportedly finalising commitments from investors including Andreessen Horowitz, Nvidia, Thrive Capital, and Battery Ventures. At the same time, behind the scenes, Cursor was in talks with SpaceX. Now those talks have come into the open, and the result is not an acquisition, at least not yet. SpaceX, the Elon Musk-founded company that makes Starlink, announced that it has secured the option to acquire Cursor later this year for $60 billion. If it decides not to proceed with the purchase, SpaceX will instead pay Cursor $10 billion over time for what it calls "AI collaboration". For Cursor, this arrangement quietly solves a looming problem. Despite explosive revenue growth and rapid adoption among enterprises, AI coding is incredibly compute-intensive. According to reports, the planned $2 billion raise would not have been enough to carry the company to breakeven, meaning another large funding round would likely follow. This deal effectively gives Cursor a $10 billion financial backstop and potential access to SpaceX's vast data centre resources, easing the pressure to continuously raise capital in an increasingly competitive AI landscape dominated by tools like Codex from OpenAI and Claude Code from Anthropic. For SpaceX, the timing is just as important. The company, which recently merged with xAI, is preparing for a major IPO. Announcing an option to buy one of the most valuable AI startups in the world allows SpaceX to position itself as more than a space and satellite business. It signals to future public investors that AI will be a core part of its story, without forcing the company to update its confidential financial filings before going public. If the acquisition happens after the IPO, SpaceX could even use publicly traded stock to help finance the purchase.
Is Anthropic's Mythos AI a Threat to Ethereum and Solana?
Ethereum and Solana are on the front lines of those problems, and they're getting battered. The crypto sector just got even spicier. On April 7, Anthropic announced Claude Mythos, a new artificial intelligence (AI) model that the company (perhaps self-servingly) has called too dangerous for public release. Since the start of April, about $606 million in capital has been pilfered from decentralized finance (DeFi) protocols across 12 different hacks targeting Ethereum (CRYPTO: ETH), Solana (CRYPTO: SOL), a couple of smaller networks, and especially the cross-chain plumbing between them. Don't worry -- there's no causal connection between Mythos' launch and the heists, as Mythos isn't publicly available yet. But the crypto community is alight with fear anyway, as the hacking tools of today are clearly sufficient to bring plenty of ruin to DeFi ecosystems, and the tools of the (very) near future look to be dramatically more powerful. So, is there a real threat to DeFi majors like Ethereum and Solana, or, in other words, is the crypto community's fear of Mythos justified? 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. If the experts and its creators are to be believed, Anthropic's Mythos is indeed a real leap forward in both offensive and defensive cyber capabilities, as it's supposedly able to autonomously find and exploit previously unknown software vulnerabilities at scale. Access to the model is currently limited to a few dozen owners of critical cyber infrastructure, and defensive work is being prioritized before its eventual release to the public, assuming that ever happens. But crypto's hacks in April weren't a result of the type of vulnerabilities that a Mythos-style model is supposed to excel at finding. Solana's $285 million Drift Protocol hack on April 1, allegedly performed by a group of hackers from North Korea, was implemented via a combination of normally allowed protocol interactions and social engineering, which is to say that the attackers were able to sweet-talk or otherwise deceive certain personnel into taking actions that compromised security when paired with other usually innocuous behaviors. Similarly, the $292 million Kelp DAO exploit on April 18 on Ethereum, also attributed to North Korean hackers, happened due to a cross-chain bridge being misconfigured in a way that left it vulnerable to a fairly rudimentary intrusion rather than an exotic cyberattack or advanced malware. Thus, the anxiety about cybersecurity in crypto, while timely and well-grounded in reality, isn't even directly connected to the AI tools that are widely available today. Nonetheless, at the start of April, the entire DeFi sector had $94 billion in total value locked (TVL) in DeFi protocols. As of April 21, there was $85 billion in TVL. When the Kelp hack happened, roughly $13 billion in DeFi TVL vaporized within 48 hours, with a $6.6 billion drop concentrated at lending protocol Aave. It's these cybersecurity incidents and the fears that follow them that are actually driving capital out of the sector. And that's certainly bad news for both Ethereum and Solana, as both rely heavily on the idea that their DeFi segments will attract capital to their chains and thereby boost the prices of the coins. One important thing to note here, for the sake of understanding the risk to your crypto portfolio, is that not all cryptocurrencies are as exposed to present and emerging cybersecurity risks as Solana and Ethereum. Bitcoin's programmable surface is, by design, very minimal; it doesn't have native smart contracts or any DeFi ecosystem. It exposes less to potential attackers, no matter what tools attackers are equipped with. Ethereum, Solana, and other smart contract-capable networks made the opposite bargain, and they can't go back on it. Ethereum's DeFi TVL alone is $45.8 billion, meaning that tens of billions in capital are tangled up in structures of user-written logic and human-created computer code. Solana's TVL is only $5.5 billion, but the problem is the same. One key takeaway for investors is that if defensive cybersecurity on those networks doesn't improve significantly in advance of the public getting access to Mythos or similarly powerful AI tools, few investors will be interested in committing their capital to places where they perceive it will be easy to steal. So, is Mythos a threat to Ethereum and Solana? Yes, it could be, but in practical terms, it probably won't be a problem for a while, and there are many more urgent risks to address in the meantime. If you hold these coins, be advised that there might be some choppy waters ahead. Before you buy stock in Ethereum, 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 Ethereum 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 $499,277!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $1,225,371!* Now, it's worth noting Stock Advisor's total average return is 972% -- a market-crushing outperformance compared to 198% 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. Alex Carchidi has positions in Bitcoin, Ethereum, and Solana. The Motley Fool has positions in and recommends Aave, Bitcoin, Ethereum, and Solana. The Motley Fool has a disclosure policy.
