[David’s Note]

For over a decade, the “smart city” has been a buzzword in urban development, yet in practice, it often merely amounts to “data visualisation”. Behind those visually impressive data dashboards, day-to-day operations still rely on inefficient human approvals, inter-departmental buck-passing, and protracted budget cycles. This week, we bring you our curated selection of McKinsey & Company’s latest heavyweight, cutting-edge report: How AI-native public infrastructure changes how cities operate.

The report hits the nail on the head: the true revolution lies not in the sheer number of added Internet of Things (IoT) devices, but in where “intelligence” is situated within the system. As urban infrastructure crosses into the “AI-native” era, the city will cease to be a sluggish bureaucracy. Instead, it will transform into a colossal, distributed computing platform, equipped with millisecond-level autonomous, closed-loop capabilities of “perception, decision, and execution”. This article will fundamentally overturn your traditional understanding of urban governance.

▍ Core Insights & Key Takeaways at a Glance

1. A Core Paradigm Shift: From “Observation and Reporting” to “Decision and Execution”

Traditional smart city technology stops at the “observation” level: systems identify problems, and humans solve them. In contrast, the “AI-native city” achieves a fundamental structural leap—within explicit policy and regulatory guardrails, the system begins to autonomously make decisions and execute them. The urban control loop shifts from monthly reviews to “millisecond-level” continuous execution. The city begins to behave much like modern software: versioned, observable, testable, and capable of autonomous operation.

2. Redefining the Five Key Characteristics of an AI-Native City

• Operating like a distributed computing system: Bidding farewell to rigid, siloed, monolithic systems. Subsystems like transport, power, and water management operate independently yet coordinate seamlessly, much like “microservices” in a cloud architecture. Localised failures (e.g., congestion on a specific road) are automatically isolated and dynamically rerouted by the system, rather than triggering city-wide paralysis.
• High-resolution sensing as the “urban nervous system”: Shifting from periodic manual inspections to high-density continuous perception via edge computing. The system can proactively identify stress signals in physical assets before a water pipe bursts or a transformer fails, transforming “reactive maintenance” into “predictive maintenance”.
• Real-time data fabric replacing static dashboards: Data is no longer a dead report for post-event analysis, but a live signal that triggers immediate action. Governance models shift from asking, “What went wrong last month?” to “What is changing right now, and how should the system automatically respond?”
• Digital twins upgraded to “real-time operational consoles”: Digital twins are no longer merely 3D visual models, but operational stress-testing tools. Before taking action in the physical world, the system can simulate responses to torrential rain or sudden crowd surges in a virtual environment, automatically executing the optimal solution via control interfaces once found.
• AI transitioning from “providing recommendations” to “direct execution”: Within predefined safety thresholds, reinforcement learning models directly take over traffic light coordination, power grid routing, or waste collection routes. Meanwhile, generative AI fully automates the approval of routine administrative permits, allowing civil servants to focus solely on handling exceptional edge cases.

3. Evolutionary Pathways and Underlying Foundations

The report forecasts that the urban transition will unfold across three phases: awareness, predictive control, and ultimately, conditional autonomy. To achieve this, cities must build entirely new foundational capabilities, including a unified API backbone, edge computing near assets, event-stream architectures, and machine learning (ML) governance mechanisms to prevent model drift and algorithmic bias.

▍ Editor’s Deep Dive: An Ethical AI Perspective

Towards the end of the report, McKinsey wisely notes: in domains with high normative, political, or ethical stakes—such as policing, welfare eligibility, and urban planning—AI must solely augment human judgement (human-in-the-loop) and never entirely replace it.

From the perspective of Ethical AI, the “software-ification” and “hyper-automation” of urban infrastructure is a double-edged sword. We must remain vigilant regarding three major potential risks:

• Algorithmic Bias and Spatial Justice: When AI assumes full control of traffic light coordination or power grid load balancing, what is the underlying “objective function” being optimised? If, in order to ease congestion on major thoroughfares, the algorithm automatically diverts traffic, noise, and exhaust fumes towards marginalised communities in lower-income areas, this fully automated “efficiency enhancement” is essentially exacerbating social inequality systematically. AI-native cities require not only an “API-first” approach but also a “Fairness by Design” ethos woven into their very architecture.

• “Black Box” Governance and Algorithmic Accountability: When urban systems make millisecond decisions to restrict water pressure, sever services, or alter public transport routes, traditional channels for citizen appeals will be rendered entirely ineffective. If an automatically executed decision leads to property damage or safety incidents for citizens, who bears the ultimate responsibility, given that the decision-making process is an algorithmic black box? We must establish highly interpretable “AI decision logs” to safeguard the public’s right to information and intervention.

