Breaking AI & Tech News Analyzed

The Leadership Shift at Anthropic Andrej Karpathy, the AI researcher who co-founded and formerly worked at OpenAI and previously led AI at Tesla, has joined Anthropic. Karpathy announced his move on X, stating that he is excited to join the team and get back to R&D.; Karpathy's Role in Pre-Training Karpathy started this week at Anthropic, where he is working on pre-training under team lead Nick Joseph. Pre-training is responsible for the large-scale training runs that give Claude its core knowledge and capabilities. Karpathy will start a team focused on using Claude to accelerate pre-training research. The Significance of Karpathy's Move Karpathy is one of the few researchers who can bridge the gap between LLM theory and large-scale training practice. Tapping him to build such a team is a clear sign from Anthropic that it believes AI-assisted research, rather than pure compute, is how it stays competitive with OpenAI and Google. Karpathy's Background Co-founded OpenAI and worked on deep learning and computer vision until 2017 Led Tesla's Full Self-Driving (FSD) and Autopilot programs from 2017 to 2022 Returned to OpenAI for one year before leaving in 2024 to start Eureka Labs, a startup dedicated to applying AI assistants to education Anthropic's Recent Hires Anthropic has also brought on Chris Rohlf to its frontier red team, which stress-tests advanced AI models against severe threats. Rohlf is a veteran of the cybersecurity industry with more than 20 years of experience. The Future of AI Research Karpathy's move to Anthropic and the company's focus on AI-assisted research signal a new direction in the AI landscape. As Karpathy stated, "I think the next few years at the frontier of LLMs will be especially formative."

Enhanced VoiceOver Capabilities Apple announced several new accessibility updates to features like VoiceOver, voice control, live recognition, and real-time caption generation for videos that will be powered by Apple Intelligence. The company also said that Vision Pro users will be able to control a compatible wheelchair with their eyes using this update. Improved Image Recognition The company said that Apple Intelligence’s image-recognition feature in VoiceOver will understand the image better and describe it in greater detail. For instance, it can look at a bill and read out the details such as the amount and due date. Apple said that the updated feature can also better describe photographs and personal records. Live Recognition and Voice Commands Users can now use an iPhone to activate the Live Recognition feature, which uses a camera to identify content in the frame, and also ask follow-up questions to know more. Users with low vision can also assign Magnifier to the action button, which presents content on a high-contrast interface. Plus, they can use voice commands such as “zoom in” or “turn on flashlight” to access features. Natural Language Processing The voice command update is not limited to the Magnifier app. Apple said that users can describe tasks in natural language to take action on what they see on the screen. For instance, in Apple Maps, they can say, “Tap the guide about best restaurants,” or in Files, they can say “Tap the purple folder.” Reader Updates The company is also updating Reader, which can now handle documents like scientific papers with multiple columns, images, and tables. Users can get AI-powered summaries or read the text in the native language with custom fonts and colors retained. AI-Generated Subtitles and Wheelchair Control Apple is adding AI-generated subtitles for videos that don’t have pre-generated captions. The company announced a new project where Vision Pro users can control their compatible wheelchairs with their eyes. Additional Features and Availability The name-recognition feature, which notifies users with hearing disabilities when someone says their name, now supports 50 languages. Apple will roll out large text support to tvOS. These features will be available to users later this year, likely as part of Apple’s upcoming iOS release.

