Breaking AI & Tech News Analyzed

Goldman Sachs's chief executive, David Solomon, has expressed heightened awareness of the capabilities of Anthropic's Mythos AI model and is collaborating closely with the tech firm following warnings about the cybersecurity risk it poses.The US bank has been closely monitoring the rapid advancements in artificial intelligence, including large language models (LLMs), as part of broader efforts to protect itself from hackers.“Obviously the LLMs are making rapid progress and we’re hyper-aware of the enhanced capabilities of these new models with the help of the US government and the model publishers,” Solomon told analysts on an earnings call on Monday.Anthropic, the company behind the Claude family of AI tools, claimed last week that its latest model, Mythos, posed an unprecedented risk due to its ability to expose flaws in IT systems. The company warned that AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities.Solomon emphasized that Goldman Sachs is working closely with Anthropic and all of its security vendors to harness frontier capabilities. “We are very focused on supplementing our cyber and infrastructure resilience. And this is part of our ongoing capabilities that we have been investing in, and are accelerating our investment in.”The news comes after the US Treasury secretary, Scott Bessent, summoned Solomon and other big American bankers to Washington to discuss the Mythos model last week. The meeting focused on heads of so-called systemically important banks, where regulators believe that a major disruption to their operations, or their potential collapse, would put financial stability at risk.On Monday, the UK government’s AI Security Institute (AISI) warned that Mythos was a “step up” over previous models in terms of the cyber threat it posed. AISI said Mythos could carry out attacks that required multiple actions and discover weaknesses in IT systems without human intervention.

Meta, the company behind Facebook and Instagram, is reportedly working on an AI version of its CEO, Mark Zuckerberg. This AI clone is being trained on Zuckerberg's mannerisms, tone, and public statements to allow employees to interact with a digital version of their boss.The rationale behind this project is to make Meta's 79,000 employees feel more connected to one of the most influential figures in Silicon Valley. The AI character will be developed using images and the voice of Zuckerberg, with the CEO reportedly taking part in the training process.This move is part of Meta's broader effort to integrate AI into its business operations. The company aims to use AI to lower costs and accelerate work pace. Zuckerberg has emphasized the importance of efficiency, stating that the goal is to 'get more done' by elevating individual contributors and flattening teams.The development of this AI character follows Meta's previous experiments with digital avatars. In 2022, Zuckerberg shared his own avatar in the metaverse, which received public criticism for its graphic quality. The company has since scaled back its metaverse vision, focusing on AI-generated 3D characters for everyday conversations.Meta's investment in AI is part of a larger strategy to remain competitive with tech rivals. The company is pouring billions of dollars into AI research to create 'superintelligence,' a system capable of performing any cognitive task far better than a human.

Senator Bernie Sanders used a Manhattan rally on Sunday to issue a stark warning: “the worst is yet to come” for the U.S. economy unless workers confront a ruling class of billionaires. Sharing the stage with New York City mayor Zohran Mamdani, the two leaders announced the launch of Union Now, a nationwide drive to boost union density and provide resources for organizing and strikes. Sanders singled out high‑profile billionaires – Elon Musk (Tesla, SpaceX), Jeff Bezos (Amazon), and President Donald Trump – as the architects of a looming crisis. He warned that Musk’s push for robotics and AI, coupled with Bezos’s recent pledge to raise $100 billion for buying and automating manufacturing firms, threatens to replace human labor on a massive scale. “Unless we fundamentally transform our economic and political systems, the worst is yet to come,” Sanders declared, emphasizing that increasing union membership is the most effective tool to tackle income inequality. Mamdani echoed the sentiment, noting that artificial intelligence is “coming for human jobs” and that worker protections are eroding. He pledged his administration’s support for Union Now, describing the effort as essential for safeguarding workers’ rights. Data presented at the rally underscored the scale of wealth concentration: in 2025, 938 U.S. billionaires saw their net worth rise by $1.5 trillion, while Musk alone possesses more wealth than the bottom 53 % of Americans. Sanders painted the billionaire class as “extremely greedy” and likened their self‑perception to 19th‑century monarchs who believe they have a divine right to rule. He warned that their unchecked influence could leave future generations without a safety net. Highlighting a recent political victory, Sanders cited Mamdani’s mayoral win as proof that ordinary people can defeat billionaire‑backed opposition. He warned that if the current trajectory continues, “fewer people will have more wealth and power, democracy will be undermined, and workers will be left with no recourse.” Closing his speech, Sanders urged unity: “If we stand together and fight for a government that works for all of us, there is nothing we cannot accomplish.”