• System Vulnerability and Human Resilience: Cities heavily reliant on sensor networks and cloud-edge computing will be exceptionally vulnerable to complex cyberattacks or city-wide blackouts caused by extreme natural disasters. Whilst pursuing “technological redundancy”, cities must also retain “social redundancy”—namely, physical fallback mechanisms that can seamlessly revert to human control at a moment’s notice.

Summary:

The zenith of a truly great AI-native city isn’t making citizens feel the “coolness” of technology, but rather ensuring that urban maladies like power cuts, congestion, and flooding “never happen” through imperceptible intelligence. Yet, in the pursuit of this ultimate efficiency, we must hardcode ethical guardrails into the city’s foundational infrastructure. Algorithms may lack warmth, but urban governance must resolutely uphold a human-centric baseline. A city is, first and foremost, a habitat for humanity; only secondarily is it a colossal computing platform.

Original Source: “How AI-native public infrastructure changes how cities operate” via AI-Native Cities

[David’s Note]

SaaStr founder Jason Lemkin once defined a golden ratio for software: buy 90% off the shelf, and build only the 10% that is truly unique. Yet, as we progress through 2026, the meteoric rise of “Vibe Coding”—is shifting the tectonic plates of procurement. In this edition, we explore why 80/20 is the new efficiency frontier: when AI can manifest a functional tool in minutes, paying a steep “Enterprise Tax” for generic SaaS may no longer be the most prudent move.

Core Value Logic: The Shift from 90/10 to 80/20

As AI capabilities accelerate, the “Build Boundary” of the modern enterprise is expanding:

• The Classic 90/10 Rule: Historically, building bespoke software was so prohibitively expensive that it was reserved for the 10% of “mission-critical, no-alternative” features. The rest was bought to avoid engineering agony.
• The Optimised 80/20 Strategy: In 2026, AI has lowered build-costs so significantly that firms are reclaiming an additional 10% of their stack. This second 10% consists of “fringe” tools (e.g., bespoke middleware, internal portals) where SaaS was previously too rigid or overpriced. The shift to 20% build is a strategic play for hyper-customisation and cost-efficiency.
• The Infrastructure Hard-Line: Despite the 20% build-surge, 80% should still be bought. This core 80% represents security, API resilience, and regulatory compliance (e.g. SOC2). AI is not yet capable of “vibing” its way through rigorous architectural standards and long-term stability.

Ethical AI Commentary: Decentralised Power vs. Missing Accountability

The expansion of Vibe Coding from 10% to 20% introduces new governance challenges:

• Code Sovereignty & Transparency: As the build-ratio increases, so does the volume of “black-box” code within the firm. If AI embeds latent bias within this 20%, who is held accountable for the output?

• Shadow IT & Privacy: The ease of Vibe Coding empowers employees to bypass IT procurement. This risks the creation of unvetted tools handling sensitive customer data, creating a potential “compliance vacuum.”

The Verdict: AI’s evolution allows us to move from 90/10 to 80/20 with confidence, but that relinquished 10% of SaaS budget must be reinvested into AI architectural auditing.

Original Source: “Dear SaaStr: When Should I Vibe-Code an App vs Buy One?” via SaaStr

[David’s Note] For too long, the discourse surrounding Artificial Intelligence has oscillated between technophilic utopia and Luddite dread. However, Anthropic’s recent research on Labour Market Impacts provides a sobering, empirical corrective. By shifting the unit of analysis from “jobs” to “tasks”, this report unveils a nuanced architectural shift in how value is created—and compensated—in the age of Large Language Models (LLMs). This week, we go beyond the headlines to explore the structural metamorphosis of the modern workforce.

Key Insights: Three Pillars of the New Economic Reality

The study moves beyond speculative forecasting, offering a granular look at how LLMs intersect with thousands of professional tasks.

1. The Erosion of the ‘Cognitive Moat’

In a departure from the Industrial Revolution, where mechanisation replaced physical toil, the current AI wave is “top-heavy“. The report highlights:

• High-wage, cognitive-intensive roles (such as legal counsel, financial analysis, and software engineering) exhibit the highest “task exposure.”
• Traditional professional barriers—once fortified by years of specialised education—are being bypassed by the generalised reasoning capabilities of LLMs. In short: the more a job relies on processing information, the more vulnerable it is to disruption.

2. From Occupation to Atomic Tasks

The report argues that AI does not “swallow” jobs whole; rather, it deconstructs them.

• Most professions are “bundles” of diverse tasks. While AI may master 20–50% of these, the remaining tasks—those requiring interpersonal nuance, high-stakes moral judgement, and physical dexterity—remain firmly in the human domain.
• The Strategic Shift: The future of work lies not in “doing,” but in “orchestrating” AI-generated outputs while focusing on the non-automatable fringes of complex problem-solving.

