News & Updates
The latest news and updates from companies in the WLTH portfolio.
Polymarket took down wagers tied to rescue of downed Air Force officer | TechCrunch
A Democratic congressman had harsh criticism for Polymarket for allowing users to bet on the date the United States would confirm the rescue of Air Force service members shot down over Iran. In a social media post on Friday, Representative Seth Moulton wrote, "They could be your neighbor, a friend, a family member. And people are betting on whether or not they'll be saved. This is DISGUSTING." (President Donald Trump announced early Sunday that the second service member, a weapons system officer, has been rescued.) Moulton also described Polymarket as a "dystopian death market" and noted that Donald Trump Jr. is an investor. The congressman recently banned his staff from participating in prediction markets like Polymarket and Kalshi. Polymarket responded that it had taken the market down "immediately" for not meeting the company's integrity standards. "It should not have been posted, and we are investigating how this slipped through our internal safeguards," the company said. Polymarket previously saw hundreds of millions of dollars traded on contracts tied to the bombing of Iran by the United States and Israel.
Agility at Sea: Geopolitical Shocks Reshape Energy Shipping
Conflict in the Middle East has practically halted shipping traffic through the Strait of Hormuz. The narrow passageway is not technically closed, but the risks now dissuade most attempts at passing through. Tanker traffic has dropped roughly 90 percent. Around 400 vessels, many just sitting full of oil, wait anchored or sheltering in the Gulf. Maersk, MSC, Hapag-Lloyd, and CMA CGM all suspended transits. Insurers pulled war risk coverage almost immediately, and freight rates experienced massive hourly volatility. The Strait moves about a fifth of the world's oil supply on a normal day; now, however, the amount is negligible. This de facto closure of Gulf trade seriously impacts importers in Asia and Europe, with an even heavier impact on fast-growing economies like those in Africa. Africa's demand for LPG and refined products has outstripped almost every other region in terms of growth. This indicates an optimistic future for a prosperous Africa, but currently, it has seriously strained markets and shortened the timelines for commoditized purchases. Delays from rerouting, geopolitical crisis, or lack of supply can cause major inconveniences. The gap between when there is a demand signal and when delivery is expected is now quite short, and the Hormuz crisis has made that gap a first-order problem. Additionally, port infrastructure has not kept up with the rapid growth for materials and energy supply. Most of the legacy ports along West and East Africa can only dock vessels in the 10,000-15,000 DWT range. New deep-water terminals are set to open across Morocco, Senegal, and Nigeria. These are genuine upgrades that will scale trade significantly; however, they will only serve industrial corridors without replacing the distributed, higher-frequency supply that most of the continent actually depends on. Similar to value chains in parts of South and Southeast Asia, growing demand requires faster, more varied delivery, forcing smaller parcels to arrive more often. More frequent shipping is quite expensive even in normal times, but when a corridor like Hormuz tightens, it can kill markets that rely on the smaller weekly imports. A few commodities traders are handling the crisis better than others. They are the ones who built flexible vessel portfolios before this Hormuz crisis ever happened. Companies like Vitol, Trafigura, Glencore, and BGN Group have been able to maintain commercial structures that let them shift parcel sizes and route options whenever conditions might suddenly change. Petredec, for example, a specialist in LPG, designed its fleet around varied port constraints currently being faced in Africa. BW LPG, the largest VLGC fleet operator in the world, has been adding mid-sized chartering capacity alongside its core large-vessel business. And Geneva based BGN Group has gone even further by venturing with South Korea's HMM on a pair of 88,000 cubic-metre VLGCs for high-volume routes. They have also partnered with Al Seer for a further five VLGCs, while maintaining smaller tonnage for shallower ports. Trammo and other LPG specialists are building similar depth. BGN Groups' Shipping Director Ozan Turgut has said that the system was established to address an increasing imbalance between energy needs and port infrastructure across developing countries. According to Mr. Turgut: "Hybridized shipping serves two purposes. First, it allows energy to reach far more demand centers without relying on costly rail or road transfers. Second, it allows us to supply rapidly growing economies whose energy demand is rising faster than port modernization. In many of these markets, ports were simply not built for today's vessel sizes and a hybrid shipping model helps bridge that gap." Another development in hybridized logistics strategies is exemplified by NYK's recently delivered Lucent Pathfinder. The ship is a dual-fuel VLGC that can run on LPG or heavy fuel oil, with meaningfully lower emissions than conventional diesel. BGN Group chartered the ship, sustainably expanding their portfolio. Former U.S. Federal Maritime Commissioner Rob Quartel has argued, "The West's oil dependence is a strategic liability; our rivals are building the zero-emission future now." There is a third layer that has not received enough attention in the African context, namely, coastal redistribution. Battery-electric vessels are not made for long-haul voyages, but the economics of electric vessels do start to make sense for shorter coastal hops, inter-port feeder runs, the Gulf of Guinea corridor, and routes between East African ports. Research has suggested that battery electrification becomes cost-competitive on routes under roughly 1,500 km. Nonetheless, the challenge in Africa remains the lack of charging infrastructure. Port development programmes and multilateral lenders could overcome that gap faster if the sector treated coastal electrification as logistics infrastructure rather than a climate project. Conversely, digital infrastructure is streamlining physical infrastructure like ports and customs authorities. As Dr. Oluseye Akomolede, founder and CEO of Arcadia ECS, notes, "AI is about to become both a major driver of global energy demand and an essential tool for managing the infrastructure that delivers it. As shipping networks, ports, and logistics systems grow more complex, AI-driven monitoring, optimization, and security technologies will play an increasingly important role in keeping supply chains resilient." Arcadia ECS is developing deep-learning systems that can secure ports, energy facilities, and other critical infrastructure without slowing operations. The hybrid model of working large vessels for deeper water terminals, mid-sized vessels for legacy ports, and electrified feeders for short distance hopping redistribution is not new in shipping theory. But the cost of not having this capacity is growing significantly. The Hormuz crisis highlights a broader global challenge facing fast-growing energy importers. Traders and logistics companies that built portfolio flexibility in advance are keeping cargo moving in the midst of the uncertainty. For emerging market buyers, hybridization presents a tempo advantage where cargo arrival will be reliable, and supply chains can hold even with increased diversification. Yet this model goes beyond the confines of emerging markets with the Hormuz crisis demanding increased flexibility across the board. Indeed, network and logistics flexibility will prove a crucial component to traders' competitive edge where reliability is no longer guaranteed at scale.
Why is Meta pausing work with Mercor?
