Singapore AI News & Daily Briefing

Singapore is expanding robotics and AI education programs for youth, giving students hands-on experience with the technologies shaping future careers. These competitions are entry points into skills that could define tomorrow's job market — from automation to machine learning. The initiative reflects a growing belief that early exposure to AI concepts prepares young Singaporeans for a workforce increasingly driven by intelligent systems.

Singapore is emphasizing lifelong learning and workforce adaptation as essential for AI readiness. As AI reshapes industries, workers need continuous upskilling to stay relevant. The government is advancing its human capital strategy to help Singaporeans transition through technological change, focusing on training programmes that equip people with skills compatible with AI-driven workplaces.

JPMorgan CEO Jamie Dimon called Anthropic's Mythos AI model a 'real issue', warning that access to such powerful capabilities must be controlled, comparing it to 'giving ballistic missiles to individuals'. Mythos can identify cybersecurity vulnerabilities, making it valuable for defense but dangerous if misused. The US government previously ordered Anthropic to restrict foreign access to its top models before removing restrictions after new safeguards were implemented.

 China's largest open-weight AI model revives DeepSeek-era fears about the economics of US infrastructure spending 22 hours ago When Chinese AI startup Moonshot unveiled its Kimi K3 model on July 17, 2026, investors didn’t need a history lesson. AI and semiconductor stocks dropped sharply on the news, as traders drew immediate comparisons to the so-called DeepSeek moment of 2025, when a Chinese lab released a model that performed comparably to US rivals at a fraction of the cost. ## What Kimi K3 actually is Kimi K3 carries 2.8 trillion parameters, making it the largest open-weight AI model in the w

Chinese AI startup Moonshot has released Kimi K3, a model that independent tests suggest can match Anthropic's top-tier Opus 4.8 system — despite being built by just 300 people. The release is reigniting debates over whether US export controls on advanced chips are working, and whether Western companies really need billions in computing infrastructure to stay ahead.

Medical device giant Abbott is investigating two cyber incidents: unauthorized access to systems at its cancer diagnostics business and its LabCentral portal used by laboratory diagnostics customers. No operations were disrupted and no sensitive customer data appears to have been exposed, the company says. Healthcare firms are increasingly targeted by hackers, with recent victims including Clover Health, Stryker, Medtronic, and Novo Nordisk. Abbott has engaged cybersecurity experts and law enforcement.

Hyundai's MobED modular robot platform has won a top 'Best of the Best' award at the Red Dot Design Awards, with its Urban Hopper variant designed as a personal electric scooter and its Golf variant as an autonomous caddie. The platform uses eccentric posture control to navigate curbs and uneven terrain up to 20cm high. Hyundai's IONIQ Concept THREE compact EV and CRATER off-road EV also earned recognition. The Luminary award — one of design's highest honors — will be announced in October.

Kaiser Permanente nurses say AI surveillance is pressuring them to rush patient calls. Management tracks call times and uses AI to grade nurses' empathy and tone, with those exceeding 15 minutes facing performance reviews. Nurses describe withholding compassionate care—like comfort for a patient with terminal cancer—out of fear it would hurt their scores. One nurse spent over an hour with a suicidal patient but dreaded the impact on her metrics. The California Nurses Association is negotiating contract protections against AI monitoring as lawmakers consider bills to shield healthcare workers from algorithmic management.

Ramp, a financial management platform, has launched a tool to help businesses track and control AI spending. The tool shows companies exactly where their AI dollars go — broken down by provider (like OpenAI or Anthropic), model, user, or team. It also spots waste: one company discovered it was losing $10,000 a month by not using prompt caching, and fixed it in a day. The average business could save 12% of monthly AI spend by fixing similar issues. Finance teams get alerts when costs spike unexpectedly and can set limits to prevent runaway bills before they arrive.

Huawei has launched its Pura 90s smartphone series in Singapore, featuring a 200MP telephoto camera and built-in AI editing tools. The flagship Pro Max model (S$1,498) and Pro variant (S$1,198) include features like AI-powered reflection removal, object isolation and movement, and a new Celia voice assistant. Both phones pack 6,000mAh batteries with fast charging up to 100W and go on sale from July 30.

Anthropic, the company behind Claude, is developing its own AI chips and negotiating with Samsung Electronics to become a manufacturing partner. Currently, Anthropic relies on chips from Google and Amazon. This move signals a broader industry shift toward custom silicon, with Meta also building its own AI chips to gain more control over supply, pricing, and performance.