The Rise of AI-Generated Hate OperationsScroll through any Facebook feed in Britain and, between the baby announcements and petty neighbourhood beefs, you're likely to come across an account with a union jack profile picture and a vague, generic name like Britain Today. These accounts – and there are hundreds, possibly thousands of them – present themselves as the work of British patriots. In one typical, AI-generated video, a middle-aged man claims his local cafe "has stopped serving pork, bacon and sausages just to avoid offending people". Another post from the same account includes a sepia-tinted set of images of Victorian London, mourning a time when the city "was English, first-world and beautiful". Alongside this type of reactionary nostalgia, it's not unusual to see memes that call Islam a "cancer", decry Muslims praying in public as an "invasion of the west" or promote the "great replacement theory".The Financial Incentives Behind AI Hate ContentFor the past seven months, I have been investigating who is really behind pages like these. The answer, it turns out, is often young, entrepreneurial men from south Asia. They tend to have zero interest in UK politics, but the content they create often boosts far-right talking points in Britain and contributes to the increasingly hostile atmosphere for immigrants and British Muslims. They're part of a booming cottage industry producing commercial AI slop.The financial incentives for creating this kind of content are huge, particularly for creators in the global south. At the Bureau of Investigative Journalism, we looked in detail at two very successful "sloperations" targeting British audiences from Pakistan and Sri Lanka. They make money from the online ads that Meta places next to high-performing content. Meta shares a proportion of the ad revenue with the creators and also makes direct payments to creators to reward posts that receive a lot of engagement.Once you hone your algorithmic rage bait, there's very good money to be made from slop. The Pakistani creator, a devout Muslim who we are not naming for his own safety, told us he makes $1,500 (£1,119) a month from one of his pages alone; Geeth Sooriyapura, the Sri Lankan creator, claimed to have made $300,000 over the course of his Facebook career. We weren't able to verify these figures, but both men were certainly making many times the average income in their countries.The Economic Impact of AI-Generated PropagandaTheir success represents the seductive promise of "passive income" culture, a pervasive modern gospel that says you should quit your job and make easy money online. The proponents of this philosophy also often sell courses as an additional revenue stream: Sooriyapura claimed that 2,500 people, mainly other Sri Lankans, have graduated from his content academy.Rightwing propaganda and Islamophobia are, of course, not new. But two key structural factors have made it particularly pervasive on social media.The Technological and Policy EnablersFirst, the wide availability of generative AI tools. These are used at every stage of the content creation process: to brainstorm ideas, to write captions and, most importantly, to create compelling images and videos. This is particularly helpful if, like the Pakistani creator, you do not speak English well. In one video we reviewed from Sooriyapura's Facebook course, he told his students that AI-generated videos can help political content go viral up to 10 times faster.Second is Meta's retreat from content moderation. Over the past couple of years, the major social platforms have made mass redundancies on the trust and safety teams that monitored and took down harmful content. This was partly motivated by pressure from the Trump administration, which believed that platforms had engaged in heavy-handed censorship of content during the Biden presidency.Social media companies justify the moderation job cuts by pointing to their use of AI to find harmful content more efficiently. But our reporting shows there is masses of deeply offensive content on there which anyone could find in a few minutes, if they bothered to look.The Future of Online Hate Speech and Platform AccountabilityAfter we spoke to the Pakistani creator, he said it was a "good thing" we had informed him about the nature of his posts and he deleted many of them. Sooriyapura told us that he did not encourage his students to "spread violence" and that he just educates "people on Facebook monetisation and audience-targeting".The Pakistani creator didn't cover his tracks particularly well. It took me a couple of hours and a little help from Osint Industries, a platform that collates information on social media accounts, to definitively confirm that the person who ran the Islamophobic slop account also had personal accounts in his own name sharing verses from the Qur'an. These are actions that Meta easily could have taken itself. But why would it spend good money implementing its own policies when there is so little political or regulatory pressure to do so?When we contacted Meta in both these cases, it took down many of their pages and sent a one-line statement: "We have clear community standards that prohibit hate speech, harassment, harmful misinformation and inauthentic behaviour and we have removed these accounts for violating our policies." I've been a tech journalist long enough to have been through this process with Meta and other social platforms many times before. The Sri Lanka network is, depressingly, back up and running, having faced minimal consequences after a bit of downtime.Meta can, and should, be doing more to take these kinds of accounts down. But as long as its core product is an algorithmic feed that financially rewards content that provokes extreme emotions, others will always appear in its place.