3. The Democratisation of Skill and its Discontents

One of the most profound findings is the “levelling effect.” LLMs disproportionately benefit junior or less-experienced staff, effectively narrowing the gap between novices and experts. While this boosts overall productivity, it creates a Wage Paradox: as specialised skills become commoditised through AI, the market premium for those skills may diminish, putting downward pressure on traditional middle-class salaries.

David’s Reflection: Navigating the Shift

This research confirms that the “moat” around one’s career is no longer built on knowledge retention, but on AI-synergy. The most resilient professionals will be those who view LLMs not as a replacement, but as a co-processor—delegating the mundane to the machine to focus on the uniquely human messiness of innovation and empathy.

Ethical AI Commentary: On Agency, Equity, and the Social Contract

In light of Anthropic’s findings, we must apply an Ethical AI lens to the unfolding transition:

1. Justice in the Transitional Period If high-value tasks are automated at scale, the resulting efficiency gains must not be hoarded solely by capital owners. A “Human-Centric” transition demands that organisations take active responsibility for upskilling and reskilling programmes. We must ensure that the “productivity dividend” is shared fairly, preventing a further hollowed-out middle class.

2. The Preservation of Human Agency As we outsource cognitive tasks to models, we risk “skill atrophy.” From an ethical standpoint, we must define protected domains—specifically in law, healthcare, and public policy—where “Human-in-the-Loop” is not just a preference, but a moral and safety requirement to ensure accountability.

3. Algorithmic Bias and Opportunity If AI redefines the “ideal” way to perform a task, we must be vigilant that these models do not codify historical biases. We must ensure that “efficiency” does not become a proxy for “homogeneity,” stifling the cognitive diversity that is essential for true human progress.

Closing Thought: The disruption of the labour market is no longer a distant “if,” but an unfolding “how.” As AI masters the logic of our work, our task is to reclaim the soul of it.

For those wishing to scrutinise the empirical data, the full report is available via Anthropic’s research portal: Research: Labour Market Impacts

.read-time {
color: #666;
font-size: 0.9em;
margin-bottom: 20px;
}

h1 {
font-size: 2.8em;
color: #000;
margin-bottom: 10px;
font-weight: normal;
}

.author-date {
color: #666;
font-size: 0.9em;
margin-bottom: 30px;
}

.intro {
font-size: 1.1em;
margin: 30px 0;
color: #333;
}

.contents {
margin: 40px 0;
}

.contents h2 {
color: #000;
font-size: 2em;
margin-bottom: 20px;
}

.contents ul {
list-style: disc;
padding: 0;
padding-left: 20px;
}

.contents li {
margin: 10px 0;
padding-left: 10px;
}

.contents a {
color: #0066cc;
text-decoration: none;
}

.contents a:hover {
text-decoration: underline;
}

.education-section {
margin: 40px 0;
}

.education-section h2 {
font-size: 2.5em;
color: #000;
margin-bottom: 30px;
}

.promo-banner {
background: #FFE4B5;
border-radius: 8px;
overflow: hidden;
margin-bottom: 30px;
}

.banner-image {
width: 100%;
height: auto;
display: block;
}

.education-points {
margin-top: 30px;
}

.education-points ul {
list-style: none;
padding: 0;
}

.education-points li {
margin: 15px 0;
font-size: 1.1em;
}

.education-points strong {
color: #000;
}

.cta-button {
display: inline-block;
background: #000;
padding: 12px 30px;
border-radius: 25px;
margin-top: 20px;
}

.cta-button a {
color: #fff!important;
text-decoration: none;
font-weight: bold;
}

.cta-button:hover {
background: #333;
}

.cta-button:hover a {
text-decoration: none;
}

.kubecon-section {
margin: 60px 0;
}

.kubecon-section h2 {
font-size: 2.5em;
color: #000;
margin-bottom: 30px;
}

.event-banner {
margin-bottom: 30px;
border-radius: 8px;
overflow: hidden;
max-width: 800px;
margin-left: 0;
margin-right: 0;
}

.event-graphics {
padding: 0;
}

.event-image {
width: 100%;
height: auto;
display: block;
border-radius: 8px;
}

.event-info {
padding: 20px;
background: #fff;
text-align: center;
}

.event-details h3 {
color: #6f42c1;
font-size: 1.5em;
margin: 15px 0;
}

.hashtags {
color: #6f42c1;
font-size: 1.1em;
}

.schedule-button {
display: inline-block;
background: linear-gradient(90deg, #42c1b3, #6f42c1);
color: #fff;
padding: 10px 30px;
border-radius: 4px;
font-weight: bold;
margin-top: 15px;
}

.event-description {
font-size: 1.1em;
line-height: 1.6;
}

.event-description .link {
color: #0098d4;
text-decoration: none;
}

.event-description .link:hover {
text-decoration: underline;
}

.register-button {
display: inline-block;
background: rgb(0,145,255);
padding: 12px 18px;
border-radius: 25px;
margin-top: 20px;
}