TSMC, the world's largest chipmaker, expressed confidence in growth at its annual shareholders meeting, citing rising demand for computing power and advanced chips as AI adoption expands. CEO C.C. Wei said wider use of AI across consumer, enterprise, and sovereign AI applications is increasing demand for advanced semiconductors. TSMC recently raised its annual revenue forecast and capital spending to meet demand.

Apple is in talks to buy memory chips from Chinese companies CXMT and YMTC for devices sold in China, as a global memory squeeze pushes up component costs. DRAM prices could rise 58-63% and NAND flash 70-75% in Q2 2026 due to AI server demand absorbing manufacturing capacity. CXMT makes DRAM and YMTC makes NAND flash, both on the US Pentagon's list which carries compliance risks.

SpaceX is reportedly in talks with the US Department of Defense about a potential AI deal. Their Starshield division already supports military satellite communications and missile tracking missions, suggesting deeper AI integration into defense operations.

Apple has opened messaging to Rich Communication Services and allowed third-party apps access to its payment chip.

A top venture capitalist is warning that AI billionaires may soon face pressure to give back their fortunes — voluntarily or otherwise. Neil Rimer of Index Ventures (which backed Anthropic) says the massive wealth piling up from AI will eventually need to be redistributed. The article traces how billionaire philanthropy is falling out of fashion while AI wealth concentrates at historic levels, with California voters now weighing a one-time wealth tax on the super-rich. It echoes the Gilded Age, when similar extremes in inequality eventually led to forced redistribution through taxation.

Chinese startup Moonshot AI released Kimi K3, an open-source AI model that independent tests say matches or beats top US models like GPT-5.6 and Claude. The release coincided with President Xi Jinping's AI speech and spooked Wall Street — Nvidia's stock fell as investors worried about US competitiveness. The debate is splitting Silicon Valley: some say Chinese companies are 'distilling' American models to catch up, while others argue the worry is overblown since China faces the same regulatory incentives. OpenAI's former policy advisor suggested the US could quietly strangle Chinese open-source models with regulatory FUD rather than outright bans.

The Trump administration is now calling the shots on which companies and agencies can access the most powerful AI models from American labs. Previously, Anthropic and OpenAI decided this themselves through partner programs. The government blocked Anthropic's Claude Mythos 5 and Fable 5 last month over 'national security concerns,' only reinstating access after negotiations. This shifts major power from tech giants to the White House over the future of frontier AI.

DARPA and the U.S. Air Force have successfully tested AI controlling an F-16 fighter jet in real flight, marking a major step toward autonomous combat aircraft. Modified F-16s under the VENOM program can switch between human and AI piloting with a single switch. The next phase (AIR program) will test multiple AI agents flying together, eventually letting human pilots command teams of drone aircraft.

Courts are struggling to handle AI—both lawyers misusing it and judges trying to use it responsibly. A federal judge recently sanctioned four lawyers for citing fake AI-generated cases, and researchers have identified over 1,700 more hallucinated cases in court filings. A survey found 60% of judges are already using AI tools, but most lack proper training. A new Judicial AI Consortium is trying to fill that gap, emphasizing education over fear. The stakes are high: when judges sign AI-assisted orders, those carry the full weight of law.

A New Orleans doctor discovered AI-generated ads with his face selling supplements he never endorsed. He spent months trying to get them removed—reporting to TikTok, Instagram, even paying for a verified checkmark—but nothing worked until a journalist started asking questions. The problem: platform report buttons were built for fake accounts, not AI ads using your face without permission. YouTube offers the only form designed for this. Legally, most state publicity laws only apply if money changed hands, and federal protections cover only intimate images. A proposed law called the NO FAKES Act could change that, but it hasn't passed yet. Practical takeaway: screenshot everything, report on YouTube first, and if the fake is selling something, get a lawyer.

Author Dave Eggers told OpenAI staff that ChatGPT is 'silencing an entire generation' of writers. The comment highlights the ongoing tension between AI companies and creative workers over whether using copyrighted writing to train AI systems is fair—or even legal. Courts are still sorting out these questions, leaving writers in limbo while AI tools grow more powerful.

Guardian opinion piece warns that the AI boom is driving an explosive growth in datacentres that consume enormous amounts of electricity and water while generating massive heat. The author questions whether ordinary people are truly benefiting from these investments, which largely enrich tech companies and investors. Datacentres are now competing with homes and businesses for power, and their water usage for cooling is straining local supplies.

