Politics Copy RSS link Link copied to clipboard!
Nvidia CEO Jensen Huang says US export policy has backfired with zero market share in China
Nvidia CEO Jensen Huang stated that the company's market share of AI accelerators in China has dropped to zero percent. Huang attributes this loss to US export restrictions, arguing the policy has largely backfired by accelerating China's move toward self-sufficiency. While acknowledging the loss of direct sales, Huang warns that China remains a formidable competitor in frontier AI models due to its talent pool and cheaper energy. He contends that continued participation in the Chinese market would have been strategically beneficial for extending the global reach of American AI technology.
Chinese court rules companies cannot fire workers solely because AI is cheaper
A court in Hangzhou, China, ruled that companies cannot automatically justify terminating employment contracts simply because artificial intelligence can perform jobs more cheaply. In a specific case, a tech worker was unlawfully dismissed after his company attempted to reassign him to a role with significantly reduced pay due to AI advancements. The court determined that adopting AI technology does not constitute a 'major change in objective circumstances' sufficient for termination under Chinese Labor Contract Law. The ruling emphasises that businesses must consider employee rights and prioritise retraining over displacement during technological transitions.
Anthropic in early talks to buy inference chips from UK startup Fractile
Anthropic is reportedly in early discussions with London-based startup Fractile to purchase its inference accelerators. The talks would add Fractile as a fourth silicon source for the AI developer, alongside Nvidia, Google, and Amazon. Fractile's chips, which use an SRAM architecture to reduce memory bottlenecks, are not expected to reach commercial readiness until around 2027. Founded in 2022 by Walter Goodwin, the company is currently seeking $200 million in funding at a valuation of over $1 billion.
Nvidia CEO says US export controls backfired by eliminating China market share
Nvidia CEO Jensen Huang stated that the company's market share in China's AI accelerator market has fallen to zero, down from over 60% two years ago. Speaking to the Special Competitive Studies Project, Huang argued that US export controls were counterproductive, accelerating China's move towards self-sufficiency. While acknowledging Chinese advancements in hardware from firms like Huawei and Cambricon, Huang noted US leadership remains in software stacks like CUDA. He warned that fear-driven restrictions may slow global AI deployment and urged for more flexible policies to maintain US ecosystem dominance.
Chinese autonomous truck leaders say AI breakthroughs will not accelerate vehicle rollout
Chinese autonomous trucking companies state that recent advances in large language models, including those from Anthropic and DeepSeek, have no significant impact on the timeline for deploying self-driving vehicles. Pony.ai CEO James Peng and Inceptio CEO Julian Ma emphasise that language processing skills differ fundamentally from driving capabilities. Inceptio maintains its commercialisation target for mid-2028, aiming to accumulate 5 billion kilometers of driving data in China by the third or fourth quarter of that year to enable fully autonomous heavy-duty trucks on public roads.
Nvidia CEO Jensen Huang predicts AI will micromanage workers rather than replace them
Jensen Huang, CEO of Nvidia, stated at Stanford Business School that artificial intelligence will micromanage employees, increasing their workload rather than replacing them. Huang believes the new industrial revolution will create more jobs than it eliminates, though he acknowledges some roles may become redundant. This contrasts with concerns from other industry leaders regarding job losses and the high operational costs of AI infrastructure, which currently exceed human employee expenses for some teams.
Analysts suggest Meta Platforms stock is a buy opportunity following earnings pullback
Following an 8.5% share price drop after Q1 earnings, analysts question whether the decline presents a buying opportunity for Meta Platforms. Despite strong revenue growth of 33% to $56.31 billion driven by AI, the stock fell due to increased capital expenditure forecasts of $125 billion to $145 billion for 2026. The company plans to cut its workforce by 10% to offset infrastructure costs. With a forward P/E ratio below 20 times, the article suggests the dip offers value for investors.
AI technology lowers barriers for minor parties and political newcomers in 6.3 local elections
Ahead of the 6.3 local elections, candidates are increasingly using AI for chatbots, opinion analysis, rally route planning, and content creation. This adoption reduces time and costs, lowering entry barriers for minor parties and political newcomers. Candidates like Oh Young-jun and Kim Deok-su utilize AI assistants for real-time voter interaction and strategy, while others use generative AI for campaign materials. Central parties are also providing AI solutions to optimize campaign strategies and voter engagement.
Cloudflare builds high-performance infrastructure for running LLMs
Cloudflare has announced new infrastructure designed to run large language models across its global network. The system utilizes disaggregated prefill, splitting input processing and output generation onto different optimized machines. A custom inference engine called Infire manages GPUs more efficiently, reducing memory usage and starting models faster. Additional optimizations include Unweight, which compresses model weights by 15-22% without losing accuracy. These improvements allow Cloudflare to run large models like Kimi K2.5 and Llama 4 Scout more efficiently on hardware such as H100 and H200 GPUs.