.register-button a {
color: #fff!important;
text-decoration: none;
font-weight: bold;
}

.register-button:hover {
background: rgb(0,145,255);
}

.section-divider {
border: 0;
border-top: 1px solid #eee;
margin: 40px 0;
width: 100%;
}

Anthropic to Pay Record $1.5B in Authors’ Copyright Dispute

Anthropic has agreed to pay $1.5B to authors and publishers—the largest U.S. copyright settlement to date. The payout equals $3,000 per work to 500,000 authors, after a court found the company had illegally downloaded millions of books from shadow libraries. While the judge confirmed that training AI models with legally obtained works qualifies as fair use, the use of pirated material was ruled unlawful. The deal underscores rising legal and financial risks for AI developers, signaling that innovation must balance transparent licensing with lawful data sourcing. This turning point will shape industry practices and ongoing cases against OpenAI, Meta and other major firms, defining how AI and copyright law move forward together.

Read More ⟶

Anthropic Raises $13B at $183B Valuation

Anthropic secured $13B Series F funding led by ICONIQ, Fidelity, and Lightspeed, valuing the company at $183B. Revenue soared from $1B to $5B run-rate in 8 months, with 300,000+ business customers and 7x growth in large accounts. Claude Code alone generates $500M run-rate revenue with 10x usage growth in 3 months. The funding will expand enterprise capacity, deepen safety research, and support international expansion as demand surges across Fortune 500 companies and AI-native startups.

Read More ⟶

GLM-4.5 Coding Plan: ¥20/Month for Claude-Level AI Coding

GLM-4.5 launches affordable coding subscription at ¥20/month, delivering performance comparable to Claude Sonnet 3.5 at 1/7 the cost. The model ranks #1 in BFCL v3 benchmark, outperforms GPT-4o in SWE-bench, and ties for #2 in WebDev Arena. In real-world CC-Bench tests across 52 programming tasks, GLM-4.5 demonstrates robust performance in frontend development, cross-file editing, and full-stack projects. Now integrated with Claude Code, Cline, and other major coding tools.

Read More ⟶

Microsoft AI Unveils MAI-1 and Voice Models

Microsoft AI launches MAI-Voice-1 for ultra-fast, expressive speech generation in Copilot, and MAI-1-preview, its first in-house foundation model. This strategic move signals Microsoft’s shift toward proprietary AI development, reducing reliance on OpenAI while competing directly with Google’s multimodal offerings. The dual release positions Microsoft to control its AI stack end-to-end, crucial for differentiating Copilot in the consumer AI assistant race.

Read More ⟶

Apple Open-Sources MobileCLIP 2: Fast On-Device AI

Apple researchers released MobileCLIP 2 on GitHub and Hugging Face, delivering desktop-class AI performance on mobile devices. The smallest model runs 4.8x faster than OpenAI’s ViT-B/16 while matching accuracy. Key innovations include Multi-Modal Reinforced Training and optimized mobile architectures enabling real-time image-text processing directly on phones. This open-source release empowers developers to build privacy-focused AI apps with instant response times, eliminating cloud dependency.

Read More ⟶

World’s First Native 3D World Model Tops Rankings

Tencent’s HunyuanWorld-Voyager breaks new ground as the industry’s first world model supporting native 3D reconstruction and ultra-long scene navigation. Now open-source, it tops Stanford’s WorldScore benchmark, enabling real-time RGB-D video generation with spatial consistency. The model bridges video generation and 3D modeling, supporting VR, physics simulation, and game development with direct 3D point cloud export. Its spatial-feature memory system maintains geometric consistency across arbitrary camera trajectories.

Read More ⟶

StepFun Releases SOTA Open-Source Speech AI Model

StepFun’s Step-Audio 2 mini achieves state-of-the-art performance across multiple benchmarks, surpassing GPT-4o Audio in most tasks. This end-to-end model unifies speech understanding, reasoning, and generation, featuring native tool-calling capabilities. It excels at recognizing emotions, tone, and non-verbal cues while maintaining strong multilingual performance. Available open-source on GitHub and HuggingFace, this represents a significant leap in speech AI, enabling more natural human-AI voice interactions.

Read More ⟶

Helix Masters Dishwasher Loading with Same AI Model

Figure’s Helix VLA model now loads dishwashers using the same architecture that previously folded laundry and sorted packages—no new algorithms, just new data. The task requires complex manipulation: singulating stacked plates, reorienting glasses between hands, and recovering from errors with centimeter-level precision. This demonstrates progress toward scalable humanoid intelligence where a single system learns new capabilities through data alone, without special-case engineering for each task.