Meta has suspended its work with Mercor, an AI data startup, as the company investigated a supply-chain incident that exposed training secrets. The collaboration was paused while Meta assessed the scope of the exposure and whether the impacted materials could influence downstream model training. The issue matters because large-scale AI systems depend on data pipelines that are often built through multiple vendors. When a third-party data provider is compromised, the risk can spread beyond that vendor: secrets used for training or proprietary processes can become difficult to fully contain. Around the same time, other AI players were also dealing with the fallout from incidents involving AI development tooling and data sources. Separately, OpenAI said it was investigating a security incident, underscoring that multiple parts of the AI stack -- data sourcing, model development, and tooling -- are increasingly targeted. For businesses and researchers relying on external data partners, the Mercor pause is a reminder that procurement and vendor security have become part of model governance, not an afterthought. Companies typically need to verify: AI competition often pushes teams to move quickly with third-party datasets and partners. That speed can collide with the realities of cybersecurity. The Meta-Mercor pause highlights that even when the primary model is not breached, downstream training operations can still be affected -- and companies may respond by cutting access immediately and restarting only after controls are verified.
What did Anthropic change about OpenClaw?
Anthropic has changed its pricing and access policy for third-party agent tooling. Claude subscriptions will no longer cover usage on OpenClaw (and similar third-party tools), starting in early April, as Anthropic moves to better manage capacity. The practical effect is that developers who relied on running Claude via OpenClaw under a flat-rate subscription will face added cost or restricted functionality. Anthropic framed the move as a capacity-management decision, not a change in OpenClaw itself. Multiple updates describe the same direction: Anthropic will stop treating third-party tool execution as part of standard subscription coverage. The policy change is scheduled to begin on April 4 (with specific time references in different coverage), and it applies to tools like OpenClaw. For users, that means the "it's free because I pay for Claude Pro/Max" assumption no longer holds once the workload is shifted into third-party agent frameworks. OpenClaw is part of a broader ecosystem of agentic tooling that lets LLMs take actions across external workflows. When an AI provider narrows which execution contexts are included in subscription plans, it can reshape: The OpenClaw policy shift is also unfolding in a climate where Anthropic has faced controversies around Claude Code and leaked source code, plus related security concerns in the broader agent tooling landscape. Together, those developments point to a tightening phase for AI platform operators: controlling capacity, reducing risk, and clarifying what is -- and isn't -- covered by subscription terms.
Polymarket pulls controversial Iran rescue markets after intense backlash
The industry continues to expand with new players entering, though regulators are asserting greater oversight through lawsuits and licensing restrictions. Polymarket has removed a betting market tied to the rescue of U.S. service members in Iran, after intense backlash and criticism from lawmakers this weekend. The market allowed users to wager on when the U.S. would confirm the rescue of two airmen after an F-15E fighter jet was shot down over Iran. The crew members have since been rescued. Rep. Seth Moulton, a Democrat from Massachusetts, criticized the listing in a post on X, calling it "disgusting" and arguing it reduced a military rescue effort to a financial trade. Moulton has taken a hard line on prediction markets, recently banning his staff from using platforms such as Polymarket and Kalshi over concerns that financial incentives could influence policy decisions. A Polymarket spokesperson said the listing did not meet its integrity standards removed shortly after it appeared. The company added that it is reviewing how the market passed internal safeguards. The incident comes as prediction markets face rising pressure in Washington. A group of congressional Democrats last month introduced legislation that would ban contracts tied to elections, war and government actions. Separately, several senators have urged the Commodity Futures Trading Commission to prohibit markets linked to individual deaths, citing national security concerns. Regulators are also asserting authority over the sector. The CFTC said this week it filed lawsuits against three states over efforts it believes attempt to bypass federal oversight of prediction markets. Industry scrutiny has expanded beyond politics. The NFL has asked operators to avoid offering contracts it views as objectionable or open to manipulation, including bets tied to officiating decisions or events known in advance. Still, the market is expanding. Kalshi has late last month secured a license to offer margin trading to institutional investors, while new players are entering the market. Among them is JPMorgan, whose CEO, Jamie Dimon, has signaled that it is looking to enter the fray.
SpaceX's Wildest Bet Yet: Data Centers in Orbit and the $350 Billion Question
The idea sounds like something from a science fiction pitch meeting that went off the rails. Racks of servers floating in low Earth orbit, cooled by the vacuum of space, powered by unfiltered solar energy, beaming processed data back to the ground. But SpaceX isn't treating it as fiction. And the investors lining up behind the company's staggering valuation aren't either. SpaceX is now valued at roughly $350 billion following its latest tender offer, a figure that makes it the most valuable private company on the planet. That number has prompted a reasonable question from analysts and industry watchers: what exactly justifies it? The Starlink satellite internet business is growing fast, and the launch business remains dominant. But a new element has entered the calculus -- orbital data centers -- and it's generating both excitement and deep skepticism across the aerospace and technology sectors. As TechCrunch reported, SpaceX has been quietly exploring the feasibility of deploying compute infrastructure in orbit, piggybacking on the Starlink constellation's existing architecture. The concept would repurpose the ultra-cold thermal environment of space as a natural cooling mechanism for high-performance processors -- one of the most expensive operational costs for terrestrial data centers. If the physics and the economics check out, SpaceX could open an entirely new revenue stream that dwarfs its current businesses. That's a monumental "if." The global data center market is projected to exceed $500 billion annually by 2030, driven largely by the explosive demand for artificial intelligence training and inference workloads. Companies like Microsoft, Google, Amazon, and Meta are spending tens of billions each year building massive ground-based facilities, many of which face growing constraints: power availability, water for cooling, permitting delays, and community opposition. The appeal of moving some of that compute off-planet is not purely theoretical. It addresses real bottlenecks. SpaceX's advantage here is structural. No other entity on Earth can launch payloads to orbit at the cost and frequency that SpaceX can with its Falcon 9 and Starship vehicles. Starship, once fully operational for regular commercial missions, could slash per-kilogram launch costs to levels that make orbital hardware deployment economically plausible for applications beyond traditional satellites. The company already manufactures and operates thousands of Starlink satellites, giving it deep institutional knowledge of building space-grade electronics at scale. But building a satellite that routes internet traffic is a fundamentally different engineering challenge than building one that runs AI inference workloads or processes massive datasets. The power requirements alone are staggering. Current Starlink satellites generate roughly 3 kilowatts from their solar arrays. A single modern AI server rack on the ground can draw 40 to 100 kilowatts. Scaling solar arrays to meet those demands in orbit would require satellites far larger and heavier than anything in the current constellation. Thermal management cuts both ways, too. Space is cold -- but it's also a vacuum, meaning heat can only be radiated away, not convected. Designing radiator systems capable of rejecting tens of kilowatts of waste heat from densely packed processors is an unsolved engineering problem at the scale SpaceX would need. NASA and the Department of Defense have studied high-power thermal rejection in orbit for decades. Progress has been incremental, not exponential. Then there's latency. For AI training jobs that don't require real-time response, orbital processing might work. For latency-sensitive applications -- financial trading, autonomous vehicle inference, real-time content delivery -- the physics of signal propagation from low Earth orbit introduces delays that ground-based facilities simply don't have. SpaceX's Starlink network already battles latency perceptions; adding compute on top of the communication layer doesn't eliminate the problem. None of this has stopped investors from pricing in the possibility. SpaceX's $350 billion valuation, as TechCrunch noted, reflects not just what the company earns today but a broad bet on its ability to create new markets. Starlink is expected to generate over $10 billion in revenue this year. The launch business adds several billion more. But a $350 billion private valuation implies future