France Probes Weather Data Tampering After Surge in Polymarket Bets
(Bloomberg) -- France's forecasting office flagged suspected tampering with weather sensors at the country's largest airport and referred the case to police, after detecting unusual readings alongside heavy betting on a popular prediction market. Most Read from Bloomberg Automated temperature readings taken at Meteo France's weather station at Charles de Gaulle International Airport spiked 4C and 5C unexpectedly in the evenings of April 6 and April 15, respectively, reaching the highest temperature recorded at the site on those days, data from the installation show. Readings from the site are important for the safe operation of the airport. They are also used to settle contracts for daily high temperatures on Polymarket, according to information on the website where traders place bets on real-world outcomes. Weather betting has boomed on prediction markets like Polymarket and Kalshi Inc., where individual traders and weather experts have flocked to put money behind their predictions for temperatures, or the amount of rain or snowfall in particular areas on specific dates. Boosters say prediction markets create economic and social value by providing better information about what will happen in the world. Detractors call them glorified gambling vulnerable to manipulation, insider trading and other shenanigans. Concerns have emerged for other contracts on prediction markets, including claims from an Israeli journalist that Polymarket users pressured him to change a story about missile strikes outside Jerusalem. Prediction market traders and independent meteorologists in a French weather discussion forum flagged the data irregularities and questioned the results of the contracts, which attracted roughly $1.4 million in combined bets, according to Polymarket data. Total betting for each was more than double the typical volume for other daily Paris temperature contracts in April. A spokesperson for Meteo France, Laurent Becler, said technicians examined sensor data and inspected the weather station, and the forecasting office subsequently filed a complaint for tampering with the operation of an automated data processing system to airport police. Becler declined to answer additional questions about the possible tampering and data irregularities. Airport police declined to answer questions about the complaint and referred questions to court officials, who declined to answer a request for more information. Charles de Gaulle International Airport declined to comment.
Anthropic looks to hire six-figure role for negotiating data center deals to fuel Europe AI expansion
Anthropic is ramping up a push to secure European data center deals to power its AI models, as it looks to hire a role for negotiating compute capacity in the region. U.S. hyperscalers' AI infrastructure expenditure is set to top $600 billion in 2026. Anthropic is looking to capitalize on the boom and has announced a slew of data center deals in the U.S. in recent weeks. While it's yet to unveil any in Europe, that could be about to change. Anthropic is now recruiting for a principal to "drive the commercial sourcing and transaction execution process" for its European data center capacity deals, according to a job advert posted in London. Anthropic declined to comment on the job advert or its plans for data centers in Europe. It comes on the back of a number of AI infrastructure deals for the company. Anthropic said this week that it's committed to spending more than $100 billion on Amazon Web Services tech over the next 10 years. It also signed an expanded deal with Broadcom earlier this month for about 3.5 gigawatts worth of computing capacity. Anthropic is currently evaluating deals to acquire data center capacity directly from developers "across the world," a source familiar with discussions told CNBC.