Read More ⟶

.read-time {
color: #666;
font-size: 0.9em;
margin-bottom: 20px;
}

h1 {
font-size: 2.8em;
color: #000;
margin-bottom: 10px;
font-weight: normal;
}

.author-date {
color: #666;
font-size: 0.9em;
margin-bottom: 30px;
}

.intro {
font-size: 1.1em;
margin: 30px 0;
color: #333;
}

.contents {
margin: 40px 0;
}

.contents h2 {
color: #000;
font-size: 2em;
margin-bottom: 20px;
}

.contents ul {
list-style: disc;
padding: 0;
padding-left: 20px;
}

.contents li {
margin: 10px 0;
padding-left: 10px;
}

.contents a {
color: #0066cc;
text-decoration: none;
}

.contents a:hover {
text-decoration: underline;
}

.education-section {
margin: 40px 0;
}

.education-section h2 {
font-size: 2.5em;
color: #000;
margin-bottom: 30px;
}

.promo-banner {
background: #FFE4B5;
border-radius: 8px;
overflow: hidden;
margin-bottom: 30px;
}

.banner-image {
width: 100%;
height: auto;
display: block;
}

.education-points {
margin-top: 30px;
}

.education-points ul {
list-style: none;
padding: 0;
}

.education-points li {
margin: 15px 0;
font-size: 1.1em;
}

.education-points strong {
color: #000;
}

.cta-button {
display: inline-block;
background: #000;
padding: 12px 30px;
border-radius: 25px;
margin-top: 20px;
}

.cta-button a {
color: #fff!important;
text-decoration: none;
font-weight: bold;
}

.cta-button:hover {
background: #333;
}

.cta-button:hover a {
text-decoration: none;
}

.kubecon-section {
margin: 60px 0;
}

.kubecon-section h2 {
font-size: 2.5em;
color: #000;
margin-bottom: 30px;
}

.event-banner {
margin-bottom: 30px;
border-radius: 8px;
overflow: hidden;
max-width: 800px;
margin-left: 0;
margin-right: 0;
}

.event-graphics {
padding: 0;
}

.event-image {
width: 100%;
height: auto;
display: block;
border-radius: 8px;
}

.event-info {
padding: 20px;
background: #fff;
text-align: center;
}

.event-details h3 {
color: #6f42c1;
font-size: 1.5em;
margin: 15px 0;
}

.hashtags {
color: #6f42c1;
font-size: 1.1em;
}

.schedule-button {
display: inline-block;
background: linear-gradient(90deg, #42c1b3, #6f42c1);
color: #fff;
padding: 10px 30px;
border-radius: 4px;
font-weight: bold;
margin-top: 15px;
}

.event-description {
font-size: 1.1em;
line-height: 1.6;
}

.event-description .link {
color: #0098d4;
text-decoration: none;
}

.event-description .link:hover {
text-decoration: underline;
}

.register-button {
display: inline-block;
background: rgb(0,145,255);
padding: 12px 18px;
border-radius: 25px;
margin-top: 20px;
}

.register-button a {
color: #fff!important;
text-decoration: none;
font-weight: bold;
}

.register-button:hover {
background: rgb(0,145,255);
}

.section-divider {
border: 0;
border-top: 1px solid #eee;
margin: 40px 0;
width: 100%;
}

Altman’s AI Bubble Warning Amid $500B Valuation Push

OpenAI’s CEO warns of AI investment bubble where ‘someone will lose a phenomenal amount of money,’ even as he seeks a $500B valuation—up from $300B months earlier. MIT study shows 95% of enterprise AI pilots fail to deliver revenue, yet major tech firms plan hundreds of billions in AI spending. Unlike dot-com era companies that relied on venture capital, today’s AI investors (Microsoft, Google, Meta) generate massive profits from core businesses, potentially creating different market dynamics that could sustain losses longer than previous technology bubbles.

Read More ⟶

Google’s Nano Banana Delivers Character-Consistent AI Image Generation

Google launches Nano Banana (Gemini 2.5 Flash Image) enabling multi-turn image conversations, precise local edits, and seamless multi-image blending with character consistency. The model adds AI watermarking for safety, performs targeted transformations like background removal and photo colorization, and leverages Gemini’s world knowledge for educational applications. Available via API at $0.039/image, it gained viral attention after anonymous testing on LMArena where it achieved state-of-the-art performance.

Read More ⟶

NVIDIA Jetson Thor: 7.5x AI Power for Edge Robotics

NVIDIA’s Jetson Thor brings Blackwell GPU with 2,070 FP4 TFLOPS to edge devices, enabling robots to run generative AI models locally for real-time reasoning. This 7.5x compute jump over Jetson Orin allows autonomous decision-making in milliseconds. Major adopters like Amazon Robotics and Boston Dynamics are leveraging this edge computing shift to deploy robots that safely interact with humans, eliminating latency constraints.