revenue streams that don't yet exist -- and orbital compute is the most frequently cited candidate. SpaceX isn't the only company chasing this idea. Lumen Orbit, a startup backed by Y Combinator, has been developing plans for orbital data center nodes aimed specifically at AI workloads. Their pitch focuses on the same advantages: free cooling, abundant solar power, and freedom from terrestrial grid constraints. The company is early-stage, but its existence signals that serious technical minds believe the concept has legs. Meanwhile, established players are watching. Microsoft's Azure Space division has explored hybrid cloud architectures that incorporate orbital assets, though the company hasn't announced plans for space-based compute at scale. Amazon's Project Kuiper, primarily an internet constellation competitor to Starlink, could theoretically evolve in similar directions given Amazon Web Services' dominance in cloud computing. But neither company has SpaceX's launch cost advantage, and that advantage may prove decisive. The military implications are significant and often underappreciated. The U.S. Department of Defense has expressed growing interest in distributed orbital compute for intelligence processing, particularly for handling the massive data streams generated by reconnaissance satellites and sensor networks. Processing that data in orbit -- rather than downlinking raw feeds to ground stations -- could dramatically reduce bandwidth requirements and accelerate decision timelines. SpaceX already holds substantial government contracts through Starshield, its defense-oriented Starlink variant. Orbital compute would deepen that relationship considerably. Wall Street, to the extent it can analyze a private company, is divided. Some analysts argue that SpaceX's valuation is reasonable when you account for the optionality embedded in its technology platform. A company that controls the cheapest launch vehicle, the largest satellite constellation, and a potential orbital compute network has compounding advantages that are nearly impossible to replicate. Others counter that the valuation has gotten ahead of engineering reality -- that orbital data centers remain a concept, not a product, and that the capital expenditure required to make them work could consume profits from the existing businesses for years. Elon Musk, characteristically, has said little publicly about the specifics. He has referenced the potential for Starlink to evolve beyond connectivity in various forums, and SpaceX job postings have occasionally hinted at compute-related hardware development. But the company hasn't made a formal announcement, filed public patents specifically for orbital data center architectures, or disclosed R&D spending in this area. The opacity is standard for SpaceX, which controls its narrative more tightly than almost any company of its size. The energy angle deserves scrutiny. One of the strongest arguments for orbital compute is that it sidesteps the terrestrial energy crisis facing data center operators. In northern Virginia, the world's largest data center market, utilities have warned that power demand from planned facilities could exceed available grid capacity within a few years. Similar constraints are emerging in Dublin, Singapore, Amsterdam, and other major hubs. Solar power in orbit is available nearly 24 hours a day in certain orbital configurations, with no atmospheric losses and no competition for grid resources. But converting that solar energy into usable power for high-performance computing at orbital scale requires solar arrays of enormous size, advanced power conditioning systems, and battery storage for eclipse periods. The International Space Station's solar arrays produce about 120 kilowatts -- enough for a single modest server rack by today's AI standards. A commercially meaningful orbital data center would need orders of magnitude more. Starship's payload capacity makes deploying large structures possible, but the engineering integration remains formidable. So where does this leave SpaceX's valuation? The honest answer: it depends on your time horizon and your appetite for risk. If SpaceX can demonstrate even a small-scale orbital compute capability within the next three to five years -- say, processing satellite imagery or running inference models in orbit before downlinking results -- it would validate the concept enough to justify continued investment. The total addressable market, if orbital data centers become even a niche segment of the broader cloud computing industry, would be enormous. A 2% share of a $500 billion market is $10 billion in annual revenue. If the engineering proves too difficult, too expensive, or too slow, SpaceX still has a dominant launch business and a rapidly growing internet service. The company isn't betting everything on orbital compute. It's treating it as an option -- one that happens to have an extraordinary payoff if it works. That's the calculus investors are making. And right now, at $350 billion, they're paying a premium for the possibility that the wildest idea in SpaceX's portfolio might actually be the most valuable one.
'Confusion, chaos and distrust': Oregon challenges Trump's order restricting vote-by-mail
Add Yahoo as a preferred source to see more of our stories on Google. PORTLAND, Ore. (KOIN) - Oregon Attorney General Dan Rayfield filed a lawsuit on Friday challenging an executive order from President Trump that limits voting by mail. Attorney General Rayfield joined 22 other attorneys general and one governor in an effort to block Trump's March 31 executive order. The order directs the Department of Homeland Security to create verified voter lists using federal data, including Social Security. Those lists would be transferred to states, including Oregon, to determine who is eligible to vote. Rayfield argues the order weaponizes the United States Postal Service by giving it rule-making power to determine who gets a ballot through the mail and who doesn't. "The United States Postal Service has one job: to deliver the mail. President Trump is trying to give it a second one -- deciding which Americans get a ballot," said Rayfield. "That is not the postal service's role, it is not the federal government's role, and it is not constitutional," Rayfield argued in a statement. "Trump has spent years weaponizing federal agencies to prop up his false story that fraud cost him the 2020 election. He votes by mail. Oregonians vote by mail. And Oregon will keep running its own elections." The lawsuit argues that the executive order violates the separation of powers as the U.S. Constitution gives states the authority to conduct elections, not the president. The attorneys general further that the executive order weaponizes the Postal Service by directing it to withhold ballots from voters that are not on a federally-approved list. The attorneys general say the order would require states to upend their existing election procedures for upcoming elections and conduct statewide voter education efforts "at a dangerously quick pace - potentially within weeks of primary elections and mere months before the beginning of mail voting for the 2026 general election." The attorneys general warn that the executive order will "create confusion, chaos and distrust" in state elections while potentially disenfranchising eligible voters. Oregon Gov. Tina Kotek (D) issued a press release Friday in support of the lawsuit, saying, "Today, Oregon is moving to block President Trump's unconstitutional voter suppression effort," adding, "His attack on the fundamental right of every American to vote has nothing to do with election integrity and everything to do with silencing people so he can ultimately influence election results." In a statement shared with KOIN 6 News, White House spokesperson Abigail Jackson defended the order, saying, "Only Democrat politicians and operatives would be upset about lawful efforts to secure American elections and ensure only eligible American citizens are casting ballots. President Trump campaigned on securing our elections and the American people sent him back to the White House to get the job done." As reported by The Associated Press, critics say Trump's executive order would offer little time to go through voter rolls before ballots are sent out this fall for elections. Critics also question whether the administration's voter lists would be reliable. AP notes that mail voting has existed for more than a century and was increasingly popular in Democratic and Republican states until 2020, when Trump hurled baseless claims of mass voter fraud in mail-in voting. These claims come as Trump himself has voted by mail as recently as last month in a Florida special election. Oregon has had mail-in voting since 1998. The state legislative fiscal office says there have been very few cases of fraud, and not enough to sway any elections. The state already uses bar codes and signature verification for mail-in ballots, which is something the president's order also stipulates. The March executive order comes after Trump signed a similar order last year to overhaul election rules; however, the order was blocked by courts. Since then, the Trump administration has requested voter rolls from several states, including Oregon. Oregon's lawsuit was later dismissed. "Now the administration is trying again, this time using the U.S. Postal Service," the Oregon attorney general's office said. Rayfield is joined in the lawsuit by the attorneys general of Arizona, California, Colorado, Connecticut, Delaware, District of Columbia, Illinois, Maine, Maryland, Massachusetts, Michigan, Minnesota, New Jersey, Nevada, New Mexico, New York, North Carolina, Rhode Island, Vermont, Virginia, Washington, Wisconsin, and the governor of Pennsylvania. Copyright 2026 Nexstar Media, Inc. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.