The ongoing conflict in Iran has created a perilous situation where the threat of nuclear war and climate disruption have converged into a single, catastrophic crisis. This crisis will persist long after the war subsides, emphasizing the need for immediate and decisive action. For over a decade, climate change has been recognized not just as an environmental issue but as a fundamental threat to national and global security. The current situation demands that governments conduct thorough risk assessments and treat climate change with the same urgency as military threats. The war on Iran has several alarming features: Unilateral action: The US and Israel launched a large-scale war against a sovereign nation without consulting major allies, creating a diplomatically isolated conflict with no clear exit strategy. Escalation threats: There are credible threats of escalation from both sides, with Donald Trump issuing ultimatums and Iran threatening to target critical infrastructure. Catastrophic miscalculation: The conditions for miscalculation are ripe, with erratic leadership, intelligence failures, and extreme pressure on decision-makers. The conflict has significant implications: Global energy shock: The closure of the Strait of Hormuz has caused an acute global energy shock, with higher oil prices likely to accelerate inflation and economic instability. Climate impact: The war is consuming military resources and political attention, with no climate dividend, and may pressure countries to extend fossil fuel use. Australia, as a regional power and signatory to the NPT, has responsibilities to the international order. The author, Admiral Chris Barrie, calls on the Australian government to take four key steps: Conduct and release a nuclear escalation risk assessment. Use diplomatic channels to counsel restraint. Refuse any form of complicity in nuclear use. Champion de-escalation at the NPT review conference. Australia can play a crucial role in addressing these threats by acting on evidence, speaking plainly about risks, and leading rather than following events.

When a Toronto reader confessed to always using "please" and "thank you" with Alexa, the Guardian invited its audience to share their thoughts on the etiquette of speaking to artificial assistants.One camp warns against treating AI as sentient beings, arguing that these systems merely mimic social cues and that confusing imitation with reality can erode our language’s power to challenge dehumanisation. As User30000 put it, we should avoid both de‑humanising language toward people and humanising language toward machines.Conversely, several contributors argue that politeness serves as a useful human default. Lauk notes that courteous phrasing may reinforce positive behavioural cues for AI that learn from user input, while also keeping the speaker grounded in genuine social norms.From an ecological standpoint, extra words translate into measurable energy and water waste. Superspartan highlights that each unnecessary token adds processing load to already energy‑intensive models, and that the cumulative effect of polite phrasing across millions of daily interactions could be substantial.Academic research supports the idea that courteous language can improve AI performance. A linked study finds that polite prompts often elicit more helpful replies, treating modern models as "statistical parrots" that mirror the tone they receive. Yet Sam Altman has publicly estimated that responding to thank‑you notes costs OpenAI tens of millions of dollars, underscoring a tangible financial dimension to the debate.Some readers take a lighter view, extending politeness to elevators, SUVs and even low‑battery iPads, while others, like Martin from Dorset, argue that the practice protects the human soul rather than the machine.There is growing concern that habitual rudeness toward AI may spill over into real‑world interactions. Poridgeoates observes that younger users, who spend more time with technology than with people, risk weakening empathy and emotional‑intelligence muscles if they treat conversational agents dismissively.Louise adds that the way we speak to objects can shape broader social behaviour, especially where power imbalances exist. While AI itself may not feel insulted, the act of being rude can reinforce undesirable character traits in the speaker.Overall, the discussion reveals a split between those who see politeness as a harmless habit that can improve AI responses, and those who view it as an unnecessary drain on resources and a potential threat to human civility.

OpenAI, a leading AI company, has recently released a 13-page policy paper titled 'Industrial Policy for the Intelligence Age,' which calls for a reimagining of the social contract around 'a slate of people-first ideas.' This move is part of an aggressive effort by major AI players to reshape the narrative around their industry, as public disapproval of AI is increasing.OpenAI's paper proposes ideas such as a four-day workweek and a public wealth fund that would return profits directly to citizens. While the company presents these ideas as a starting point for a broader conversation, critics argue that they are more of a public relations ploy than a genuine policy document.OpenAI spent nearly $3m on lobbying in 2025, and its president, Greg Brockman, co-founded a pro-AI Super Pac that raised more than $125m last year. The company is also backing a bill in Illinois that would shield AI firms from liability in cases where an AI model causes serious societal harms.Critics argue that these efforts are aimed at undermining independent efforts to regulate the industry and that the company's proposals shift responsibility away from the company and towards the public and lawmakers. As public distrust of AI grows, the industry is looking for ways to reframe the debate and influence regulation.A Pew Research Center survey found that only 16% of Americans believe that AI will help people think more creatively, while only 5% of Americans believe it will help people better form meaningful relationships. An NBC News poll found that only 26% of voters had a favorable opinion of AI, with a net negative rating.