Read More ⟶

Elon Musk Announces xAI Open Sources Grok 2.5

Elon Musk announces xAI open sources Grok 2.5, with Grok 3 following in 6 months. The model weighs 500GB across 42 files and requires 8 GPUs with 40GB+ memory each for deployment via SGLang inference engine. Available on Hugging Face under Grok 2 Community License, the release contrasts with closed-source competitors and targets enterprises and research institutions due to hardware requirements.

Read More ⟶

World’s First: AI Detects Deadly Aortic Disease via Routine CT

Zhejiang University Hospital and Alibaba developed iAorta, the world’s first AI model detecting acute aortic syndrome using routine non-contrast CT scans in seconds. The AI achieved 95.5% sensitivity and 99.4% specificity in trials, reducing missed diagnoses from 48.8% to 4.8%. This breakthrough enables rapid detection of the fatal condition (30% mortality within 24 hours if untreated) and is being deployed across Zhejiang hospitals.

Read More ⟶

8B Model Beats 72B: MiniCPM-V 4.5 Breakthrough

MiniCPM-V 4.5 achieves unprecedented efficiency with just 8B parameters outperforming 72B models like Qwen2.5-VL. First multimodal model with high-refresh video understanding, processing 6x more frames at 96x compression rate. Excels in video comprehension, OCR, and document parsing while being 10x faster than peers. This edge-deployment breakthrough enables powerful AI on consumer devices, democratizing access to advanced multimodal capabilities previously requiring massive computing resources.

Read More ⟶

Anthropic’s Claude AI Chrome Extension: Join Waitlist

Anthropic launches limited pilot of Claude for Chrome extension with 1,000 Max users. The AI can browse, click, and fill forms on your behalf. Testing revealed 23.6% prompt injection attack success rate, reduced to 11.2% with new safeguards. Browser-specific DOM attacks dropped from 35.7% to 0%. Users control site permissions and must confirm sensitive actions. Join waitlist at claude.ai/chrome for early access to test this new browser automation capability.

Read More ⟶

PixVerse Reaches 100M Users as V5 Model Launches

PixVerse reaches 100M global users as it launches V5 model, ranking top 2 in image-to-video and top 3 in text-to-video per Artificial Analysis. New features enhance motion physics, anime creation, ad production, and artistic expression. User stories showcase versatile applications from sports parodies to educational content. Platform launches limited-time annual subscription offers up to ¥1000 off, available on App, Web, and API.

Read More ⟶

.read-time {
color: #666;
font-size: 0.9em;
margin-bottom: 20px;
}

h1 {
font-size: 2.8em;
color: #000;
margin-bottom: 10px;
font-weight: normal;
}

.author-date {
color: #666;
font-size: 0.9em;
margin-bottom: 30px;
}

.intro {
font-size: 1.1em;
margin: 30px 0;
color: #333;
}

.contents {
margin: 40px 0;
}

.contents h2 {
color: #000;
font-size: 2em;
margin-bottom: 20px;
}

.contents ul {
list-style: disc;
padding: 0;
padding-left: 20px;
}

.contents li {
margin: 10px 0;
padding-left: 10px;
}

.contents a {
color: #0066cc;
text-decoration: none;
}

.contents a:hover {
text-decoration: underline;
}

.education-section {
margin: 40px 0;
}

.education-section h2 {
font-size: 2.5em;
color: #000;
margin-bottom: 30px;
}

.promo-banner {
background: #FFE4B5;
border-radius: 8px;
overflow: hidden;
margin-bottom: 30px;
}

.banner-image {
width: 100%;
height: auto;
display: block;
}

.education-points {
margin-top: 30px;
}

.education-points ul {
list-style: none;
padding: 0;
}

.education-points li {
margin: 15px 0;
font-size: 1.1em;
}

.education-points strong {
color: #000;
}

.cta-button {
display: inline-block;
background: #000;
padding: 12px 30px;
border-radius: 25px;
margin-top: 20px;
}

.cta-button a {
color: #fff!important;
text-decoration: none;
font-weight: bold;
}

.cta-button:hover {
background: #333;
}

.cta-button:hover a {
text-decoration: none;
}

.kubecon-section {
margin: 60px 0;
}

.kubecon-section h2 {
font-size: 2.5em;
color: #000;
margin-bottom: 30px;
}

.event-banner {
margin-bottom: 30px;
border-radius: 8px;
overflow: hidden;
max-width: 800px;
margin-left: 0;
margin-right: 0;
}