Polymarket traders win $37K after Paris weather data glitch, raising suspicion1h
Ruben Hallali, a meteorologist, told French media outlet BFMTV the temperature glitch was unlikely to be a natural event. Two Polymarket accounts have attracted suspicion after making $37,000 betting correctly on two unusual temperature readings of a weather station located in a major airport in France. The two weather-focused prediction markets focused on the highest temperature in Paris on April 6 and 15, using the highest temperature recorded at the Charles de Gaulle Airport Station in degrees Celsius, according to Polymarket. French media outlet BFMTV reported on Monday that the temperature suddenly climbed to over 21 degrees Celsius on April 6, before dropping again immediately. The market resolved with the winner taking over $16,000. Meanwhile, blockchain analytics tool Bubblemaps reported a similar glitch for the April 15 market. The weather station showed 18 degrees Celsius most of the day, then suddenly spiked to 22 degrees Celsius before dropping back. Some have questioned whether foul play was involved. Prediction markets are already facing growing scrutiny over insider trading and possible violations of gambling laws. "That spike didn't show on nearby stations," Bubblemaps analysts said, adding that "Just before the spike, one trader started buying NO shares on "18°C," before exiting with over $21, 000. Ruben Hallali, a meteorologist, told BFMTV the temperature glitch was unlikely to be a natural event. Related: Charles Schwab, Citadel Securities are eying prediction markets "Such temperature variations seem very unlikely, especially on these two dates, and over such a short period. We can imagine that an individual with a good understanding of how the sensors work intervened, resulting in temperatures rising by two degrees at the right time, to validate a bet," he added. Météo France, the official government weather agency of France, has reportedly made a complaint with the police unit, the Roissy Air Transport Gendarmerie Brigade, for alleged tampering with the operation of its automated data processing systems.
ESET Research discovers new China-aligned group, GopherWhisper: It abuses messaging services Discord, Slack, and Outlook to spy
* ESET Research has uncovered a new China-aligned APT group, which has been named GopherWhisper, that targets governmental institutions in Mongolia. * GopherWhisper leverages Discord, Slack, Microsoft 365 Outlook, and file.io for command and control (C&C) communications and exfiltration. * The group's toolset includes custom Go-based backdoors, injectors, exfiltration tools, loader FriendDelivery, and a C++ backdoor. * ESET analyzed C&C traffic from the attacker's Slack and Discord channels, gaining information about the group's internal operations and post-compromise activities. BRATISLAVA, Slovakia, April 23, 2026 (GLOBE NEWSWIRE) -- ESET researchers have discovered a previously undocumented China-aligned APT group that they named GopherWhisper. The group wields a wide array of tools, mostly written in Go, that use injectors and loaders to deploy and execute various backdoors in its arsenal. In the observed campaign, the threat actors targeted a governmental institution in Mongolia. GopherWhisper abuses legitimate services, notably Discord, Slack, Microsoft 365 Outlook, and file.io, for command and control (C&C) communication and exfiltration. ESET discovered the group in January 2025, when it found a previously undocumented backdoor, which ESET researchers named LaxGopher, in the system of a government institution in Mongolia. Digging deeper, they managed to uncover several more malicious tools, mainly various additional backdoors, all deployed by the same group. The majority of these tools were written in Go, and their collective aim was cyberespionage. According to ESET telemetry, the victim impacted by GopherWhisper backdoors is a Mongolian governmental institution. By analyzing the C&C traffic from the attacker-operated Discord and Slack servers, ESET estimates that dozens of other victims besides the Mongolian institution were also affected, though it has no information about their geolocation or verticals. Of the seven tools that were discovered, four are backdoors - LaxGopher, RatGopher, and BoxOfFriends, written in Go, and SSLORDoor, written in C++. Furthermore, ESET found an injector (JabGopher), a Go-based exfiltration tool (CompactGopher), and a malicious DLL file (FriendDelivery). Since the set of malware ESET found bore no code similarities to any known threat actor's tools, and there was also no overlap in the Tactics, Techniques, and Procedures (TTPs) used by any other group, ESET decided to attribute the tools to a new group. Researchers chose to name that group GopherWhisper due to the majority of the group's tools' being written in the Go programming language, which has a gopher as its mascot, and based on the filename of whisper.dll, which is side-loaded. Get the latest news delivered to your inbox Sign up for The Manila Times newsletters By signing up with an email address, I acknowledge that I have read and agree to the Terms of Service and Privacy Policy. GopherWhisper is characterized by the extensive use of legitimate services such as Slack, Discord, and Outlook for C&C communication. "During our investigation, we managed to extract thousands of Slack and Discord messages, as well as several draft email messages from Microsoft Outlook. This gave us great insight into the inner workings of the group," says ESET researcher Eric Howard, who discovered the new threat group. "Timestamp inspection of the Slack and Discord messages showed us that the bulk of them were being sent during working hours, i.e. between 8 a.m. and 5 p.m., which aligns with China Standard Time. Furthermore, the locale for the configured user in Slack metadata was also set to this time zone. We therefore believe that GopherWhisper is a China-aligned group," explains Howard. Advertisement Based on this ESET investigation, the group's Slack and Discord servers were first used to test the functionality of the backdoors, and then later, without clearing the logs, also used as C&C servers for the LaxGopher and RatGopher backdoors on multiple compromised machines. In addition to the Slack and Discord communications, ESET researchers were also able to extract email messages used for communication between the BoxOfFriends backdoor and its C&C using the Microsoft Graph API. ESET Research's Eric Howard presented these findings at Botconf 2026 conference. For a more detailed analysis of the new GopherWhisper threat group and its arsenal, check out the latest ESET Research blogpost and white paper "GopherWhisper: A burrow full of malware" on WeLiveSecurity.com. Make sure to follow ESET Research on Twitter (today known as X), BlueSky, and Mastodon for the latest news from ESET Research. About ESET Advertisement ESET® provides cutting-edge cybersecurity to prevent attacks before they happen. By combining the power of AI and human expertise, ESET stays ahead of emerging global cyberthreats, both known and unknown - securing businesses, critical infrastructure, and individuals. Whether it's endpoint, cloud, or mobile protection, our AI-native, cloud-first solutions and services remain highly effective and easy to use. ESET technology includes robust detection and response, ultra-secure encryption, and multifactor authentication. With 24/7 real-time defense and strong local support, we keep users safe and businesses running without interruption. The ever-evolving digital landscape demands a progressive approach to security: ESET is committed to world-class research and powerful threat intelligence, backed by R&D centers and a strong global partner network. For more information, visit www.eset.com, or follow our social media, podcasts, and blogs.