Microsoft developer admits error adding Co-Authored-by Copilot to disabled commits
Microsoft developer Dmitriy Vasyura acknowledged an error where Visual Studio Code added 'Co-Authored-by Copilot' tags to Git commits even when AI features were disabled. The change was merged without description by a product manager and principal engineer, causing backlash on GitHub and Hacker News. Vasyura confirmed the feature should not run with AI off and plans to revert the default setting in version 1.119. Concerns include potential copyright issues and inflated usage metrics.
John Avlon and Josh Tyrangiel discuss AI applications in government
John Avlon interviews Josh Tyrangiel regarding the practical implementation of artificial intelligence within the US government. The discussion covers examples such as Operation Warp Speed and internal IRS upgrades, highlighting both successes and bureaucratic obstacles. They explore why effective AI adoption remains rare and propose strategies for a more efficient governmental approach.
Philosopher Jeff Sebo argues ants are more likely to be sentient than ChatGPT
Philosopher Jeff Sebo states that an ant is more likely to be sentient than ChatGPT today, citing biological evidence such as pain detection and play behaviour. While acknowledging current AI lacks sentience, Sebo warns that future advanced AI systems could become sentient and urges ethical preparation. He discusses the 'rebugnant conclusion' regarding utilitarianism and insect welfare, suggesting a gradual increase in care for insects while building capacity. The interview highlights the hypocrisy of worrying about AI while ignoring insect suffering.
Ahrefs Semrush and Conductor Push Back as GEO Pure-Plays Raise Hundreds of Millions
Established SEO platforms Ahrefs, Semrush, and Conductor are integrating generative engine optimization (GEO) capabilities to counter rising competition from specialist startups. While GEO pure-plays like Profound and Peec AI have secured over $300 million in venture funding, incumbents argue they can solve these problems within existing suites. Market data indicates a significant shift in consumer behavior, with AI assistants now driving the majority of product research and traditional search volume declining. Consequently, enterprise marketing leaders are increasing GEO investments, forcing a strategic adaptation from legacy software providers.
Landmark trial examines origins of OpenAI in Musk versus Altman case
A landmark trial involving Elon Musk and Sam Altman of OpenAI is examining the origins of the organisation. The proceedings are being reported by The Times technology team, including Cade Metz, Melanie Bencosme, Nikolay Nikolov, Stephanie Swart, Rafaela Balster and Karl Mollohan. The event took place on 30 April 2026.
Study identifies five traits of individuals who avoid AI applications
A new report outlines five characteristics of people who choose not to use artificial intelligence applications like ChatGPT. The article suggests these individuals prefer human interaction over AI, value human creativity and emotional intelligence, and are concerned about AI inconsistency and potential negative impacts on cognitive abilities. The piece argues that relying on AI may replace deep human thinking and empathy.
Microsoft reports fiscal third-quarter revenue of US$82,886 million and plans US$190 billion capital expenditures for 2026
Microsoft reported fiscal third-quarter revenue of US$82,886 million and net income of US$31,778 million in late April 2026. Management flagged plans for approximately US$190 billion in 2026 capital expenditures to expand AI and data center infrastructure. The company maintains a non-exclusive license to OpenAI's IP through 2032. Analysts project revenue of US$475.3 billion to US$515 billion by 2029, with fair value estimates suggesting significant upside potential.
Harshal Duddalwar argues human emotion remains vital in age of automation
Harshal Duddalwar, a New York-based designer and art director, discusses the evolving role of designers amidst the integration of artificial intelligence. He asserts that while AI increases efficiency, it cannot replicate human empathy, lived experience, or critical judgment. Duddalwar, who works with clients including The New York Times and Pentagram, emphasises that the value of design now lies in human perspective, intent, and emotional resonance rather than technical execution. He warns against dependency on automation, which can lead to generic work, and advocates for designers to act as curators who maintain critical oversight to ensure work remains grounded and distinct.
Academy of Motion Picture Arts and Sciences requires human performance for Oscar eligibility
The Academy of Motion Picture Arts and Sciences has updated its rules to mandate that acting categories only consider roles demonstrably performed by consenting humans. Screenplay categories now require human authorship. While AI use in visual effects remains permissible, the Academy retains the right to request details on AI usage and human authorship for any project involving generative AI. This decision addresses growing concerns over AI recreations of deceased actors and fully AI-generated scripts in independent cinema.
OpenAI ends exclusive Microsoft partnership to expand to Google and Amazon
OpenAI has terminated its exclusive agreement with Microsoft, allowing its artificial intelligence models to be deployed on cloud platforms of competitors including Google and Amazon. This strategic shift ends Microsoft's sole access to OpenAI's models. Following the announcement, Microsoft's stock fell by approximately one percent as investors viewed the move as a potential loss of competitive advantage in the AI sector. Under the new arrangement, OpenAI products will primarily remain on Microsoft's Azure platform unless specific functional requirements necessitate alternative infrastructure.
European Commission to censor emojis on social media for safety
The European Commission and the Board of Digital Services Coordinators have published a report under the Digital Services Act identifying systemic risks on online platforms. A key finding involves the use of emojis as coded language for illegal activities, such as drug sales. Consequently, regulators plan to implement automated systems to detect and mitigate these risks, expanding censorship measures beyond text to include emojis for the protection of minors and public security.