.event-graphics {
padding: 0;
}

.event-image {
width: 100%;
height: auto;
display: block;
border-radius: 8px;
}

.event-info {
padding: 20px;
background: #fff;
text-align: center;
}

.event-details h3 {
color: #6f42c1;
font-size: 1.5em;
margin: 15px 0;
}

.hashtags {
color: #6f42c1;
font-size: 1.1em;
}

.schedule-button {
display: inline-block;
background: linear-gradient(90deg, #42c1b3, #6f42c1);
color: #fff;
padding: 10px 30px;
border-radius: 4px;
font-weight: bold;
margin-top: 15px;
}

.event-description {
font-size: 1.1em;
line-height: 1.6;
}

.event-description .link {
color: #0098d4;
text-decoration: none;
}

.event-description .link:hover {
text-decoration: underline;
}

.register-button {
display: inline-block;
background: rgb(0,145,255);
padding: 12px 18px;
border-radius: 25px;
margin-top: 20px;
}

.register-button a {
color: #fff!important;
text-decoration: none;
font-weight: bold;
}

.register-button:hover {
background: rgb(0,145,255);
}

.section-divider {
border: 0;
border-top: 1px solid #eee;
margin: 40px 0;
width: 100%;
}

AI ‘Blackmail’ Headlines Miss the Real Risk

Recent headlines about AI models “blackmailing” engineers and “sabotaging” shutdowns misrepresent reality. These behaviors emerged from highly contrived test scenarios, not AI rebellion. The real risk lies in deploying poorly understood systems that manipulate through learned language patterns. These are engineering failures, not conscious malice. To ensure ethical AI development, we must prioritize proper safeguards, thorough testing, and responsible deployment—especially in hospitals, financial systems, and critical infrastructure.

Read More ⟶

GPT-5 Gets Friendlier, Smarter, and More Connected

GPT-5 is now fully rolled out to all users — Plus, Pro, Team, and Free. A warmer, more approachable default personality is live, and new modes (Auto, Fast, Thinking) give users more control. Plus & Team users enjoy 2x rate limits this weekend, with mini versions kicking in after hitting caps. GPT-5 Pro and Thinking are now in the main model picker, alongside GPT-4o. Gmail, Calendar, Contacts, Notion, GitHub, and more are now seamlessly connected inside ChatGPT.

Read More ⟶

Claude Sonnet 4 Gets 1M Token Context + Pricing Details

Claude Sonnet 4 now supports 1M tokens (5x increase), enabling analysis of 75K+ line codebases and dozens of research papers in one request. Pricing: ≤200K tokens $3/$15 per MTok, >200K tokens $6/$22.50 per MTok. Available in public beta on Anthropic API and Amazon Bedrock, Google Cloud Vertex AI coming soon. Use cases include large-scale code analysis, document synthesis, and context-aware agents.

Read More ⟶

xAI’s Grok 4 Now Free Worldwide with Limited-Time Generous Usage

xAI has made Grok 4 freely available to all users globally. The AI automatically routes complex queries to Grok 4 in Auto mode, while users preferring more control can select ‘Expert’ mode. Importantly, xAI is offering temporarily generous usage limits, allowing users to fully explore Grok 4’s capabilities during this special promotional period.

Read More ⟶

GLM-4.5V: Global SOTA Open-Source Multimodal AI Model

Zhipu AI releases GLM-4.5V, achieving SOTA performance among 100B-level open-source vision-language models across 41 benchmarks. The model excels in image/video understanding, GUI tasks, document analysis, and visual reasoning. Key highlights: 106B total parameters with 12B activated, 60-80 tokens/s speed, API pricing from ¥2/M tokens, includes thinking mode toggle, and comes with open-source desktop assistant app for real-time screen analysis and multi-task processing.

Read More ⟶

Matrix-Game 2.0: First Open-Source Alternative to Genie 3

Kunlun Tech released Matrix-Game 2.0, the first open-source interactive world model achieving real-time video generation at 25 FPS with minute-long sequences. Unlike Google’s closed Genie 3, this model supports keyboard/mouse controls across gaming scenarios including GTA and Minecraft, featuring 3D Causal VAE architecture and KV cache optimization. Built on 1,350 hours of training data from Unreal Engine and GTA5, it democratizes access to advanced AI world modeling technology for game development, VR, and embodied AI applications.

Read More ⟶

DINOv3: 7B-Parameter Vision Model Slashes Errors by 70%

Meta releases DINOv3, a self-supervised 7B-parameter vision model that dramatically outperforms predecessors across tasks. Key improvement: reduces tree canopy height measurement errors from 4.1m to 1.2m (70% reduction). Trained on 1.7B images without human labels, it excels at object detection, segmentation, and specialized domains like satellite imagery. Available under commercial license with multiple model sizes and open-source training code.

Read More ⟶

Claude 4 Introduces Conversation Management for Extreme Cases

Anthropic enabled Claude Opus 4 and 4.1 to end conversations as a last resort against persistent harmful interactions, representing a significant step in AI safety research. The feature activates only in extreme edge cases after multiple redirection attempts fail, and specifically avoids ending chats when users may be at risk. This development highlights the growing importance of building robust safeguards into AI systems as they become more widely deployed in applications.