Claude, OpenClaw and the new reality: AI agents are here -- and so is the chaos
The age of agentic AI is upon us -- whether we like it or not. What started with an innocent question-answer banter with ChatGPT back in 2022 has become an existential debate on job security and the rise of the machines. More recently, fears of reaching artificial general intelligence (AGI) have become more real with the advent of powerful autonomous agents like Claude Cowork and OpenClaw. Having played with these tools for some time, here is a comparison. First, we have OpenClaw (formerly known as Moltbot and Clawdbot). Surpassing 150,000 GitHub stars in days, OpenClaw is already being deployed on local machines with deep system access. This is like a robot "maid" (Irona for Richie Rich fans, for instance) that you give the keys to your house. It's supposed to clean it, and you give it the necessary autonomy to take actions and manage your belongings (files and data) as it pleases. The whole purpose is to perform the task at hand -- inbox triaging, auto-replies, content curation, travel planning, and more. Next we have Google's Antigravity, a coding agent with an IDE that accelerates the path from prompt to production. You can interactively create complete application projects and modify specific details over individual prompts. This is like having a junior developer that can not only code, but build, test, integrate, and fix issues. In the realworld, this is like hiring an electrician: They are really good at a specific job and you only need to give them access to a specific item (your electric junction box). Finally, we have the mighty Claude. The release of Anthropic's Cowork, which featured AI agents for automating legal tasks like contract review and NDA triage, caused a sharp sell-off in legal-tech and software-as-a-service (SaaS) stocks (referred to as the SaaSpocalypse). Claude has anyway been the go-to chatbot; now with Cowork, it has domain knowledge for specific industries like legal and finance. This is like hiring an accountant. They know the domain inside-out and can complete taxes and manage invoices. Users provide specific access to highly-sensitive financial details. The key to making these tools more impactful is giving them more power, but that increases the risk of misuse. Users must trust providers like Anthorpic and Google to ensure that agent prompts will not cause harm, leak data, or provide unfair (illegal) advantage to certain vendors. OpenClaw is open-source, which complicates things, as there is no central governing authority. While these technological advancements are amazing and meant for the greater good, all it takes is one or two adverse events to cause panic. Imagine the agentic electrician frying all your house circuits by connecting the wrong wire. In an agent scenario, this could be injecting incorrect code, breaking down a bigger system or adding hidden flaws that may not be immediately evident. Cowork could miss major saving opportunities when doing a user's taxes; on the flip side, it could include illegal writeoffs. Claude can do unimaginable damage when it has more control and authority. But in the middle of this chaos, there is an opportunity to really take advantage. With the right guardrails in place, agents can focus on specific actions and avoid making random, unaccounted-for decisions. Principles of responsible AI -- accountability, transparency, reproducibility, security, privacy -- are extremely important. Logging agent steps and human confirmation are absolutely critical. Also, when agents deal with so many diverse systems, it's important they speak the same language. Ontology becomes very important so that events can be tracked, monitored, and accounted for. A shared domain-specific ontology can define a "code of conduct." These ethics can help control the chaos. When tied together with a shared trust and distributed identity framework, we can build systems that enable agents to do truly useful work. When done right, an agentic ecosystem can greatly offload the human "cognitive load" and enable our workforce to perform high-value tasks. Humans will benefit when agents handle the mundane. Dattaraj Rao is innovation and R&D architect at Persistent Systems. Welcome to the VentureBeat community! Our guest posting program is where technical experts share insights and provide neutral, non-vested deep dives on AI, data infrastructure, cybersecurity and other cutting-edge technologies shaping the future of enterprise.
Do AI Models Actually Feel Things? Anthropic's Bombshell Research Says: Sort Of
A landmark new paper from Anthropic's Interpretability team reveals that Claude and likely every major LLM harbors internal "emotional" representations that causally shape behavior. This changes everything about how we build, align, and think about AI. When you type a message to an AI assistant and it responds with "I'm happy to help!" or "I'm sorry, that must be frustrating" what's actually happening inside the model? For years, the assumed answer was: nothing. It's all statistical mimicry. Pattern matching. Words without weight. Anthropic's interpretability researchers have now published evidence that this answer is, at minimum, incomplete and possibly wrong in ways that should make the entire AI industry pause. Published in early April 2026, the paper "Emotion Concepts and their Function in a Large Language Model" is a deep dive into the internals of Claude Sonnet 4.5. What the team found wasn't just surface behavior it was structural. Inside the...
