OpenAI is making strides in superapp development and AI model advancements, while Anthropic focuses on expanding its text and code tools. Other notable developments include Amazon's acquisition of Fauna Robotics, Google Cloud's improved Identity-Aware Proxy integration, and Cloudflare's upgraded server fleet. Security and privacy concerns surrounding AI coding agents and large language models are also highlighted.
OpenAI is shutting down its Sora AI video app and website following legal challenges.
OpenAI's decision to shut down Sora, its AI video app, marks a rare retreat in the company's aggressive expansion strategy. The shutdown follows legal pushback and the collapse of a $1B deal with Disney, signaling potential challenges in monetizing cutting-edge AI applications. This raises questions about the balance between innovation and regulation in the AI space. As the industry matures, how can companies navigate the increasingly complex legal landscape while delivering groundbreaking products?
Amazon acquired Fauna Robotics, the maker of human-like service robots.
Amazon has acquired Fauna Robotics, a company specializing in human-like service robots, marking a significant expansion of its ambitions beyond e-commerce and cloud computing. This acquisition signals Amazon’s intent to play a larger role in the robotics space, which is increasingly intersecting with AI, automation, and workplace productivity. With Fauna’s technology, Amazon could integrate advanced robotics into logistics, customer service, or even home automation. As the boundaries between digital and physical automation blur, how might this acquisition reshape Amazon’s competitive edge and influence the broader robotics industry?
Anthropic reportedly hit $19 billion in annualized revenue by focusing on text, code, and enterprise tools.
Anthropic has achieved a remarkable milestone, reportedly hitting $19 billion in annualized revenue, driven largely by its focus on text, code, and enterprise tools. This achievement underscores the growing demand for practical, high-reliability AI solutions in business contexts, particularly in sectors where safety and precision are paramount. As Anthropic continues to scale, it sets a benchmark for other AI companies navigating the balance between innovation and monetization. With OpenAI’s recent pivot away from experimental projects, this data point raises an important question: is the future of AI more about refining core capabilities for enterprise adoption, or will we see a resurgence of bold, consumer-facing innovation?
OpenAI is reportedly building a superapp combining ChatGPT, Codex, and browser functionality.
OpenAI is reportedly developing a superapp that consolidates ChatGPT, Codex, and its browser into a single platform, signaling a major shift in how users interact with AI. This move aligns with a broader trend of AI companies aiming to become central hubs for productivity, commerce, and information access. By integrating coding, search, and conversational AI, OpenAI could redefine the boundaries of digital assistance. How do you envision the role of such a superapp in your workflow, and what features would make it indispensable for professionals?
OpenAI reportedly finished a new model codenamed 'Spud,' which Sam Altman described as potentially accelerating the economy.
OpenAI has reportedly completed work on a new model codenamed 'Spud,' with Sam Altman indicating it could 'really accelerate the economy.' While details remain scarce, the codename suggests a focus on scalability, efficiency, or perhaps a breakthrough in reasoning capabilities. Given OpenAI’s pivot toward profitability, such a model could play a key role in powering enterprise tools, agents, or productivity applications. As AI models become increasingly intertwined with economic growth, how might the release of 'Spud' influence industry adoption and investment in AI-driven solutions?
Anthropic released Auto Mode in research preview, enabling Claude to autonomously execute actions with built-in safeguards.
Anthropic has introduced Auto Mode in a research preview for Claude, marking a significant leap toward autonomous AI agents. This feature enables Claude to autonomously execute actions while incorporating built-in safeguards to filter risky behaviors and prompt injections. As AI systems increasingly interact with real-world tools and workflows, this development addresses critical safety concerns while pushing the boundaries of practical AI deployment. The built-in guardrails suggest a more mature approach to agentic AI, balancing autonomy with risk mitigation. How do you see these safeguards shaping the future adoption of autonomous AI in enterprise environments?
OpenAI shifted away from its in-chat checkout feature to prioritize product discovery and merchant-directed purchasing flows.
OpenAI has decided to sunset its in-chat checkout feature in favor of a stronger focus on product discovery and merchant-driven purchasing flows. This shift highlights OpenAI's response to low adoption rates and a clearer understanding of user engagement patterns. For businesses leveraging AI platforms, this signals a broader trend toward ecosystem-driven commerce rather than transactional interactions within chat interfaces. It also underscores the importance of aligning AI features with tangible user needs and monetization pathways. What does this mean for companies building on top of OpenAI's platform in terms of their own product strategies?