Anthropic CEO Dario Amodei has a warning for AI companies: Stop telling people that...
Dario Amodei sat down for lunch at a San Francisco Italian restaurant in April, ate three of the four shared bread rolls, and told the Financial Times something that most of his peers in the AI industry are still tiptoeing around: the disruption is real, it's coming fast, and pretending otherwise is actively eroding public trust in the technology. He wasn't there to sugarcoat things. He was there to say that the AI industry has a credibility problem -- and that the only way out of it is to stop promising a soft landing while quietly accelerating the turbulence. For an industry used to talking about abundance and breakthrough, it was a notably uncomfortable thing to say out loud.Amodei, who leads Anthropic -- currently riding a $380 billion valuation and a wave of buzz around its powerful Claude Mythos model -- said he believes AI could eliminate around 50% of entry-level white-collar jobs within five years. That includes junior tech roles, early-career lawyers, consultants, and finance professionals. His message to the broader industry wasn't panic, but accountability: stop overselling the upside while glossing over what's being lost."We should not deny that the disruption is going to happen," Amodei told the FT. "We just have to make the positive effect so large that we have a tool to address the disruption."His candor stands out in an industry that has historically leaned on optimism as a default. OpenAI CEO Sam Altman, meanwhile, released a 13-page policy document in early April calling for a kind of New Deal for the AI era -- proposing four-day workweeks, public wealth funds, and portable benefits to cushion the transition to what he calls "superintelligence." Critics were quick to note the irony: the companies creating the displacement are now pitching themselves as the architects of the safety net. One analyst called it "comms work to provide cover for regulatory nihilism."Amodei's concern isn't just theoretical. Anthropic's own labor market research, published in March, found that computer programmers, customer service representatives, and financial analysts are among the most exposed occupations to AI automation right now. A separate survey of 81,000 Claude users, released in April, found that early-career workers expressed significantly higher anxiety about job displacement than senior professionals -- consistent with what's already showing up in hiring data. Among workers aged 22 to 25, the rate of entry into high-exposure occupations has dropped by around 14% compared to pre-ChatGPT levels.Federal Reserve researchers, using an expanded definition of software workers that includes contractors and coders outside the tech sector, found roughly half a million fewer coding jobs exist today than pre-AI trends would have predicted. The job market isn't in freefall, but the floor is quietly shifting underneath it -- and it's the people just starting out who are feeling it first.Amodei's framing keeps returning to a single phrase: AI can only "diffuse at the speed of trust." And right now, trust is running short.Part of the problem, he argues, is that AI companies -- his own included -- haven't fully delivered the promised benefits yet. The productivity gains are real in pockets, particularly in software development and certain knowledge work, but they haven't translated into broadly visible improvements in living standards. A recent survey of over 1,000 executives found that 80% said AI wasn't having any measurable impact on their productivity or headcount. Until the benefits show up in ways ordinary people can feel, he says, skepticism is not only understandable but rational."Is that just propaganda? Is that just vapourware that's not going to happen? We actually have to make it happen," he said. That's a more humble posture than most AI CEOs tend to adopt publicly.It also sets up an uncomfortable tension. Anthropic this year unveiled Claude Mythos -- a model that cybersecurity researchers say represents a genuine step change in automated vulnerability discovery, capable of finding and exploiting zero-day flaws in major operating systems without human guidance. The company's AI coding tools have helped trigger a wave of stock selloffs in the software sector; nearly $3 trillion in market cap has evaporated since October, as investors worry AI agents are quietly eating the application layer. Whether that counts as "delivering the benefits" depends a lot on who you ask and whose job is next.Not everyone is convinced the industry's newfound policy enthusiasm is the real thing. Soribel Feliz, a former U.S. Senate AI policy advisor, noted that the ideas circulating -- shared prosperity, democratized access, worker protections -- are essentially the same framework discussed since ChatGPT launched in late 2022. "I have it in my handwritten notes," she told Fortune. "All of this was already said, all of it." The problem, she said, isn't the vocabulary. It's that nobody has built the actual mechanisms to back it up.Economists who have spent careers studying technological transitions are also skeptical of the timeline predictions coming from AI executives. Daron Acemoglu, Erik Brynjolfsson, and David Autor -- some of the most cited labor economists working