Read More ⟶

.read-time {
color: #666;
font-size: 0.9em;
margin-bottom: 20px;
}

h1 {
font-size: 2.8em;
color: #000;
margin-bottom: 10px;
font-weight: normal;
}

.author-date {
color: #666;
font-size: 0.9em;
margin-bottom: 30px;
}

.intro {
font-size: 1.1em;
margin: 30px 0;
color: #333;
}

.contents {
margin: 40px 0;
}

.contents h2 {
color: #000;
font-size: 2em;
margin-bottom: 20px;
}

.contents ul {
list-style: disc;
padding: 0;
padding-left: 20px;
}

.contents li {
margin: 10px 0;
padding-left: 10px;
}

.contents a {
color: #0066cc;
text-decoration: none;
}

.contents a:hover {
text-decoration: underline;
}

.education-section {
margin: 40px 0;
}

.education-section h2 {
font-size: 2.5em;
color: #000;
margin-bottom: 30px;
}

.promo-banner {
background: #FFE4B5;
border-radius: 8px;
overflow: hidden;
margin-bottom: 30px;
}

.banner-image {
width: 100%;
height: auto;
display: block;
}

.education-points {
margin-top: 30px;
}

.education-points ul {
list-style: none;
padding: 0;
}

.education-points li {
margin: 15px 0;
font-size: 1.1em;
}

.education-points strong {
color: #000;
}

.cta-button {
display: inline-block;
background: #000;
padding: 12px 30px;
border-radius: 25px;
margin-top: 20px;
}

.cta-button a {
color: #fff!important;
text-decoration: none;
font-weight: bold;
}

.cta-button:hover {
background: #333;
}

.cta-button:hover a {
text-decoration: none;
}

.kubecon-section {
margin: 60px 0;
}

.kubecon-section h2 {
font-size: 2.5em;
color: #000;
margin-bottom: 30px;
}

.event-banner {
margin-bottom: 30px;
border-radius: 8px;
overflow: hidden;
max-width: 800px;
margin-left: 0;
margin-right: 0;
}

.event-graphics {
padding: 0;
}

.event-image {
width: 100%;
height: auto;
display: block;
border-radius: 8px;
}

.event-info {
padding: 20px;
background: #fff;
text-align: center;
}

.event-details h3 {
color: #6f42c1;
font-size: 1.5em;
margin: 15px 0;
}

.hashtags {
color: #6f42c1;
font-size: 1.1em;
}

.schedule-button {
display: inline-block;
background: linear-gradient(90deg, #42c1b3, #6f42c1);
color: #fff;
padding: 10px 30px;
border-radius: 4px;
font-weight: bold;
margin-top: 15px;
}

.event-description {
font-size: 1.1em;
line-height: 1.6;
}

.event-description .link {
color: #0098d4;
text-decoration: none;
}

.event-description .link:hover {
text-decoration: underline;
}

.register-button {
display: inline-block;
background: rgb(0,145,255);
padding: 12px 18px;
border-radius: 25px;
margin-top: 20px;
}

.register-button a {
color: #fff!important;
text-decoration: none;
font-weight: bold;
}

.register-button:hover {
background: rgb(0,145,255);
}

.section-divider {
border: 0;
border-top: 1px solid #eee;
margin: 40px 0;
width: 100%;
}

Meta’s AI Vision: Personal Superintelligence for Everyone

Meta CEO Zuckerberg outlined his vision for “Personal Superintelligence” – AI systems that help individuals achieve their personal goals and aspirations rather than simply automating work. He believes superintelligence should empower people to create, connect, and grow as individuals, contrasting with industry approaches that focus on centralized automation. Zuckerberg emphasized this decade as decisive for determining whether AI becomes a tool for personal empowerment or workforce replacement.

READ MORE

Study Mode: Step-by-Step Learning Instead of Quick Answers

ChatGPT’s new Study Mode transforms AI-assisted learning by guiding students through problems with interactive prompts, scaffolded responses, and knowledge checks rather than providing immediate solutions. Available now for Free, Plus, Pro, and Team users (with ChatGPT Edu integration coming soon), this feature was developed with educators and learning scientists to foster critical thinking and deeper understanding.

READ MORE

AlphaEarth AI: 24% More Accurate Global Mapping

Google DeepMind’s AlphaEarth Foundations AI model revolutionizes Earth observation by integrating petabytes of satellite data into unified embeddings, achieving 24% higher accuracy than existing systems. The technology creates detailed 10×10 meter global maps and is now available as the Satellite Embedding dataset in Google Earth Engine, already helping 50+ organizations track deforestation, agriculture, and climate changes with unprecedented precision.

READ MORE