Anthropic launched a science blog to formalize AI-driven research workflows.
Anthropic's new science blog isn't just a blog—it's a blueprint for the future of AI-driven research. By formalizing workflows, the company is paving the way for AI to become a systematic partner in scientific discovery. This aligns with the surge in AI's role across industries, where structured iteration and quality assurance are paramount. How do you see AI reshaping the research processes within your field in the next five years?
The agent era will require new distribution solutions as agents need APIs rather than app stores, potentially disrupting traditional platform monetization models.
As the agent era emerges, the traditional app store model faces obsolescence because agents require APIs for interaction rather than centralized app stores. This shift could dismantle Apple's monopolistic payment systems, as users can easily switch platforms if fees are prohibitive. The result may be a more competitive, low-margin payment layer rather than a monopolistic one. For businesses and platforms, this underscores the need to rethink monetization strategies in an agent-driven future. How will your organization adapt its distribution and revenue models to thrive in this evolving landscape?
Claude Code introduces auto mode with classifier-based permissioning for safer autonomous shell commands.
Safety and reliability are at the heart of AI-driven development, and Anthropic's new 'auto mode' in Claude Code is a game-changer. By introducing classifier-based permissioning, the system now allows safe, autonomous execution of shell commands while blocking destructive actions. This innovation directly addresses the growing need for trustworthy AI assistants in critical workflows. As AI tools become more embedded in software development, how can teams ensure both innovation and safety remain in perfect balance?
Ray Data LLM enables 2x throughput over vLLM's synchronous LLM engine at production scale.
Anyscale's Ray Data LLM delivers a 2x throughput improvement over vLLM's synchronous engine, addressing a critical bottleneck in large-scale batch inference for LLMs. Built for modern workloads prioritizing throughput over latency, this library offers scalable execution, fault tolerance, and optimized architecture for production deployments. For companies running inference at scale, this translates to substantial cost savings and improved efficiency. As AI adoption accelerates, infrastructure innovations like this are key to keeping pace with growing demands. How are you optimizing your inference pipelines to balance performance with cost?
Ossature is an open-source harness for spec-driven code generation that validates specifications and generates code with built-in verification.
Ossature introduces an open-source harness for spec-driven code generation, enabling developers to write specifications that Ossature validates, audits for ambiguities, and converts into executable plans. With verification built into the build loop and error-handling agents for repairs, this tool exemplifies the next generation of AI-assisted development. By providing only task-relevant context, Ossature improves efficiency and reduces errors in complex software projects. For engineering teams, this represents a shift toward more structured, reliable AI-driven development. How can your team integrate spec-driven tools like Ossature to improve code quality and reduce debugging time?
Directional updates in RLVR improve reasoning performance by identifying reasoning-critical tokens for test-time extrapolation and training-time reweighting.
RLVR's directional updates have demonstrated significant improvements in reasoning performance by better identifying reasoning-critical tokens. This enables both test-time extrapolation and training-time reweighting, leading to higher accuracy in complex tasks. As AI models tackle increasingly sophisticated problems, techniques that enhance reasoning capabilities are invaluable. For researchers and engineers, this represents a step toward more reliable and interpretable AI systems. How can we further optimize model training to prioritize reasoning over memorization?
TurboQuant is a quantization method from Google that reduces vector memory overhead while preserving performance.
Google's TurboQuant introduces an extreme vector compression technique that dramatically reduces memory overhead while preserving model performance. By improving key-value cache efficiency and accelerating vector search, this method addresses one of the biggest bottlenecks in large-scale AI deployments. In an era where compute costs are skyrocketing, innovations like TurboQuant offer a pathway to more sustainable and scalable AI systems. For infrastructure teams, this represents a critical tool for optimizing resource utilization. How will you incorporate model compression techniques to reduce costs and improve efficiency in your AI workflows?
Semantic calibration emerges as a byproduct of next-token prediction, enabling base models to assess confidence in open-domain QA tasks.