Moulton rips Polymarket after platform took bets on search for the now-rescued US airman shot down over Iran - The Boston Globe
"They could be your neighbor, a friend, a family member. And people are betting on whether or not they'll be saved," he wrote. "This is DISGUSTING." Polymarket, in a response to Moulton's post two hours later, said that the now-suspended market should never have been opened on its platform. "We took this market down immediately as it does not meet our integrity standards," the company wrote. "It should not have been posted, and we are investigating how this slipped through our internal safeguards." Moulton doubled down, accusing the company of promoting bets "on the lives of our troops." He posted another screenshot showing various ongoing trades on Polymarket, including markets betting on a US ground invasion of Iran. "Your integrity standards are severely lacking," he wrote. Prediction markets such as Polymarket and Kalshi, which allow users to bet on the outcome of real-world developments, have come under criticism for their loose guardrails and consumer protections, as well as the risk of insider trading. President Trump announced on social media Sunday morning that the missing airman was rescued Saturday night. The rescue of the US airman followed an intense search after Friday's crash of the F-15E Strike Eagle, while Iran had promised a reward for anyone who turned in an "enemy pilot." Trump said that the service member was "seriously wounded and really brave" and rescued from "deep inside the mountains" in Iran. Trump said a second crew member was rescued in "broad daylight" within hours of the crash. The fighter jet was the first known American aircraft to crash in Iranian territory since the US and Israel launched the war with strikes on Iran on Feb. 28.
2 Stocks About To Make a Killing on the SpaceX IPO
SpaceX is a major customer of Nvidia, especially now that it owns xAI as well. SpaceX is getting ready to go public, and it could be the biggest offering in history. Elon Musk's space company, which recently merged with his AI company, xAI, filed confidentially to go public last week, and SpaceX is reportedly seeking a valuation of up to $2 trillion. The company was valued at $1.25 trillion in the merger earlier this year, though that number was determined by the board of directors and investment bankers. The last time investors bought shares of the company was at the end of 2025, and its valuation was $800 billion. 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 " SpaceX is reportedly seeking to raise as much as $75 billion, which would make the largest IPO raise on record. At that valuation, it would be one of the most valuable companies in the world. The IPO is clearly a windfall for its investors, including Elon Musk, whose net worth could top $1 trillion after it goes public, but there are other overlooked winners in the stock market that investors should be aware of, as these stocks could get a meaningful boost from the IPO. Keep reading to see two of them. Image source: SpaceX. Alphabet (NASDAQ: GOOG) (NASDAQ: GOOGL) is best known as the parent of Google and the world's biggest advertising company, but it's also a prolific investor. It owns stocks in more than 20 publicly traded companies, as well as a number of start-ups, including through its Google Ventures arm. Alphabet invested $900 million in SpaceX in 2015, getting a stake of about 7% in Musk's space company, and it's believed to own approximately that much today. In its Q1 2025 report, Alphabet disclosed an $8 billion unrealized gain, which was widely believed to be related to SpaceX's ballooning valuation. If Alphabet still owns 7% of SpaceX, that investment would be worth $140 billion if SpaceX reaches a valuation of $2 trillion. Even for a company the size of Alphabet, that's a huge gain, and would be more than its net income last year. That stake would also be especially valuable to Alphabet right now, as it would become liquid, and the company is set to plow roughly $175 billion into capital expenditures this year to support its AI ambitions. A windfall from SpaceX would certainly help with that. Nvidia (NASDAQ: NVDA) isn't an investor in SpaceX, but it could benefit from the blockbuster IPO in another way. SpaceX is a major customer of Nvidia, which has become the most valuable company in the world due to its dominance of the AI GPU market. Elon Musk has expressed admiration for Nvidia and CEO Jensen Huang on several occasions, and recently said that both SpaceX and Tesla would continue to purchase Nvidia chips at scale. SpaceX's acquisition of xAI also puts it in a position to spend even more on Nvidia chips than it would have as funding xAI, which owns the social media site X and the chatbot Grok, with SpaceX's profits from Starlink, was a major part of the rationale for the merger. Musk wants xAI to compete with OpenAI and Anthropic, which will require large investments in Nvidia chips. It's a good bet, then, that a significant portion of the $75 billion that SpaceX could raise will be spent on Nvidia chips. Over the longer term, Musk also envisions putting data centers in space. While that might sound like a fantastical idea, he aims to use the funding from the IPO to launch data center satellites, arguing that space-based data centers can be more easily powered and cooled. Though the idea has been met with skepticism, if Musk pulls it off, it could spark even more demand for Nvidia chips. Whatever happens with the data center plan, Nvidia figures to be a winner from SpaceX's payday. The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now... and Nvidia 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 $532,066!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $1,087,496!* Now, it's worth noting Stock Advisor's total average return is 926% -- a market-crushing outperformance compared to 185% 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. Jeremy Bowman has positions in Nvidia. The Motley Fool has positions in and recommends Alphabet, Nvidia, and Tesla. The Motley Fool has a disclosure policy.
The UK government reportedly wants Anthropic to expand its presence in London
While the US and Anthropic are in the midst of a major dispute, the UK is trying to sway the San Francisco-based AI company to expand its presence on English soil. According to a report from The Financial Times, staffers at the UK's Department for Science, Innovation and Technology have worked on proposals that include expanding Anthropic's office in London, along with a potential dual stock listing. The UK's strategy follows a public fallout between Anthropic and the US Department of Defense earlier this year. After the AI company said it wouldn't budge on certain AI guardrails, the Department of Defense pulled its contract and eventually designated Anthropic a supply chain risk. While the designation is currently temporarily blocked by a court-ordered injunction, the feud is far from over. In the meantime, the UK's efforts to court Anthropic have ramped up in the recent weeks thanks to the company's disagreements with the US, according to FT's sources. With no end in sight for the debacle with the Department of Defense, Anthropic's CEO, Dario Amodei, is expected to visit the UK in May, according to FT. However, even in London, Anthropic will have to compete against OpenAI, which already committed to expanding its footprint in the English capital in February.