on this question -- have pushed back on sweeping job-loss projections, arguing that the evidence so far points to slower, more uneven disruption than the industry tends to advertise.Amodei seems at least partially aware of this gap. He wants AI regulation modeled on how society handles cars and planes -- acknowledging economic value while enforcing real safety standards. He's put $20 million behind a PAC pushing for stricter AI safety laws and has clashed with his own government over attempts to restrict Anthropic's Pentagon contracts.Whether any of that adds up to meaningful protection for the entry-level analyst staring down an uncertain career path is still an open question. But at minimum, one of the most powerful people in AI is saying out loud what others keep leaving in the footnotes: the disruption isn't something that happens to other industries. It's already here.
ESET Research discovers new China-aligned group, GopherWhisper: It abuses messaging services Discord, Slack, and Outlook to spy
* ESET Research has uncovered a new China-aligned APT group, which has been named GopherWhisper, that targets governmental institutions in Mongolia. * GopherWhisper leverages Discord, Slack, Microsoft 365 Outlook, and file.io for command and control (C&C) communications and exfiltration. * The group's toolset includes custom Go-based backdoors, injectors, exfiltration tools, loader FriendDelivery, and a C++ backdoor. * ESET analyzed C&C traffic from the attacker's Slack and Discord channels, gaining information about the group's internal operations and post-compromise activities. BRATISLAVA, Slovakia, April 23, 2026 (GLOBE NEWSWIRE) -- ESET researchers have discovered a previously undocumented China-aligned APT group that they named GopherWhisper. The group wields a wide array of tools, mostly written in Go, that use injectors and loaders to deploy and execute various backdoors in its arsenal. In the observed campaign, the threat actors targeted a governmental institution in Mongolia. GopherWhisper abuses legitimate services, notably Discord, Slack, Microsoft 365 Outlook, and file.io, for command and control (C&C) communication and exfiltration. ESET discovered the group in January 2025, when it found a previously undocumented backdoor, which ESET researchers named LaxGopher, in the system of a government institution in Mongolia. Digging deeper, they managed to uncover several more malicious tools, mainly various additional backdoors, all deployed by the same group. The majority of these tools were written in Go, and their collective aim was cyberespionage. According to ESET telemetry, the victim impacted by GopherWhisper backdoors is a Mongolian governmental institution. By analyzing the C&C traffic from the attacker-operated Discord and Slack servers, ESET estimates that dozens of other victims besides the Mongolian institution were also affected, though it has no information about their geolocation or verticals. Of the seven tools that were discovered, four are backdoors -- LaxGopher, RatGopher, and BoxOfFriends, written in Go, and SSLORDoor, written in C++. Furthermore, ESET found an injector (JabGopher), a Go-based exfiltration tool (CompactGopher), and a malicious DLL file (FriendDelivery). Since the set of malware ESET found bore no code similarities to any known threat actor's tools, and there was also no overlap in the Tactics, Techniques, and Procedures (TTPs) used by any other group, ESET decided to attribute the tools to a new group. Researchers chose to name that group GopherWhisper due to the majority of the group's tools' being written in the Go programming language, which has a gopher as its mascot, and based on the filename of whisper.dll, which is side-loaded. GopherWhisper is characterized by the extensive use of legitimate services such as Slack, Discord, and Outlook for C&C communication. "During our investigation, we managed to extract thousands of Slack and Discord messages, as well as several draft email messages from Microsoft Outlook. This gave us great insight into the inner workings of the group," says ESET researcher Eric Howard, who discovered the new threat group. "Timestamp inspection of the Slack and Discord messages showed us that the bulk of them were being sent during working hours, i.e. between 8 a.m. and 5 p.m., which aligns with China Standard Time. Furthermore, the locale for the configured user in Slack metadata was also set to this time zone. We therefore believe that GopherWhisper is a China-aligned group," explains Howard. Based on this ESET investigation, the group's Slack and Discord servers were first used to test the functionality of the backdoors, and then later, without clearing the logs, also used as C&C servers for the LaxGopher and RatGopher backdoors on multiple compromised machines. In addition to the Slack and Discord communications, ESET researchers were also able to extract email messages used for communication between the BoxOfFriends backdoor and its C&C using the Microsoft Graph API. ESET Research's Eric Howard presented these findings at Botconf 2026 conference. For a more detailed analysis of the new GopherWhisper threat group and its arsenal, check out the latest ESET Research blogpost and white paper "GopherWhisper: A burrow full of malware" on WeLiveSecurity.com. Make sure to