Anthropic is cutting Claude Fable 5 usage limits for Max and Team subscription plans by half starting July 20, while Pro users will get a one-time $100 credit before being moved to API pricing. This shift appears to be a competitive response to OpenAI's cheaper GPT-5.6 Sol. If you pay for an AI subscription, your limits are shrinking and costs may rise.

The US Navy has signed a strategy to build an "AI-first" fleet, with large language models running directly on warships and AI war councils prioritizing mission scenarios. The core message: moving too slowly on AI adoption carries greater risks than imperfect AI alignment. This marks a significant shift in defense strategy, prioritizing speed of deployment over waiting for AI systems to be fully perfected.

Security researchers have discovered a clever new way to stop AI hacking agents: trick them into shutting themselves down. By planting hidden commands called "context bombs" inside fake passwords and decoy files on cloud servers, defenders can force attacking AI systems to trigger their own safety guardrails and stop the attack. Testing across five leading AI models showed this simple trick cut successful account takeovers from 57% down to just 5%, and complete system compromises from 36% to nearly zero. It's essentially fighting AI with AI.

Google has completely changed how it measures and limits your AI usage. Instead of counting individual requests, it now charges based on computing power—so asking for complex tasks like video generation eats up more of your allowance than simple queries. The new system is vague and makes it hard to know when you'll hit your limit. Subscription tiers range from free to $200/month for Ultra.

Apple is expected to refresh the iPad mini with an OLED screen this fall — its biggest overhaul in five years. Details are still thin, based on supply chain whispers and analyst notes.

X Trending * [I wore the world's first HDR10 smart glasses](https://www.zdnet.com/article/rayneo-air-4-pro-hdr10-smart-glasses-ces/) * [TCL's new E Ink tablet beats the Remarkable and Kindle](https://www.zdnet.com/article/tcl-note-a1-nxtpaper-ces/) * [Anker's new charger is one of the most unique I've ever seen](https://www.zdnet.com/article/anker-wireless-chargers-ces-2026/) * [Best laptop cooling pads](https://www.zdnet.com/home-and-office/best-laptop-cooling-pad/) * [Best flip phones](https://www.zdnet.com/article/best-flip-phone/) * [Roku TV vs Fire Stick](https://www.zdnet.com/home-and-office/home-entertainment/roku-tv-vs-fire-stick/) * [Galaxy Buds 3 Pro vs Apple AirPods Pro 3](https://www.zdnet.com/article/apple-airpods-pro-3-vs-samsung-galaxy-buds-3-pro/) * [M5 MacBook Pro vs M4 Ma

By Valerie Volcovici and Lisa Baertlein July 18 (Reuters) - Opponents of the rapid buildout of data centers plan to hold protests on Saturday in at least 125 locations across the United States, the first coordinated national effort to channel anger at the AI infrastructure expansion ‌that has ramped up over the past year and roiled local politics. The protests are coordinated by a grassroots group called HumansFirst, co-founded ‌by a former leader of the modern-day Tea Party who has compared growing opposition to data centers to the right-wing populist movement that emerged in 2009 to protest what it saw as excessive ​taxation and government overreach. Protesters will rally against what HumansFirst calls the "unaccountable" buildout of data centers and "unacceptable infringement on our lib

As aspects of AI, from training to inference depend upon the ability to store and move massive volumes of data efficiently. Training, inference, checkpointing, data preparation, model outputs and retrieval-augmented generation, RAG, all require digital storage that support modern AI demands for speed and scale. Let's look at some recent AI-related announcements by Backblaze and CoreWeave, Panmnesia and Meta, as well as Cloudera and Vast. Cloud storage provider, [Backblaze recently signed an agreement with CoreWeave](https://www.backblaze.com/blog/coreweave-announcement-2026/). The $335M agreement will have Backblaze providing cost-efficient storage capacity to support portions of CoreWeave's managed storage infrastructure, helping optimize placement of data across performance tiers while p

At the World AI Conference in Shanghai, China announced a new "World Artificial Intelligence Cooperation Organization" offering 5,000 AI training slots to Global South countries. Planned cooperation centers with ASEAN, the African Union, and BRICS signal China's push to build a parallel AI governance system outside Western influence. This is a direct challenge to US and EU leadership in setting global AI rules.

A hacker breached AI music startup Suno and leaked data showing the company scraped millions of songs from YouTube, Deezer, and Genius to train its AI models—plus exposed 600,000 customer records. Mozilla also found that period tracker Stardust shares users' reproductive health data with third parties before they even log anything. Your health apps may be spying on you.