Big banks seeking a piece of spacex's IPO must subscribe to Musk's grok
Some of the banks have agreed to spend tens of millions on the chatbot, and they have already started integrating Grok into their IT systems It's not uncommon for large companies doing big deals to make demands of their bankers and lawyers. But Elon Musk has made a particularly bold demand of his Wall Street advisers ahead of the initial public offering (IPO) of his company SpaceX. Musk is requiring banks, law firms, auditors and other advisers working on the IPO to buy subscriptions to Grok, his artificial intelligence (AI) chatbot, which is part of SpaceX, said four people with knowledge of the matter, who were not authorised to speak publicly about confidential discussions. Some of the banks have agreed to spend tens of millions on the chatbot, and they have already started integrating Grok into their IT systems, three of the people said. Musk and a SpaceX spokesman did not respond to requests for comment. Also Read ECM fees rise sharply even as investment banking revenues slump in Q1premium SpaceX targets more than $2 trillion valuation ahead of record-breaking IPO Tesla sales rise after year of Musk boycotts but still miss expectations Powerica makes muted debut, lists at 7% discount; misses GMP estimates Elon Musk, US SEC head toward trial over Twitter investor fraud case For almost any major initial public offering, banks find ways to ingratiate themselves with the company going public, as well as its chief executive. But after several years with few significant public offerings coming to market, Wall Street has been salivating for a deal like SpaceX, which is forecast to be one of the largest in history. The IPO is expected to raise more than $50 billion at a valuation above $1 trillion, which means the banks could generate fees in excess of $500 million for advising on the deal. Musk's ability to secure business from the banks for his AI chatbot also shows the enormous sway of the world's richest man over a banking sector clamoring for his business now and into the future. The banks' purchases of Grok subscriptions were not merely good-will gestures, said three people with knowledge of the arrangements. Musk insisted that they purchase the chatbot services. He has also asked the banks to advertise on X, which is also owned by SpaceX, but was less adamant about that request, according to two of those people. For now, five banks are expected to work on the offering -- Bank of America, Citigroup, Goldman Sachs, JPMorgan Chase and Morgan Stanley. The law firms Gibson Dunn and Davis Polk are also advising on the deal. Musk's agreement with banks is a big score for SpaceX, which merged with xAI in February and whose Grok is a distant fourth in the AI race behind OpenAI's ChatGPT, Claude and Google's Gemini. Musk has marketed Grok as the antidote to political correctness and said his chatbot would not be "woke," unlike its competitors. In recent months, Grok has been mired in controversy after sharing antisemitic content as well as generating nonconsensual sexualised images of women and girls. Despite its problems, Musk has continued to promote the chatbot, regularly urging his more than 237 million followers on X to "try Grok." Grok generates revenue mostly from individuals rather than from businesses. The subscriptions from the banks will give the so-called enterprise part of the AI arm a boost ahead of SpaceX's IPO. In its most recent financial report to investors before the SpaceX merger, xAI reported roughly $1 billion in revenue from its AI operations, according to a person who viewed the results. Starlink, SpaceX's satellite internet service, generates billions of dollars in so-called free cash flow from its operations, said a person familiar with its financials. Documents obtained by The New York Times showed that Starlink recorded about $8 billion in revenue in 2024. It's unclear which bank, if any, will have the lead role in the deal, a position that carries prestige and often an outsize share of the fees, according to two people familiar with the negotiations. SpaceX, which confidentially filed IPO paperwork with the Securities and Exchange Commission this week, left the names of banks off the filing, one of the people said. Conditions apply * Some banks have agreed to spend tens of millions of dollars and begun integrating Grok into their internal IT systems * With SpaceX expected to raise over $50 billion at a valuation above $1 trillion, the banks could generate fees in excess of $500 million for advising on the deal * Beyond Grok, Musk has also asked banks to advertise on X * Subscriptions from the banks could strengthen Grok's corporate revenue profile ahead of listing More From This Section US troops fighting for Jesus, says Hegseth; Pope Leo XIV disagrees Trump vows strikes on Iran's infra after US pilot rescued in 'daring op' OPEC+ agrees in principle to raise output amid Iran war disruptions Iraq thanks Iran for allowing oil tankers through Strait of Hormuz Donald Trump says US will target Iran's infrastructure on Tuesday
Climate chaos accelerating
Those on the frontlines urgently need a surge in adaptation finance to face the here-and-now impacts of the climate crisis. Planet Earth is being pushed beyond its limits. Every key climate indicator is flashing red. That is the clear message of the latest report from the World Meteorological Organization. The report confirms that the Earth's energy imbalance -- the gap between heat absorbed and heat released -- is the highest on record. In other words, our planet is trapping heat faster than it can shed it. Greenhouse gas concentrations are higher than at any point in hundreds of thousands of years. Global temperatures continue to rise. Humanity has just endured the eleven hottest years on record. When history repeats itself eleven times, it is no longer a coincidence. It is a call to act. Meanwhile, oceans are absorbing epic levels of heat, fueling ever stronger storms. Glaciers and sea ice are vanishing. And sea-levels are relentlessly rising. These findings are not confined to charts and graphs. They are written into the daily lives of people. In families struggling as droughts and storms drive up food prices; Those on the frontlines urgently need a surge in adaptation finance to face the here-and-now impacts of the climate crisis. And in this age of war, climate stress is also exposing another truth: our addiction to fossil fuels is destabilizing both the climate and global security. Now more than ever, we must accelerate a just transition to renewable energy. Renewables deliver climate security, energy security and national security. Today's report should come with a warning label: climate chaos is accelerating and delay is deadly. The way ahead must be grounded in science, common sense and the courage to act.
2 Stocks About To Make a Killing on the SpaceX IPO | The Motley Fool
SpaceX is a major customer of Nvidia, especially now that it owns xAI as well. SpaceX is getting ready to go public, and it could be the biggest offering in history. Elon Musk's space company, which recently merged with his AI company, xAI, filed confidentially to go public last week, and SpaceX is reportedly seeking a valuation of up to $2 trillion. The company was valued at $1.25 trillion in the merger earlier this year, though that number was determined by the board of directors and investment bankers. The last time investors bought shares of the company was at the end of 2025, and its valuation was $800 billion. SpaceX is reportedly seeking to raise as much as $75 billion, which would make the largest IPO raise on record. At that valuation, it would be one of the most valuable companies in the world. The IPO is clearly a windfall for its investors, including Elon Musk, whose net worth could top $1 trillion after it goes public, but there are other overlooked winners in the stock market that investors should be aware of, as these stocks could get a meaningful boost from the IPO. Keep reading to see two of them. Alphabet (GOOG 0.20%) (GOOGL 0.57%) is best known as the parent of Google and the world's biggest advertising company, but it's also a prolific investor. It owns stocks in more than 20 publicly traded companies, as well as a number of start-ups, including through its Google Ventures arm. Alphabet invested $900 million in SpaceX in 2015, getting a stake of about 7% in Musk's space company, and it's believed to own approximately that much today. In its Q1 2025 report, Alphabet disclosed an $8 billion unrealized gain, which was widely believed to be related to SpaceX's ballooning valuation. If Alphabet still owns 7% of SpaceX, that investment would be worth $140 billion if SpaceX reaches a valuation of $2 trillion. Even for a company the size of Alphabet, that's a huge gain, and would be more than its net income last year. That stake would also be especially valuable to Alphabet right now, as it would become liquid, and the company is set to plow roughly $175 billion into capital expenditures this year to support its AI ambitions. A windfall from SpaceX would certainly help with that. Nvidia (NVDA +0.87%) isn't an investor in SpaceX, but it could benefit from the blockbuster IPO in another way. SpaceX is a major customer of Nvidia, which has become the most valuable company in the world due to its dominance of the AI GPU market. Elon Musk has expressed admiration for Nvidia and CEO Jensen Huang on several occasions, and recently said that both SpaceX and Tesla would continue to purchase Nvidia chips at scale. SpaceX's acquisition of xAI also puts it in a position to spend even more on Nvidia chips than it would have as funding xAI, which owns the social media site X and the chatbot Grok, with SpaceX's profits from Starlink, was a major part of the rationale for the merger. Musk wants xAI to compete with OpenAI and Anthropic, which will require large investments in Nvidia chips. It's a good bet, then, that a significant portion of the $75 billion that SpaceX could raise will be spent on Nvidia chips. Over the longer term, Musk also envisions putting data centers in space. While that might sound like a fantastical idea, he aims to use the funding from the IPO to launch data center satellites, arguing that space-based data centers can be more easily powered and cooled. Though the idea has been met with skepticism, if Musk pulls it off, it could spark even more demand for Nvidia chips. Whatever happens with the data center plan, Nvidia figures to be a winner from SpaceX's payday.