New research reveals that semantic calibration emerges naturally from next-token prediction training, allowing base models to meaningfully assess confidence in open-domain question-answering tasks. This finding suggests that even uncalibrated models possess an inherent ability to gauge certainty, which could revolutionize how we interpret AI outputs. For applications requiring trustworthy AI, this development is a game-changer, offering insights into model reliability without additional training. How can we leverage this emergent property to build more transparent and dependable AI systems?
OpenAI has raised an additional $10 billion, bringing its total funding to over $120 billion, with a targeted $600 billion compute spend by 2030.
OpenAI has secured an additional $10 billion from top-tier investors, pushing its total funding to over $120 billion. With a targeted $600 billion compute spend by 2030 and a renewed focus on profitable initiatives ahead of a potential IPO, this funding round underscores the scale and ambition of the AI industry. For competitors and partners, this signals intensified competition and the need for strategic alignment with OpenAI's ecosystem. It also raises questions about the sustainability of such massive investments. How will the AI landscape evolve as companies like OpenAI scale their infrastructure and ambitions to this magnitude?
A US District Judge questioned the government's ban on Anthropic, suggesting it may be punitive rather than justified.
A US District Judge has raised serious doubts about the government's ban on Anthropic, describing the action as potentially punitive rather than justified. This ruling, part of ongoing efforts to challenge the ban, highlights the tension between national security concerns and fair regulatory practices. For AI companies navigating government restrictions, this case could set a critical precedent. The judge's remarks suggest a need for clearer, more consistent policies governing AI model usage. How should companies and policymakers strike the right balance between security and innovation in AI regulation?
Databricks launched Lakewatch, an AI-powered SIEM platform for threat detection, alongside acquisitions of Antimatter and SiftD.ai.
Databricks has entered the cybersecurity market with Lakewatch, an AI-powered SIEM platform designed for threat detection. Complemented by acquisitions of Antimatter and SiftD.ai to enhance secure agent deployment, this move signals Databricks' ambition to extend its AI leadership into critical infrastructure protection. For enterprises, this development offers a new layer of AI-driven security that integrates with existing data workflows. As cyber threats grow more sophisticated, AI-native security solutions are becoming essential. How can organizations leverage platforms like Lakewatch to stay ahead of evolving threats?
Anthropic's Economic Index Report shows a shift in Claude usage from high-value tasks to lower-wage personal queries.
Anthropic's latest Economic Index Report reveals a notable shift in Claude usage, with a decline in high-value tasks and an increase in lower-wage personal queries. This trend suggests a democratization of AI access, as more users explore personal applications and lighter-use cases. For businesses, this signals both an opportunity to capture new markets and a challenge to maintain engagement with high-value enterprise solutions. How can companies adapt their AI products to serve both professional and personal use cases effectively?
EVA is a framework for evaluating voice agents that assesses complete multi-turn spoken conversations using a realistic bot-to-bot architecture.
ServiceNow's EVA framework introduces a novel approach to evaluating voice agents by simulating realistic multi-turn conversations between bots. This method provides a more accurate assessment of voice AI capabilities than traditional benchmarks, which often fail to capture real-world interactions. For developers building voice interfaces, EVA offers a critical tool for improving robustness and user experience. As voice AI becomes more prevalent, standardized evaluation methods like this are essential for progress. How can we design better evaluation frameworks to ensure voice AI meets the demands of diverse use cases?
Anthropic's Claude can now operate a Mac directly by controlling the screen, clicking, typing, and navigating apps.
Anthropic's latest updates to Claude mark a pivotal shift in AI's role in our daily workflows. The ability to directly control a Mac's interface—clicking, typing, and navigating apps—bridges the gap between AI assistance and tangible productivity. This move underscores the industry's pivot from purely generative tasks to interactive, hands-on automation. With reliability still evolving (~50%), this technology heralds a future where AI doesn't just suggest answers but executes them. How soon do you envision AI agents becoming a seamless, reliable part of your daily toolkit?
OpenAI is in advanced talks with Helion Energy to secure 5GW fusion power by 2030.
OpenAI's move to secure 5GW of fusion power by 2030 isn't just about energy—it's about redefining the future of AI infrastructure. With Sam Altman stepping back from the board to focus on this initiative, the stakes couldn't be higher. Fusion power could unlock unprecedented scalability for AI data centers, addressing both performance and sustainability concerns. As AI continues to consume more resources, how will energy innovations like this shape the next wave of technological progress?