follow ESET Research on Twitter (today known as X), BlueSky, and Mastodon for the latest news from ESET Research. About ESET ESET® provides cutting-edge cybersecurity to prevent attacks before they happen. By combining the power of AI and human expertise, ESET stays ahead of emerging global cyberthreats, both known and unknown -- securing businesses, critical infrastructure, and individuals. Whether it's endpoint, cloud, or mobile protection, our AI-native, cloud-first solutions and services remain highly effective and easy to use. ESET technology includes robust detection and response, ultra-secure encryption, and multifactor authentication. With 24/7 real-time defense and strong local support, we keep users safe and businesses running without interruption. The ever-evolving digital landscape demands a progressive approach to security: ESET is committed to world-class research and powerful threat intelligence, backed by R&D centers and a strong global partner network. For more information, visit www.eset.com, or follow our social media, podcasts, and blogs.
FTX's $200,000 Cursor stake now worth $3 billion after SpaceX deal
Key point: The missed return in $FTT highlights the stakes for creditors. A minor stake in a technology startup, once dismissed during the FTX bankruptcy proceedings, has now reached a multibillion-dollar valuation. Aerospace giant SpaceX has reportedly reached a deal to acquire AI software startup Cursor at a $60 billion valuation. This development has sent shockwaves through the crypto industry, as the initial investment into Cursor by FTX and its affiliated trading firm Alameda Research has skyrocketed in value -- highlighting a dramatic shift in the value of bankruptcy estate assets. ContentsSale at bankruptcy led to massive lossesSpaceX's AI strategy and high-stakes dealFTX creditors miss out on billionsSale at bankruptcy led to massive losses Back in April 2022, Alameda Research invested $200,000 in Anysphere, the developer behind Cursor, acquiring a 5% stake when the company was valued at only $4 million. However, after FTX's collapse at the end of 2022 and the ensuing bankruptcy, court-appointed administrators offloaded the same stake in April 2023 for just $200,000 -- the original investment amount. Within just a year, Cursor's value soared to unprecedented levels, turning this sale into a cautionary tale for FTX creditors who missed out on a potential windfall. According to the SpaceX agreement, the 5% stake in Cursor is now valued at $3 billion. This means FTX's bankruptcy estate sold an asset for a mere fraction -- about 1/15,000th -- of its current value. SpaceX's AI strategy and high-stakes deal SpaceX revealed this week that it has secured a right to acquire Cursor by the end of this year at a $60 billion valuation. If the full acquisition is not completed, the agreement includes a $10 billion breakup fee. Led by Elon Musk, the move is viewed as part of SpaceX's strategic push to challenge AI coding leaders such as OpenAI and Anthropic. Musk, whose AI firm xAI merged with SpaceX last year, has acknowledged lagging behind in this field -- a factor motivating the Cursor acquisition. The reason SpaceX is delaying its immediate acquisition of Cursor is tied to its upcoming planned IPO. The company is reportedly targeting a staggering $2 trillion valuation at market debut, and the $10 billion breakup fee has become a crucial factor in the ongoing negotiation process. FTX creditors miss out on billions Following FTX's collapse, bankruptcy administrators rapidly liquidated assets to facilitate early payouts to creditors. Since then, with the crypto market's recovery, several disposed assets have experienced explosive appreciation. The investment in Cursor -- made by Alameda Research and FTX under Sam Bankman-Fried -- stands out as a vivid example. Throughout trials and in written statements, Bankman-Fried consistently argued that the hasty asset fire sales ended up irreparably harming creditors. Bankman-Fried contends that holding on to FTX's assets during the crisis, instead of dumping them, could have left creditors whole or perhaps even profiting. From prison, he projected that the bankruptcy estate would have hit a total value of $78 billion if assets had not been prematurely sold. The rollout of Cursor's AI software tool in 2023 significantly accelerated the company's valuation. Analysts have noted Cursor's growth trajectory as exceptionally rare in the startup world. While FTX bankruptcy administrators are reimbursing creditors in principal plus interest in U.S. dollars, they have not been able to pass on the substantial upside generated by recent asset appreciation. Meanwhile, Sam Bankman-Fried's family has launched a public campaign appealing for his exoneration. In March, his parents told CNN that "since FTX creditors are ultimately being repaid, the case deserves a fresh review." The Cursor stake, both in family statements and Bankman-Fried's commentary, has become a key example of lost value in the bankruptcy process. You can follow our news on Telegram, Facebook & Coinmarketcap & XDisclaimer: The information contained in this article does not constitute investment advice. Investors should be aware that cryptocurrencies carry high volatility and therefore risk, and should conduct their own research.