The Lead A US jury has handed a resounding victory to Sam Altman and OpenAI in their long, bitter courtroom battle with Elon Musk, finding Altman, OpenAI, and its president, Greg Brockman, not liable for Musk's claims that they unjustly enriched themselves and broke a founding contract. The Court Ruling and Its Implications The unanimous verdict, delivered after less than two hours of deliberation, is a stark rebuke of Musk and his lawyer's claims that Altman 'stole a charity' through his leadership of OpenAI. The jury's decision, affirmed immediately by the judge's dismissal of all charges, provides OpenAI with a stamp of approval for its for-profit plans, already in motion, and a clear path ahead to go public later this year at around a $1tn valuation. The Financial Impact Musk's demands that Altman be removed as CEO and that the for-profit arm of the company transfer some $150bn to the non-profit arm would have jeopardized the blockbuster initial public offering. A delay to OpenAI's financial bonanza may have been one of Musk's goals, as his own company, SpaceX, is reportedly planning to go public in June. The Industry Impact OpenAI's plans now seem all but guaranteed, given that the world's richest man couldn't put a stop to them. Wall Street, ever wary of upheaval and uncertainty, is likely breathing a sigh of relief, said professor Sarah Kreps, director of the Tech Policy Institute at Cornell University. She called the ruling a reflection of the tough reality that developing frontier AI is expensive and that maintaining non-profit status is not viable in the face of fierce, capital-intensive competition. The Future Outlook The trial did not deliver answers to major questions of the AI boom about safety, governance, and labor. While the jury's verdict was a 'technical' one, Musk's lawyers said he would appeal the case. The trial demonstrated that a small cabal, mostly men, rules the AI industry, and its central element was not a fight over AI's benefit to humanity but a hateful vendetta that Musk brought against Altman.

The Growing Concern Over AI's Economic ImpactOne in three university students in Great Britain believe that artificial intelligence will eliminate jobs so rapidly that it will trigger civil unrest, according to a new survey by King's College London (KCL). This significant finding highlights the deep concerns among educated young people about the potential societal consequences of rapid technological advancement.The poll, conducted by the King's Institute for Artificial Intelligence and the KCL Policy Institute, represents the first major tracking study of attitudes toward AI across different segments of British society. It compares responses from university students, young people aged 16 to 29, employers, and the general public.Student Usage Patterns and PessimismUniversity students emerge as among the heaviest users of AI technology, with 77% reporting using it at least a few times a month—substantially higher than the 46% of workers who do so. Additionally, 27% of students use AI daily or almost daily, indicating deep integration of these tools into academic life.Despite their familiarity with AI, students express significant pessimism about its economic consequences. More than half are convinced that job losses resulting from AI will be more severe than those in a typical recession. This pessimism is particularly notable given that students generally hold more positive views about AI's overall impact on humanity compared to the general public.Key Statistics from the AI Attitudes Survey34% of university students believe AI will eliminate jobs fast enough to cause civil unrest (compared to 22% of the general public)77% of university students use AI at least a few times a month (compared to 46% of workers)27% of university students use AI daily or almost daily52% of male university students believe AI is positive for humanity (compared to 24% of the general public)9 out of 10 university students have encountered problems with AI, most commonly factual errors (37%) and made-up sources (31%)78% of students would still choose to attend university, though 30% would have selected a different subjectImplications for Education and the WorkforceThe survey reveals a significant gap between students' perceptions of their preparedness for an AI-shaped job market and their actual experiences. While 60% believe universities are capable of preparing them for this future, only 36% report actually receiving adequate preparation.This disconnect suggests that educational institutions may be struggling to adapt curricula and teaching methods to address the rapidly evolving technological landscape. The findings also highlight gender differences in how students perceive AI's impact on their cognitive abilities, with male students more likely to believe AI enhances their thinking skills while female students tend to hold the opposite view.Divergent Views on AI's FutureThe poll captures contrasting perspectives on AI's potential impact. Bobby Duffy, director of the KCL Policy Institute, emphasizes the widespread concern about AI's effect on employment, particularly at entry-level positions, and its broader implications for young people and the economy.In contrast, Bouke Klein Teeselink, a lecturer in philosophy, politics, and economics at KCL, offers a more optimistic outlook. He suggests that with appropriate training, policies, and institutional support, AI could lead to increased productivity, expanded opportunities, higher incomes, and accelerated scientific progress.These divergent views reflect the broader societal debate about artificial intelligence—balancing legitimate concerns about displacement and inequality against the potential benefits of technological advancement.