ChatGPT is evolving into a visual shopping discovery platform with product images, prices, and comparisons.
ChatGPT's transformation into a visual shopping discovery platform reflects a strategic shift from pure conversational AI to a multifaceted tool. By integrating product images, prices, and side-by-side comparisons, OpenAI is positioning ChatGPT as a central hub for commerce discovery. This move blurs the lines between search, shopping, and AI agents, creating new opportunities for brands and consumers alike. How will this evolution change the way we discover and interact with products online?
Luma Labs' Uni-1 model combines image understanding and generation, outperforming Google's Nano Banana in tests.
Luma Labs' Uni-1 model is making waves in the AI image generation space by combining understanding and generation in a single autoregressive framework. With performance gains over Google's Nano Banana and GPT Image 1.5, Uni-1 is carving out a niche in multi-image reasoning and editing. This innovation highlights the growing importance of multimodal models in creative and commercial applications. How will these advancements impact the balance of power in the AI-generated content ecosystem?
Cursor shares technical details of Composer 2 MoE model, focusing on pretraining, reinforcement learning, and benchmark development.
Cursor's technical report on Composer 2 offers a rare glimpse into the rigorous process behind building state-of-the-art AI models. By emphasizing pretraining, reinforcement learning, and benchmark development, the team demonstrates how practical coding tasks can drive model excellence. This approach underscores the shift from theoretical benchmarks to real-world applicability in AI. How can more companies adopt similar methodologies to bridge the gap between AI research and tangible engineering outcomes?
Google merges Wiz, Gemini agents, and dark web AI to enhance threat detection with 22-second attack responses.
Google's integration of Wiz, Gemini agents, and dark web AI into a unified threat detection system is a masterclass in AI-driven cybersecurity. By achieving 22-second response times to attacks, the platform sets a new benchmark for real-time defense. This innovation highlights the critical role AI plays in safeguarding digital infrastructure. How can organizations balance the need for rapid AI-driven responses with the imperative of maintaining robust security frameworks?
Meta partners with Arm to develop AI-optimized data center CPUs.
Meta's partnership with Arm to build AI-optimized data center CPUs underscores the industry's focus on hardware innovation as a critical enabler of AI advancements. This collaboration aims to create a new class of silicon tailored for AI workloads, potentially reducing costs and improving performance. As AI continues to push the limits of computational power, how will these hardware innovations reshape the competitive landscape for tech giants?
Google Cloud Looker's directory deletion API contained a validation flaw allowing Remote Command Execution (RCE) via path traversal.
A critical validation flaw in Google Cloud Looker's directory deletion API has been discovered, enabling attackers to achieve Remote Command Execution (RCE) through path traversal. By exploiting a race condition in Ruby's `FileUtils.rm_rf` and pre-placing malicious Git configs, threat actors could bypass `.git` protections and trigger full repository deletion. The post-exploitation phase revealed overpermissioned Kubernetes service accounts, allowing privilege escalation across Looker instances in the same cluster. This incident underscores the risks of insufficient input validation in cloud APIs and the cascading impact of identity-based exploits. How can organizations better balance agility with rigorous validation in their cloud deployments?
LAPSUS$ claimed responsibility for an alleged AstraZeneca data breach, offering a 3GB archive of internal data for sale.
LAPSUS$ has claimed responsibility for an alleged breach of AstraZeneca's systems, advertising a 3GB archive containing source code, infrastructure files, and internal SQL scripts. The claimed data includes Java, Angular, and Python code, as well as GitHub Enterprise user records and Terraform configurations. While the breach's legitimacy remains unverified, the scale and scope of the claimed data—spanning 1,486 directories and 5,892 files—underscore the persistent threat to intellectual property and operational integrity. How should organizations balance transparency with risk management in the event of a potential breach?
A cyberattack on Intoxalock disrupted court-mandated ignition interlock devices, preventing cars from starting in multiple states.
A cyberattack on Intoxalock, a provider of court-mandated ignition interlock devices, disrupted backend systems used by DUI-convicted drivers to calibrate their devices. The outage, which persisted for over a week, left an estimated 7-10% of users in some states unable to start their vehicles. This incident illustrates the far-reaching consequences of cyberattacks on critical infrastructure, particularly when they intersect with public safety and regulatory compliance. How can organizations in regulated industries better prepare for and respond to such disruptions?