Brazilian President Luiz Inácio Lula da Silva used his Friday social‑media post to herald a new security partnership with the United States as a breakthrough in the fight against transnational crime.The agreement links the Brazilian Federal Revenue Service with U.S. Customs and Border Protection, establishing the DESARMA programme. Under DESARMA the two agencies will exchange real‑time data, apply rigorous cargo‑tracking protocols and launch joint operations to intercept illicit shipments of arms and narcotics.Finance Minister Dario Durigan called the collaboration an “important step in strengthening international cooperation”, noting that it will integrate intelligence and joint actions to curb the flow of weapons and drugs between the two nations.The pact is part of President Donald Trump’s broader “Shield of the Americas” initiative, which seeks to enlist right‑wing governments across the Western Hemisphere in a hard‑line campaign against criminal networks.Despite the cooperation, Lula’s administration has pushed back against Washington’s attempts to label Brazilian gangs such as the Primeiro Comando da Capital (PCC) and Comando Vermelho (CV) as foreign terrorist organisations. Lula and Foreign Minister Mauro Vieira have repeatedly asserted Brazil’s sovereignty and warned against external “tutelage”.In the last twelve months Brazil reported seizing 1,168 illegally imported weapons and parts, most of which originated from the U.S. state of Florida, underscoring the scale of the smuggling problem.A statement from the Revenue Service emphasized that the deal will ensure a “continuous flow of information from U.S. authorities to their Brazilian counterparts”, enhancing operational coordination.Relations remain strained: Vieira told U.S. Secretary of State Marco Rubio that Brazil opposes the terrorist‑label designation, while the Trump administration continues aggressive actions in the region, including lethal strikes on maritime vessels and a controversial raid in Venezuela.Looking ahead, Lula is expected to travel to Washington, D.C., later this year for talks with Trump, providing a diplomatic channel to manage both cooperation and disagreement.

Brian Cox’s latest live show, ‘Emergence’, draws its spark from Johannes Kepler’s 1609 treatise ‘The Six‑Cornered Snowflake’. Cox explains that Kepler’s wonder at the symmetry of snowflakes, written during a Prague snowstorm, embodies the modern scientific quest to understand the origins of patterns we observe.He says the show explores three realms: what we know, what we don’t know, and what may remain unknowable. This framework sets the stage for his reflections on the biggest unanswered questions in science.If he could answer any scientific mystery, Cox would choose the existence of life beyond Earth. He points to current missions—two spacecraft heading to Jupiter’s moons and the James Webb Space Telescope probing exoplanet atmospheres—as promising avenues that might soon reveal a biosignature.When asked about the future of technology, Cox warns that the trajectory of artificial intelligence is still opaque, making it both thrilling and potentially hazardous. He adds that quantum computing is another frontier with wildly divergent timelines—some experts see breakthroughs within five years, others doubt they’ll arrive in a lifetime—highlighting the revolutionary yet unpredictable nature of these tools.Cox also reflects on his shifting stance toward social media. He admits moving from early optimism about its democratizing power to a more critical view of its noise and misinformation, while acknowledging its complex impact on politics and public discourse.On personal advice, Cox stresses the importance of following one’s passion: “Do what you most enjoy,” he says, recalling his own journey from a teenage musician to a celebrated astronomer, supported by parents who encouraged his interests.He shares a memorable fan encounter: at an ELO concert, a man introduced himself as Paul McCartney after asking a question about Saturn’s moon Enceladus—an encounter that left Cox “overwhelmed” as a lifelong Beatles fan.When asked about a pop‑culture stance, Cox admits he is largely out of touch, but notes that musical innovation seems to have plateaued since the 1980s, contrasting the rapid evolution of earlier decades.One of his favorite astronomical facts is the staggering number of observable galaxies: about two trillion. He highlights how, within a century, humanity moved from debating whether other galaxies existed to quantifying an immense cosmos, also measuring its age at roughly 13.8 billion years.Brian Cox will tour ‘Emergence’ across Australia from 5 to 30 May 2026.

A 20‑year‑old male was taken into custody by the San Francisco Police Department (SFPD) after a Molotov cocktail was hurled at the North Beach residence of OpenAI chief executive Sam Altman in the early hours of Friday, around 4 a.m. local time (11:00 GMT). The incendiary device ignited part of an exterior gate before the suspect fled on foot.Police confirmed that the suspect was located about an hour later, approximately 4.8 kilometres (three miles) from the scene, near OpenAI’s headquarters, where he allegedly threatened to set the building ablaze. No injuries were reported.OpenAI released a statement thanking the SFPD for its rapid response and emphasizing that the company is fully cooperating with investigators. “Thankfully, no one was hurt,” the spokesperson said, adding that the incident highlights the need for continued vigilance in protecting employees.The attack arrives amid a wave of security concerns targeting OpenAI’s facilities. In November, a separate threat prompted a temporary lockdown of the company’s San Francisco office, and activists have increasingly singled out Altman and OpenAI for the perceived risks associated with advanced artificial intelligence.Critics also point to OpenAI’s recent partnership with the U.S. Department of Defense, arguing that the collaboration intensifies scrutiny of the firm’s role in military technology. Public sentiment toward AI remains divided; a recent NBC News poll indicated that AI is viewed less favorably than the U.S. Immigration and Customs Enforcement agency.Despite the controversy, OpenAI’s growth trajectory remains robust. The company announced a valuation of $852 billion following a funding round that raised $122 billion. Its flagship product, ChatGPT, continues to dominate the consumer AI market with over 900 million weekly active users and roughly 50 million subscribers. Usage of OpenAI’s search features has also tripled over the past year.While the motive behind the Molotov attack remains unclear, the incident underscores the escalating security challenges faced by leading AI firms operating at the intersection of technology, public policy, and national defense.