The Vatican's Digital Turn: Pope Leo XIV's First Encyclical on AIIn a groundbreaking move that signals a significant shift in the intersection of faith and technology, Pope Leo XIV is preparing to release his first major papal document addressing the rapid ascent of artificial intelligence. The encyclical, titled Magnifica Humanitas (Magnificent Humanity), represents the Vatican's attempt to provide moral guidance in an era defined by digital transformation.Historic Collaboration: The 'Magnifica Humanitas' LaunchThe presentation of this document is set for 25 May at the Vatican, marking a departure from tradition. For the first time, the encyclical will be launched during a public event attended by Christopher Olah, the co-founder of Anthropic, a leading US-based AI firm currently embroiled in a high-profile lawsuit with the Trump administration over federal agency use of AI.Document Title: Magnifica Humanitas (Magnificent Humanity)Sign Date: 15 MayKey Attendees: Pope Leo XIV, Christopher Olah, theologians Anna Rowlands and Léocadie LushomboA 135-Year Parallel: From the Industrial to the Digital RevolutionThe timing of the document is deeply symbolic. Signed on 15 May, exactly 135 years after his namesake Pope Leo XIII signed the seminal Rerum Novarum, the new encyclical mirrors the historical response to the Industrial Revolution. While Leo XIII addressed the challenges of capitalism and workers' rights, Leo XIV is addressing the challenges of the technological revolution.Redefining AI Ethics Through a Moral LensThe encyclical is expected to move beyond simple warnings, aiming to offer 'workable answers' to modern challenges. Key themes include:The protection of workers' rights in an automated economy.A strong stance against the use of AI in warfare, specifically advocating for a ban on lethal autonomous weapons.Ensuring technological advancements do not override human dignity.Shaping Global AI RegulationThis strategic move by the Vatican—featuring lay speakers and the Pope in person—suggests a new era of engagement with the tech industry. By inviting figures like Christopher Olah, the Church is signaling a willingness to collaborate with industry leaders rather than simply critique them. This document is poised to become a critical reference point in the global debate on AI safety, regulation, and the ethical boundaries of machine intelligence.

The Leap: Conversational Access to Quantitative Drug‑Discovery ModelsIn a bold move to democratize high‑performance chemistry, SandboxAQ has integrated its proprietary large quantitative models (LQMs) into Anthropic’s conversational AI, Claude. The partnership eliminates the need for users to provision costly computing resources, allowing scientists to query complex quantum‑chemistry simulations in natural language.SandboxAQ Teams with Anthropic to Embed LQMs in ClaudeThe five‑year‑old Alphabet spin‑out, chaired by Eric Schmidt, announced the integration after raising $950 million from investors. The LQMs are “physics‑grounded,” meaning they are built on scientific equations and real‑world lab data rather than purely on text patterns. They can perform quantum chemistry calculations, molecular‑dynamics runs, and micro‑kinetics simulations, delivering predictions about candidate molecules before any wet‑lab work begins.Financial and Market Scale of the Quantitative Economy$950 million raised to date by SandboxAQ.The company positions its LQMs within a $50+ trillion quantitative economy spanning biopharma, finance, energy, and advanced materials.Traditional drug‑discovery projects can cost billions of dollars and take a decade to yield a viable molecule.Why a Conversational Interface Could Disrupt Pharma R&D;Historically, only computationally sophisticated teams could leverage large‑scale chemistry models, requiring on‑premise GPUs or cloud clusters. By surfacing these capabilities through natural‑language chat, SandboxAQ lowers the barrier for:Computational scientists seeking rapid hypothesis testing.Experimentalists who lack deep AI‑infrastructure expertise.Large pharmaceutical and industrial firms aiming to accelerate material discovery.Customers have reported that existing software failed to translate complex problems into actionable results, a gap SandboxAQ hopes to fill.Future Outlook: Scaling AI‑Driven Chemistry Across IndustriesWith the Claude integration, SandboxAQ expects broader adoption beyond pharma, extending into energy, finance, and advanced materials where quantitative simulations are critical. As more firms adopt conversational AI for scientific workflows, the competitive advantage will shift from model performance to usability and integration speed. The next wave may see LQMs embedded in other enterprise assistants, further blurring the line between AI chat and high‑performance scientific computing.