Crunchyroll is investigating a breach after a threat actor compromised an Okta SSO account and exfiltrated 8 million Zendesk support tickets.
Crunchyroll is investigating a breach after a threat actor compromised an Okta Single Sign-On (SSO) account belonging to a third-party support agent. The attacker exfiltrated 8 million Zendesk support tickets containing names, email addresses, IP addresses, and limited payment data, then demanded a $5 million ransom. This incident underscores the vulnerabilities inherent in third-party access and the critical importance of robust identity and access management (IAM) practices. What strategies are you implementing to secure third-party access and mitigate supply chain risks?
QualDerm Partners disclosed a data breach impacting 3.1 million individuals after unauthorized network access in December 2025.
QualDerm Partners has disclosed a data breach impacting 3.1 million individuals, revealing unauthorized network access that occurred over two days in December 2025. While details remain limited, this incident highlights the persistent threat to healthcare data and the need for robust monitoring and incident response capabilities. The long-term impact on patient trust and regulatory scrutiny underscores the critical role of proactive security measures. How can organizations in the healthcare sector better align security practices with patient privacy expectations?
Mazda confirmed a data breach exposing employee and partner data due to unauthorized access to a warehouse management system.
Mazda has confirmed a data breach impacting employee and partner data, stemming from unauthorized access to a warehouse management system used for parts sourced from Thailand. The breach, which occurred in December 2025, highlights the vulnerabilities in supply chain systems and the potential for cascading risks across global operations. For an industry increasingly reliant on digital integration, this incident serves as a reminder of the need for comprehensive security controls beyond traditional IT environments. How can manufacturers better secure their operational technology (OT) and supply chain systems against such threats?
The FCC announced a ban on consumer routers manufactured outside the US, citing national security concerns.
The FCC has announced a ban on all consumer routers manufactured outside the US, citing national security concerns and requiring 'conditional approval' for continued importation or sales. This policy shift reflects growing scrutiny of global supply chains and the potential risks of foreign-manufactured networking equipment. For organizations reliant on consumer-grade hardware, this ban may necessitate rapid adjustments to procurement strategies and supply chain dependencies. How will this policy influence your organization's approach to hardware sourcing and vendor risk management?
Apple announced Apple Business, a platform offering free MDM, hosted email, identity management, and app distribution for SMBs.
Apple has just dropped a strategic bombshell with the launch of Apple Business, a unified platform that bundles free MDM, hosted email, identity management, and app distribution for SMBs. This isn't just another feature update—it's Apple positioning itself as a serious player in the enterprise IT stack, going head-to-head with Google Workspace and Microsoft 365. By offering these core IT services for free, Apple is likely aiming to own the device layer first, then pull the rest of the enterprise stack along with it. The email space, in particular, is a critical battleground for identity, access, and workflows. How do you think this will reshape the enterprise software landscape for smaller businesses in the next 12-18 months?
Salesforce SVP outlined 8 design principles for agentic AI that works in production environments.
Salesforce SVP Shibani Ahuja has laid out a compelling argument: many AI deployments fail not because of the models themselves, but because of inadequate architecture. In a new set of 8 design principles for agentic AI, she emphasizes governance, modularity, and identity management for agents from day one. This is a wake-up call for organizations rushing to deploy AI agents on top of legacy systems. Agentic AI isn't just another layer—it requires a fundamental rethinking of how we design and secure our systems. How are you ensuring your AI agents are built on a foundation that can scale securely and transparently?
Apple's hosted email offering is framed as a strategic move into the enterprise stack.
Apple's push into hosted business email isn't just about expanding its product suite—it's a calculated entry into the heart of the enterprise stack. Email remains the anchor for identity, access, and workflows, meaning whoever owns it gains leverage over the entire environment. This move signals Apple's intent to compete not just in devices or apps, but in the foundational infrastructure that powers modern businesses. With Microsoft 365 and Google Workspace as incumbents, Apple is betting on integration and seamless user experience to win over SMBs. What implications does this have for your organization's email and collaboration strategy in the long term?
CrowdStrike expanded its Falcon platform into a unified data and AI security platform.