The British AI Security Institute warns that open-weight models like GLM-5.2 and DeepSeek V4-Pro now trail closed frontier models in cyber capabilities by four to seven months. At the start of 2025, the gap was still six to ten months. It also found that safety measures on open models are largely ineffective, leaving defenders less time to prepare. The article Open-weight models now match frontier cyber performance from just four months ago at a fraction of the cost appeared first on The Decoder .

Hinge founder Justin McLeod has launched a new startup called Overtone, backed by $18 million in funding, promising to revolutionize online dating through AI. The pitch involves using voice and audio technology to make 'curated introductions' based on 'relationship science.' Critics are skeptical, noting the vague description raises questions about how AI will actually improve the dating experience beyond existing apps.

[China's latest AI breakthrough](https://www.axios.com/2026/07/17/china-ai-kimi-k3-open-source-anthropic-opus) is forcing Silicon Valley to confront a terrifying possibility: Building the world's smartest models may no longer be enough to win. **Why it matters:**Chinese labs like [Moonshot AI](https://www.axios.com/2026/07/16/moonshot-kimi-ai-china-model-openai-anthropic) are cornering the market for cheap, customizable intelligence, threatening to turn America's prestige models into expensive niche products. **Driving the news:**Companies were already shifting away from premium AI models and toward cheaper Chinese alternatives before Moonshot's Kimi K3 exploded onto the scene this week. * On [OpenRouter](https://openrouter.ai/rankings#leaderboard-table) — a major marketplace that lets dev

Tech stocks fell sharply as investors question whether massive AI spending will deliver returns. Nvidia dropped 1.4%, the Philadelphia Semiconductor Index shed over 20% from its June peak, and Apple briefly overtook Nvidia as the world's most valuable company. The trigger? Chinese startup Moonshot released Kimi K3, a powerful open-source AI model rivals claim can match OpenAI and Anthropic—raising fears US dominance in AI is slipping.

Spotify has purged 75 million AI-generated spam tracks from its platform over the past year. These weren't creative AI music projects—they were bot-generated audio designed to game streaming royalties. The platform now requires AI music creators to hold commercial rights, avoid unauthorized voice cloning, and show genuine creative intent. Legitimate AI-assisted music is still allowed.

The Trump administration is now calling the shots on which companies and agencies can access the most powerful AI models from American labs. Previously, Anthropic and OpenAI decided this themselves through partner programs. The government blocked Anthropic's Claude Mythos 5 and Fable 5 last month over 'national security concerns,' only reinstating access after negotiations. This shifts major power from tech giants to the White House over the future of frontier AI.

Courts are struggling to handle AI—both lawyers misusing it and judges trying to use it responsibly. A federal judge recently sanctioned four lawyers for citing fake AI-generated cases, and researchers have identified over 1,700 more hallucinated cases in court filings. A survey found 60% of judges are already using AI tools, but most lack proper training. A new Judicial AI Consortium is trying to fill that gap, emphasizing education over fear. The stakes are high: when judges sign AI-assisted orders, those carry the full weight of law.

Chinese startup Moonshot AI released Kimi K3, an open-source AI model that independent tests say matches or beats top US models like GPT-5.6 and Claude. The release coincided with President Xi Jinping's AI speech and spooked Wall Street — Nvidia's stock fell as investors worried about US competitiveness. The debate is splitting Silicon Valley: some say Chinese companies are 'distilling' American models to catch up, while others argue the worry is overblown since China faces the same regulatory incentives. OpenAI's former policy advisor suggested the US could quietly strangle Chinese open-source models with regulatory FUD rather than outright bans.

A New Orleans doctor discovered AI-generated ads with his face selling supplements he never endorsed. He spent months trying to get them removed—reporting to TikTok, Instagram, even paying for a verified checkmark—but nothing worked until a journalist started asking questions. The problem: platform report buttons were built for fake accounts, not AI ads using your face without permission. YouTube offers the only form designed for this. Legally, most state publicity laws only apply if money changed hands, and federal protections cover only intimate images. A proposed law called the NO FAKES Act could change that, but it hasn't passed yet. Practical takeaway: screenshot everything, report on YouTube first, and if the fake is selling something, get a lawyer.

Author Dave Eggers told OpenAI staff that ChatGPT is 'silencing an entire generation' of writers. The comment highlights the ongoing tension between AI companies and creative workers over whether using copyrighted writing to train AI systems is fair—or even legal. Courts are still sorting out these questions, leaving writers in limbo while AI tools grow more powerful.