The Acquisition Deal Anthropic announced Monday it has acquired Stainless, a startup founded by former Stripe engineer Alex Rattray whose software is widely used by rival AI labs, including OpenAI and Google. Stainless' Technology and Impact The New York-based startup, founded in 2022, rose to prominence in the emerging AI industry for automating the creation and maintenance of software development kits, or SDKs — the libraries developers use to interact with APIs. Rattray developed software that could take API specifications and turn them into production-ready SDKs across multiple programming languages, including Python, TypeScript, Kotlin, Go, and Java. Financial Terms and Future Plans Anthropic didn’t disclose terms of the deal. However, The Information reported last week that Anthropic was in talks to acquire Stainless, which is backed by Sequoia Capital and Andreessen Horowitz, for more than $300 million. The acquisition will take a key infrastructure supplier out of the hands of Anthropic’s competitors. The company told TechCrunch it will wind down all hosted Stainless products, including its SDK generator. Impact on the AI Industry The technology is particularly valuable to companies like Anthropic, OpenAI, Google, Replicate, Runway, and Cloudflare that are building AI agents that can connect to external software and complete tasks on behalf of users. Stainless’s SDK tools are an easy way to build and maintain those connections — but going forward, the tools will only be available to Anthropic, not its competitors. Future Outlook According to Anthropic, Stainless software has powered the generation of every official Anthropic SDK since the earliest days of its API. “I started Stainless because SDKs deserve as much care as the APIs they wrap,” Rattray said in a press release posted Monday. “Anthropic was one of the first teams to bet on this with us. We have been watching what developers have built on Claude over the last few years, which made bringing our teams together an easy decision. The team gets to keep doing the work we love, on the platform where it matters most.”

The Legal Victory for OpenAI's Leadership In a decisive moment for the artificial intelligence industry, a federal jury in Oakland, California has ruled in favor of Sam Altman and Greg Brockman, OpenAI's president, in their high-stakes legal battle against Elon Musk. The nine-person jury found the OpenAI leaders not liable for unjustly enriching themselves or breaking contracts made with Musk when founding the startup. This verdict represents a significant legal victory for Altman and a stark rebuke of Musk's central claim that Altman "stole a charity" through his leadership of OpenAI. The Courtroom Decision and Its Implications The jury's finding, while non-binding and advisory, carries substantial weight as Judge Yvonne Gonzalez Rogers immediately indicated she would agree with the jury's decision. This alignment between jury verdict and judicial ruling effectively ends the legal chapter of Musk's ambitious lawsuit, which sought $134 billion to be redistributed from OpenAI's for-profit arm to its non-profit component. The case also demanded the removal of Altman and Brockman from their roles at OpenAI and the undoing of the firm's for-profit restructuring. Musk's Core Allegations Against OpenAI At the heart of the three-week trial was Musk's allegation that Altman, Brockman, and OpenAI breached their founding agreement when they restructured the company into a for-profit entity. Musk accused the defendants of breach of charitable trust and unjust enrichment, claiming that Altman had deceived him into co-founding OpenAI in 2015 as a non-profit dedicated to bettering humanity, only later to twist the organization's purpose to pursue personal gain. This narrative formed the foundation of Musk's legal challenge against the company he helped establish. OpenAI's Defense Strategy OpenAI's legal team systematically rejected all of Musk's claims, asserting that he was always aware of plans to create a for-profit entity from the company's inception. The defense highlighted that Musk's motivations stemmed from jealousy after his failed attempt to take over OpenAI in 2018, which led to his departure from the company shortly thereafter. OpenAI representatives repeatedly emphasized that the company remains overseen by its nonprofit organization and remains dedicated to what it refers to as "the mission" of helping the world with its AI technology. The Silicon Valley Showdown The trial delivered unprecedented access to the inner workings of OpenAI and featured testimony from several of Silicon Valley's most prominent executives. Beyond the primary litigants, Musk, Altman, and Brockman, Microsoft CEO Satya Nadella also took the stand, facing combative cross-examinations that revealed the intense personal and professional dynamics at play. The proceedings brought in many current and former OpenAI executives, as well as academic experts on nonprofit law and corporate governance, creating a comprehensive record of the company's founding and evolution. The Future of OpenAI Post-Verdict With this legal challenge behind them, OpenAI can now focus on its ambitious AI development initiatives without the cloud of Musk's lawsuit hanging over its leadership structure. The verdict reinforces the company's current governance model and its transition toward a for-profit entity while maintaining its nonprofit oversight. For the AI industry at large, this outcome provides stability to one of its most influential organizations during a critical period of technological advancement. The case also sets a precedent for how founding agreements in tech startups are interpreted when companies evolve their business models in response to market pressures and technological opportunities.