Anthropic: You Can't Use OpenClaw With Claude Without Paying Extra
AI firm Anthropic has clamped down on people using its Claude chatbot via the agentic AI tool OpenClaw. OpenClaw, released in late 2025, is a free and open-source AI agent that can execute tasks via third-party large language models (LLMs) like Claude, ChatGPT, Grok or Google Gemini. The app, which was even called "Clawd" when it debuted, has always had a close relationship with Anthropic's LLMs. As of 4 April 2026, Claude Pro and Max subscribers can no longer use their available credits through third-party frameworks such as OpenClaw. They are allowed to keep using tools like OpenClaw in theory, but they must now pay separately under a new "extra usage" billing system. Anthropic's Boris Cherny, head of Claude Code, chalked up the move to cost concerns in a post on X, saying the tool's subscriptions "weren't built for the usage patterns of these third-party tools." "Capacity is a resource we manage thoughtfully, and we are prioritizing our customers using our products and API," said Cherny. Cherny explained that his firm was a "big fan of open source," and that the move "is more about engineering constraints." Though the move currently only impacts OpenClaw, the policy is set to be rolled out to other third-party tools in the coming weeks. The news follows OpenClaw's creator Peter Steinberger joining OpenAI in February. The move saw OpenAI's CEO Sam Altman promise to support OpenClaw as an open source project. Steinberger claimed he had "begged" Anthropic to reconsider the recent restriction on OpenClaw, but was only able to delay the change by a week. The Claude-maker isn't the only AI giant that is shutting down popular products due to resource concerns. Last month, OpenAI shut down its AI video generator Sora, both the consumer app and the API. A spokesperson said the company wanted to free up resources for other areas of the business, such as "world simulation research to advance robotics."
Anthropic: You Can't Use OpenClaw With Claude Without Paying Extra
AI firm Anthropic has clamped down on people using its Claude chatbot via the agentic AI tool OpenClaw. OpenClaw, released in late 2025, is a free and open-source AI agent that can execute tasks via third-party large language models (LLMs) like Claude, ChatGPT, Grok or Google Gemini. The app, which was even called "Clawd" when it debuted, has always had a close relationship with Anthropic's LLMs. As of 4 April 2026, Claude Pro and Max subscribers can no longer use their available credits through third-party frameworks such as OpenClaw. They are allowed to keep using tools like OpenClaw in theory, but they must now pay separately under a new "extra usage" billing system. Anthropic's Boris Cherny, head of Claude Code, chalked up the move to cost concerns in a post on X, saying the tool's subscriptions "weren't built for the usage patterns of these third-party tools." "Capacity is a resource we manage thoughtfully, and we are prioritizing our customers using our products and API," said Cherny. Cherny explained that his firm was a "big fan of open source," and that the move "is more about engineering constraints." Though the move currently only impacts OpenClaw, the policy is set to be rolled out to other third-party tools in the coming weeks. The news follows OpenClaw's creator Peter Steinberger joining OpenAI in February. The move saw OpenAI's CEO Sam Altman promise to support OpenClaw as an open source project. Steinberger claimed he had "begged" Anthropic to reconsider the recent restriction on OpenClaw, but was only able to delay the change by a week. The Claude-maker isn't the only AI giant that is shutting down popular products due to resource concerns. Last month, OpenAI shut down its AI video generator Sora, both the consumer app and the API. A spokesperson said the company wanted to free up resources for other areas of the business, such as "world simulation research to advance robotics."
Regulatory agility key to faster access to life-saving medicines
Access to life-saving medicines often depends not only on scientific breakthroughs, but on how quickly and efficiently regulatory systems can evaluate and approve them. National regulatory authorities (NRAs) such as the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) play a vital role in safeguarding public health. They ensure that medicines, vaccines, and medical devices meet stringent standards of safety, efficacy, and quality before and after reaching the market. Without these safeguards, patients would be exposed to unsafe or ineffective products, and trust in healthcare systems would erode. At the same time, modern healthcare demands more than protection as it requires speed, adaptability, and global coordination. The research-based pharmaceutical industry works closely with NRAs to strengthen regulatory systems, accelerate access to innovative therapies, promote reliance on trusted regulatory decisions, and maintain the highest standards of quality and safety. These efforts include addressing bottlenecks, harmonizing standards such as Good Manufacturing Practices (GMP) and Good Documentation Practices (GDP), and improving preparedness for public health emergencies. The COVID-19 pandemic demonstrated what is possible when regulatory systems evolve. NRAs and the biopharmaceutical industry adopted flexible and iterative approaches to accelerate the development, evaluation, authorization, and distribution of vaccines and diagnostics. What once took years was achieved in record time, without compromising safety. These regulatory agilities included the use of digital tools, decentralized clinical trials, and rolling reviews of data. Equally important was the unprecedented level of collaboration among regulators, industry, and international platforms such as the International Coalition of Medicines Regulatory Authorities (ICMRA). This alignment enabled faster decision-making and more efficient deployment of critical health technologies. A key enabler of this speed was regulatory reliance. Under this approach, one regulatory authority considers and gives significant weight to the scientific assessments or approvals of another trusted authority or institution, such as the World Health Organization (WHO), while maintaining its own independent decision-making. A notable example was the WHO Emergency Use Listing (EUL), a risk-based mechanism used during public health emergencies to evaluate unlicensed vaccines, therapeutics, and diagnostics. Recognition of the WHO EUL allowed countries to expedite national authorizations, make efficient use of limited regulatory resources, avoid duplication, and accelerate patient access to COVID-19 vaccines. "To ensure the availability of COVID-19 vaccines and RT-PCR test kits in the country during the pandemic, we worked closely with the Philippine FDA, the Department of Health, and embassies of several countries to expedite regulatory processes and navigate border closures," said Dr. Diana Edralin, president of the Pharmaceutical and Healthcare Association of the Philippines (PHAP). Dr. Edralin emphasized that ensuring Filipinos have timely and sustained access to quality medicines remains a key priority, requiring close collaboration between government and industry. "Enhancing access to innovative medicines in the country requires a dynamic, resilient, and responsive regulatory system. This is why PHAP promotes the