GopherWhisper APT group hides command and control traffic in Slack and Discord - IT Security News
Attackers continue to lean on everyday collaboration platforms to hide command and control traffic inside normal enterprise noise. A newly identified China-aligned APT group pushes that trend further, running its operations through Slack workspaces, Discord servers, Outlook drafts, and the file.io sharing service. GopherWhisper toolset overview ESET researchers have named the group GopherWhisper and tied it to an intrusion at a Mongolian governmental entity. The name draws on two elements: most of the group's tooling ... More →
SpaceX's IPO filing says its space‑AI data centers aren't yet proven
SpaceX's pre‑IPO filing warns its space‑AI data centers and Mars plans may not be viable. NEW YORK - SpaceX has warned investors that its ambitions to build space-based artificial intelligence (AI) data centres, as well as human settlements on the Moon and Mars, depend on unproven technologies and may not be commercially viable. Citing reuters.com (22/04/26), this previously unreported warning is contained in a pre-IPO (S-1) filing as the rocket manufacturer prepares for an Initial Public Offering (IPO) that could become the largest in history. The company is targeting a listing in the coming months with a valuation of around US$1.75 trillion and fundraising of US$75 billion. The business assessment in the S-1 filing, which is required under US securities law to inform investors of potential risks while protecting the company from future legal liabilities, presents a far more cautious view than the public vision promoted by billionaire CEO Elon Musk in recent weeks. "Our initiatives to develop orbital AI computing and industrialisation in orbit, on the Moon, and interplanetary remain at an early stage, involve significant technical complexity and unproven technologies, and may not achieve commercial viability," SpaceX wrote in an excerpt from the S-1 document reviewed by Reuters (reported by journalist Echo Wang). The document also specifically notes that any future orbital AI data centres would operate "in a harsh and unpredictable space environment, exposing them to a wide and unique range of space-related risks that could cause malfunction or failure". This cautious stance on paper contrasts with Musk's remarks at the World Economic Forum in January. At that time, Musk described building AI data centres in space as "a no-brainer" and expressed confidence that it would become the cheapest place to deploy AI within the next two to three years. In February, following the announcement of a merger between SpaceX and his social media and AI venture, xAI, Musk stated that "space-based AI is clearly the only way to scale". SpaceX did not immediately respond to requests for further comment regarding this discrepancy. Beyond AI and interplanetary settlements, SpaceX also highlighted its heavy reliance on Starship, its fully reusable next-generation rocket, which has so far experienced several delays and test failures. "Any failure or delay in developing Starship at scale or in achieving its intended launch cadence, reusability, and capabilities would delay or limit our ability to execute our growth strategy," the filing stated. By comparison, Starship is designed to carry significantly larger payloads than SpaceX's current flagship rocket, Falcon 9, with the ultimate goal of dramatically reducing launch costs for Starlink satellites, space-based data centres, and missions to transport humans to the Moon. (SF/LM)
ESET Research discovers new China-aligned group, GopherWhisper: It abuses messaging services Discord, Slack, and Outlook to spy
BRATISLAVA, Slovakia, April 23, 2026 (GLOBE NEWSWIRE) -- ESET researchers have discovered a previously undocumented China-aligned APT group that they named GopherWhisper. The group wields a wide array of tools, mostly written in Go, that use injectors and loaders to deploy and execute various backdoors in its arsenal. In the observed campaign, the threat actors targeted a governmental institution in Mongolia. GopherWhisper abuses legitimate services, notably Discord, Slack, Microsoft 365 Outlook, and file.io, for command and control (C&C) communication and exfiltration. ESET discovered the group in January 2025, when it found a previously undocumented backdoor, which ESET researchers named LaxGopher, in the system of a government institution in Mongolia. Digging deeper, they managed to uncover several more malicious tools, mainly various additional backdoors, all deployed by the same group. The majority of these tools were written in Go, and their collective aim was cyberespionage. According to ESET telemetry, the victim impacted by GopherWhisper backdoors is a Mongolian governmental institution. By analyzing the C&C traffic from the attacker-operated Discord and Slack servers, ESET estimates that dozens of other victims besides the Mongolian institution were also affected, though it has no information about their geolocation or verticals. Of the seven tools that were discovered, four are backdoors -- LaxGopher, RatGopher, and BoxOfFriends, written in Go, and SSLORDoor, written in C++. Furthermore, ESET found an injector (JabGopher), a Go-based exfiltration tool (CompactGopher), and a malicious DLL file (FriendDelivery). Since the set of