CrowdStrike has taken a bold step forward with the expansion of its Falcon platform into a unified data and AI security solution. By consolidating identity, endpoint, SaaS, cloud, and AI data protection into a single architecture, CrowdStrike is moving toward a vision where one platform owns data security across every layer. The addition of AI-driven classification and automated enforcement represents a significant leap in how organizations can protect their expanding attack surfaces. In an era where data is scattered across cloud, AI tools, and SaaS apps, this unified approach could redefine enterprise security. How are you consolidating visibility and control over your organization's security posture in the face of growing complexity?
Cyera launched Browser Shield, Data Lineage, and MCP capabilities for AI security and data governance.
Cyera has just introduced three game-changing capabilities designed to bring order to the chaos of AI data usage: Browser Shield, which blocks sensitive data from being sent to AI tools like ChatGPT; Data Lineage, which tracks how data moves after access; and MCP, which lets teams build security agents on top of that data. This is a major leap forward for enterprises struggling to see and control what data goes into AI tools and what happens to it afterward. With AI adoption accelerating, having visibility and control over data flows is no longer optional. Are you confident that your organization can trace and govern every piece of data entering or leaving your AI systems?
Armis won recognition for its cyber exposure management platform at RSAC 2026.
Armis has been recognized at RSAC 2026 for its cyber exposure management platform, Centrix, which addresses a critical gap in asset visibility across IT, OT, and IoT environments. As organizations grapple with increasingly complex and distributed infrastructure, the ability to unify exposure management across all asset types has never been more important. This recognition highlights the growing demand for platforms that can provide a holistic view of an organization's attack surface. Are you confident that your security team has full visibility into every asset that could be an entry point for attackers?
Citrix patched a critical NetScaler vulnerability allowing unauthenticated attackers to read sensitive memory.
Citrix has issued an urgent patch for a critical NetScaler vulnerability that allows unauthenticated attackers to read sensitive memory and potentially hijack sessions. With the high risk of imminent exploitation, especially for internet-facing SAML deployments, this is a stark reminder of the importance of timely patching and proactive vulnerability management. The stakes are higher than ever as attackers increasingly target infrastructure that underpins critical business processes. How is your organization balancing the need for rapid patch deployment with the operational risks of downtime and disruption?
Software companies are facing a revenue durability crisis in 2026, with declining growth rates observed in firms like Bill.com and Snowflake.
The tech world is grappling with a harsh reality in 2026: revenue durability is the new north star for investors and founders alike. Recent data shows Bill.com’s growth plummeting from 90% to just 12%, while Snowflake’s dropped from 73% to 26%. This isn’t just a blip—it’s a fundamental shift in how software companies are valued. With valuations collapsing and fundraising becoming brutally competitive, the question isn’t just about growth anymore—it’s about whether that growth can sustain itself over the long term. How are you reassessing your pricing, retention, and defensibility strategies to weather this storm?
A founder gained firsthand industry experience by working as a technician in the pest control sector before building a vertical SaaS solution.
Terry Clarke took an unconventional path to building a vertical SaaS for pest control: he became a licensed technician first. By working in the trenches, he uncovered the real pain points that his future product could solve. This hands-on approach isn’t just about empathy—it’s about building solutions that resonate with users because you’ve lived their daily challenges. In a world where many founders rely on secondhand data, Terry’s story is a reminder that deep industry immersion can be the difference between a product that *sounds* good and one that *works* in the real world. What’s one industry you wish you could experience firsthand to build a better product?
Founders are advised to build a PMF (Product-Market Fit) machine by starting with a metric, ranking features by their potential impact, and continuously revising roadmaps.
Building a product isn’t about throwing features at a wall and hoping something sticks. In 2026, the most successful founders are treating PMF like a machine—start with a metric you want to move, rank your ideas by their potential impact, and ruthlessly prioritize high-impact, low-effort tasks. But the work doesn’t stop there. Constantly revising roadmaps and measuring everything ensures you’re not just shipping features, but shipping *the right* features. How are you balancing speed and precision in your product development process?
Software companies in 2026 face two primary paths to survival: accelerate growth with AI-native products or achieve over 40% operating margins.
The writing is on the wall for software companies in 2026: there are only two paths to survival. Either you double down on AI-native products to fuel explosive growth, or you ruthlessly optimize for over 40% operating margins—including stock compensation. CEOs must make a decisive choice: innovate or automate. Staying in the middle is no longer an option. The companies that thrive will be those that either redefine their market with AI or redefine their cost structures to outlast the downturn. Which path is your company pursuing, and what’s your first move?