adoption of global best practices in pharmaceutical regulation from Europe, North America, and Southeast Asia," she added. The lessons from the pandemic are clear. However, regulatory flexibility should not be limited to emergencies. Applying these best practices in routine settings can significantly shorten the time it takes for patients to benefit from new therapies. Expedited pathways such as Fast Track, Breakthrough Therapy, and Accelerated Approval allow faster access to medicines that address serious conditions or unmet medical needs, while maintaining rigorous safety standards. Meanwhile, the use of Electronic Common Technical Document (eCTD) formats, along with digital tools for compliance tracking and document management, can improve efficiency, reduce errors, and streamline submissions. At the regional level, the Philippines' ASEAN Chairmanship presents a strategic opportunity to advance the ASEAN Harmonization of Pharmaceutical Registration. This initiative seeks to align technical requirements and regulatory processes across member states, facilitating trade while ensuring the safety, quality, and efficacy of medicines. Key mechanisms such as the ASEAN Common Technical Dossier (ACTD) and the ASEAN Pharmaceutical Regulatory Framework (APRF) help reduce duplication, lower regulatory costs, and enable faster access to medicines across Southeast Asia. Together, they represent an important step toward a more integrated and efficient regional regulatory system. Ultimately, improving access to innovative medicines requires regulatory systems that are not only robust, but also adaptive. By embracing reliance, harmonization, and digital transformation, countries like the Philippines can accelerate access to life-saving therapies while upholding the highest standards of patient safety. In doing so, regulatory excellence becomes not just a safeguard but a catalyst for better health outcomes. Teodoro B. Padilla is the executive director of Pharmaceutical and Healthcare Association of the Philippines, which represents the biopharmaceutical medicines and vaccines industry in the country. Its members are at the forefront of developing, investing and delivering innovative medicines, vaccines, and diagnostics for Filipinos to live healthier and more productive lives.
Anthropic: You Can't Use OpenClaw With Claude Without Paying Extra
AI firm Anthropic has clamped down on people using its Claude chatbot via the agentic AI tool OpenClaw. OpenClaw, released in late 2025, is a free and open-source AI agent that can execute tasks via third-party large language models (LLMs) like Claude, ChatGPT, Grok or Google Gemini. The app, which was even called "Clawd" when it debuted, has always had a close relationship with Anthropic's LLMs. As of 4 April 2026, Claude Pro and Max subscribers can no longer use their available credits through third-party frameworks such as OpenClaw. They are allowed to keep using tools like OpenClaw in theory, but they must now pay separately under a new "extra usage" billing system. Anthropic's Boris Cherny, head of Claude Code, chalked up the move to cost concerns in a post on X, saying the tool's subscriptions "weren't built for the usage patterns of these third-party tools." "Capacity is a resource we manage thoughtfully, and we are prioritizing our customers using our products and API," said Cherny. Cherny explained that his firm was a "big fan of open source," and that the move "is more about engineering constraints." Though the move currently only impacts OpenClaw, the policy is set to be rolled out to other third-party tools in the coming weeks. The news follows OpenClaw's creator Peter Steinberger joining OpenAI in February. The move saw OpenAI's CEO Sam Altman promise to support OpenClaw as an open source project. Steinberger claimed he had "begged" Anthropic to reconsider the recent restriction on OpenClaw, but was only able to delay the change by a week. The Claude-maker isn't the only AI giant that is shutting down popular products due to resource concerns. Last month, OpenAI shut down its AI video generator Sora, both the consumer app and the API. A spokesperson said the company wanted to free up resources for other areas of the business, such as "world simulation research to advance robotics."
Baidu Robotaxi System Failure Triggers Massive Traffic Chaos in Wuhan - News Directory 3
The outage left several passengers stranded for extended periods. Baidu's autonomous ride-hailing service, Apollo Go, experienced a widespread system failure on April 1, 2026, causing multiple robotaxis to stall in the middle of roads in Wuhan, China. The incident resulted in significant traffic disruptions, stranded passengers and at least one reported highway collision. Local authorities in Wuhan confirmed the event after receiving a surge of reports from riders on Tuesday night. The Wuhan local traffic police department stated that preliminary investigations suggest a system malfunction was the cause of the vehicles freezing in traffic. The outage left several passengers stranded for extended periods. One rider reported on the social media platform RedNote that their vehicle broke down on an elevated highway at 9:00 p.m. Local time on April 1, 2026. The user stated they were unable to reach customer service initially and remained stuck on the overpass until 10:30 p.m., when the order was cancelled. Another passenger told Wired that their cab stopped four times during a single trip before eventually parking in front of an intersection. The vehicle's internal screens instructed riders to remain onboard for a company representative. That passenger reported it took 30 minutes to reach a customer service agent before the group decided to exit the vehicle and seek alternative transportation. The failure extended beyond passenger inconvenience to road safety. Dash cam footage shared on social media, including the platform X, appeared to show a human-driven vehicle crashing into the rear of a stranded Baidu car that had come to a stop on a busy, multi-lane highway. Wuhan serves as the site of one of the largest experiments in self-driving technology and is home to Apollo Go's largest robotaxi deployment in China. Reports on the scale of the fleet vary, with some sources stating Baidu operates more than 500 driverless cars in the city, while others indicate the fleet exceeds 1,000 vehicles operating without human drivers. The incident occurred as Baidu continues to scale its autonomous driving unit to compete with other regional providers such as WeRide and Pony. This expansion mirrors the strategies of U.S.-based companies like Alphabet, which has scaled its Waymo service in the western United States. The Wuhan local traffic police department acknowledged the reports via its official Weibo account on Wednesday, April 2, 2026. The police confirmed that they worked with Apollo Go staff to manage the initial situation and stated that passengers had safely exited the affected vehicles. Despite the police confirmation and widespread social media reporting, Baidu did not immediately respond to requests for comment from CNBC regarding the system failure or the resulting collisions. The police department noted that the incident is still undergoing further investigation to determine the exact nature of the system malfunction that caused the fleet to stall.
:max_bytes(150000):strip_icc()/GettyImages-2152484525-18bfb39b460249f3ae213d800ed5bfca.jpg)