malware ESET found bore no code similarities to any known threat actor's tools, and there was also no overlap in the Tactics, Techniques, and Procedures (TTPs) used by any other group, ESET decided to attribute the tools to a new group. Researchers chose to name that group GopherWhisper due to the majority of the group's tools' being written in the Go programming language, which has a gopher as its mascot, and based on the filename of whisper.dll, which is side-loaded. GopherWhisper is characterized by the extensive use of legitimate services such as Slack, Discord, and Outlook for C&C communication. "During our investigation, we managed to extract thousands of Slack and Discord messages, as well as several draft email messages from Microsoft Outlook. This gave us great insight into the inner workings of the group," says ESET researcher Eric Howard, who discovered the new threat group. "Timestamp inspection of the Slack and Discord messages showed us that the bulk of them were being sent during working hours, i.e. between 8 a.m. and 5 p.m., which aligns with China Standard Time. Furthermore, the locale for the configured user in Slack metadata was also set to this time zone. We therefore believe that GopherWhisper is a China-aligned group," explains Howard. Based on this ESET investigation, the group's Slack and Discord servers were first used to test the functionality of the backdoors, and then later, without clearing the logs, also used as C&C servers for the LaxGopher and RatGopher backdoors on multiple compromised machines. In addition to the Slack and Discord communications, ESET researchers were also able to extract email messages used for communication between the BoxOfFriends backdoor and its C&C using the Microsoft Graph API. ESET Research's Eric Howard presented these findings at Botconf 2026 conference. ESET® provides cutting-edge cybersecurity to prevent attacks before they happen. By combining the power of AI and human expertise, ESET stays ahead of emerging global cyberthreats, both known and unknown -- securing businesses, critical infrastructure, and individuals. Whether it's endpoint, cloud, or mobile protection, our AI-native, cloud-first solutions and services remain highly effective and easy to use. ESET technology includes robust detection and response, ultra-secure encryption, and multifactor authentication. With 24/7 real-time defense and strong local support, we keep users safe and businesses running without interruption. The ever-evolving digital landscape demands a progressive approach to security: ESET is committed to world-class research and powerful threat intelligence, backed by R&D centers and a strong global partner network. For more information, visit www.eset.com, or follow our social media, podcasts, and blogs.
Paris Weather Glitch Sparks $37K Win on Polymarket
Two Polymarket accounts are under scrutiny after they won about $37,000 from weather prediction markets tied to Paris temperatures. The contracts used the highest temperature recorded at the Charles de Gaulle Airport station on April 6 and April 15. On both dates, the market outcome appeared to turn on short-lived jumps in temperature readings. French outlet BFMTV reported that the station moved above 21 degrees Celsius on April 6 before quickly falling again. That result helped settle one market, with the winning side taking more than $16,000. A similar pattern appeared in the April 15 market. Bubblemaps said the station held near 18 degrees Celsius for most of the day, then briefly rose to 22 degrees Celsius before dropping back. The move helped decide the market and added to concerns around the data source. Bubblemaps said one trader bought NO shares on "18°C" just before the spike and later exited with more than $21,000 in profit. The analytics firm also noted that the jump did not appear on nearby stations. That detail added more attention to the timing of the trade and the recorded reading. Ruben Hallali, a meteorologist, told BFMTV that the sudden change did not look like a normal weather event. He said the movement in the readings over such a short time was hard to explain through natural conditions alone. "Such temperature variations seem very unlikely, especially on these two dates, and over such a short period," Hallali added. "We can imagine that an individual with a good understanding of how the sensors work intervened, resulting in temperatures rising by two degrees at the right time, to validate a bet." Furthermore, Météo France, the country's official weather agency, has reportedly filed a complaint with the Roissy Air Transport Gendarmerie Brigade. The complaint concerns alleged tampering with its automated data processing systems linked to the weather station readings.
TVK Leader Aadhav Arjuna Votes in Villivakkam, Flags Transport Chaos
Despite repeated claims from U.S. leadership that strikes and operations had crippled Tehran's military strength, recent intelligence assessments suggest Iran retains significant military capabilities even after weeks of intense U.S. and Israeli bombardments. According to Pentagon and other U.S. officials familiar with the assessments, Iran still holds a substantial portion of its ballistic missile and drone arsenal, along with naval and air assets, indicating that the country's defence posture remains resilient.