Adobe is integrating AI 'coworkers' into Photoshop, Express, and Acrobat that allow users to make edits using plain text descriptions.
Adobe just announced a major leap in AI integration by introducing 'AI coworkers' into Photoshop, Express, and Acrobat. These tools allow users to make edits using plain text descriptions, fundamentally changing how designers and creatives interact with software. The addition of Project Moonlight, which recognizes user styles across apps, signals a move toward seamless, concept-to-finished-work workflows powered by Adobe Firefly. This isn't just another AI feature—it's a redefinition of the creative process. How will this shift the role of designers when AI handles the execution, and what new creative boundaries might we explore?
iOS 26.4 introduces expanded customization for the Liquid Glass interface with a 'Reduce Bright Effects' setting and improved 'Reduce Motion' feature.
Apple's iOS 26.4 update brings subtle but meaningful improvements to the Liquid Glass interface with two new accessibility features: 'Reduce Bright Effects' and an enhanced 'Reduce Motion' option. These updates reflect Apple's ongoing commitment to inclusive design, ensuring that interactions are less jarring for users sensitive to visual stimuli or motion. While still incremental, they signal a trend toward more granular UI control—a topic increasingly relevant as interfaces grow more dynamic. How can we balance aesthetic innovation with accessibility without compromising either?
NYSE partners with Securitize to build a 24/7 tokenized securities platform for stocks and ETFs.
The New York Stock Exchange (NYSE) has taken a monumental step toward integrating blockchain technology into traditional finance by partnering with Securitize to launch a 24/7 tokenized securities platform for stocks and ETFs. This move underscores a critical inflection point where major exchanges are no longer treating tokenization as an experiment but as a core upgrade for trading, settlement, and market access. With blockchain rails enabling faster, more transparent, and globally accessible trading, Wall Street's modernization race is accelerating. The implications for institutional players, retail investors, and DeFi integration are profound. How do you see tokenized securities reshaping the future of capital markets in the next five years?
Solana Foundation publishes a formal privacy framework for institutional adoption, defining four configurable modes for pseudonymity, confidentiality, and anonymity.
The Solana Foundation has unveiled a groundbreaking privacy framework designed to accelerate institutional adoption of blockchain technology. By defining four configurable modes—ranging from pseudonymity to full anonymity backed by ZK proofs and MPC—the framework addresses one of the biggest hurdles for institutional DeFi participation: regulatory compliance. With Solana's high throughput enabling encrypted order books and private credit risk calculations, this protocol stands to become the infrastructure backbone for enterprises requiring selective disclosure. As privacy moves beyond binary choices, the modular approach sets a new standard. How will enterprises balance transparency with privacy in the evolving regulatory landscape?
Tether announces its first full reserve audit conducted by a Big Four accounting firm to improve transparency beyond point-in-time attestations.
Tether is making a landmark move toward full transparency by engaging a Big Four accounting firm to conduct its first-ever full reserve audit. This development marks a shift beyond the periodic attestations that have long defined its transparency efforts, signaling a potential evolution in stablecoin governance and regulatory alignment. With $13.3 trillion in USDT settled onchain and a growing presence in institutional markets via USAT, credibility is paramount. The audit could reshape trust in stablecoins and influence broader adoption in regulated financial systems. How will this level of transparency impact the stablecoin market's competitive dynamics?
Bernstein predicts Bitcoin's cycle bottom is in place and maintains a $150,000 price forecast for year-end 2026.
Bernstein's latest research suggests Bitcoin's cycle bottom is already behind us, with a bold $150,000 price target for the end of 2026. The firm's strategic bull case hinges on Bitcoin's shift toward perpetual preferred instruments like STRC, which provide non-dilutive capital and supported aggressive accumulation during the downturn. With Bitcoin's 762,099 BTC treasury and below-history NAV premium, the firm views this as a leveraged play on further upside. This institutional endorsement, paired with growing adoption of blockchain-based instruments, reinforces Bitcoin's maturing role in global finance. Do you believe institutional strategies are now the primary driver of Bitcoin's long-term value?
MoonPay's Open Wallet Standard (OWS) launches as a local-first key management protocol to serve as the wallet layer for agentic payment standards.
MoonPay has introduced the Open Wallet Standard (OWS), a local-first key management protocol designed to underpin the next generation of agentic payments. By deriving accounts for EVM, Solana, Bitcoin, and other chains from a single BIP-39 seed and encrypting keys with AES-256-GCM, OWS reduces signing latency to microseconds while eliminating custodial dependencies. With 21 founding organizations including PayPal, Circle, and the Ethereum and Solana Foundations, this standard could redefine how agents and users interact with blockchain-based systems. As agentic commerce evolves, could local-first key management be the missing link to mass adoption?
Rain expands its Visa membership into Asia-Pacific to launch consumer and corporate credit card programs backed by stablecoins.
Rain is bringing its regulated Visa membership to the Asia-Pacific region, enabling partners to launch credit card programs backed by stablecoins. This expansion is particularly significant given the region's heavy stablecoin usage and major remittance corridors, positioning stablecoins as the settlement layer behind familiar payment experiences. As Asia-Pacific emerges as a hub for digital asset innovation, this move could accelerate mainstream adoption of crypto-backed financial products. How will traditional payment networks adapt to the rise of blockchain-native settlement layers?
Email.md converts Markdown into email-safe HTML that renders correctly across all major email clients.
Email development just got easier with Email.md, a tool that converts simple Markdown into email-safe HTML compatible with all major email clients. This addresses a long-standing pain point where complex HTML and inconsistent rendering engines (like Outlook's Word engine) make email development tedious. By simplifying the process, Email.md empowers developers to focus on content creation rather than debugging rendering issues across platforms. How could your team benefit from reducing the complexity of email development?
Agentic RAG improves standard RAG by adding an LLM to rewrite queries, route to sources, evaluate responses, and loop for better accuracy.
Agentic RAG is transforming how we retrieve and process information by introducing an LLM-driven feedback loop into traditional RAG pipelines. Unlike standard RAG, which struggles with ambiguous queries or fragmented answers, Agentic RAG dynamically rewrites queries, routes to the right sources, evaluates retrieved content, and iterates until accuracy is achieved. This approach is particularly valuable for enterprise applications where precision is critical. As AI systems become more autonomous, how can organizations leverage these advancements to improve decision-making and user experiences?
Node.js worker threads can isolate CPU-heavy user code from critical system processes to prevent event loop starvation.
Inngest faced a critical challenge with Node.js-based Connect services where CPU-heavy user code caused event loop starvation, leading to 'no available worker' errors. Their solution? Moving core services like WebSocket connections and heartbeats into a separate Node.js worker thread. This separation ensures system stability even under heavy load. For teams maintaining high-performance Node.js applications, this is a reminder of the importance of isolating critical processes. How are you optimizing your Node.js applications to handle resource contention?
Hegel is a new family of property-based testing libraries using the Python Hypothesis engine for Rust, with plans for Go, C++, OCaml, and TypeScript.
Hegel, a new property-based testing library family, is set to revolutionize how we approach software testing. Built on the Python Hypothesis engine, it currently supports Rust and is expanding to Go, C++, OCaml, and TypeScript. Property-based testing shifts the paradigm from validating specific cases to defining general rules that the software must adhere to, uncovering edge cases developers might miss. As software systems grow in complexity, how can property-based testing become a standard part of your testing strategy?
Video.js v10 beta reduces bundle sizes by up to 88% and prepares the framework for AI-augmented development.
Video.js, the popular open-source media player framework, has just entered its v10 beta phase with massive performance improvements—reducing bundle sizes by up to 88%. This modernization not only enhances performance but also prepares the framework for AI-augmented development. As media applications grow more complex, these optimizations are crucial for delivering seamless user experiences. How are you leveraging modern frameworks to future-proof your media applications?
TypeScript 6.0 has been announced as a transition release to prepare developers for TypeScript 7.0, including alignment changes, new features, and deprecations.
TypeScript 6.0 is here, and it’s setting the stage for a major evolution with TypeScript 7.0 on the horizon. This release isn’t just incremental—it’s a strategic step to modernize the language, introducing better type inference, updated standard library types, and forward-looking deprecations. For engineering teams, this means preparing now for smoother migrations and leveraging modern JavaScript practices. The shift also hints at deeper performance optimizations coming in TypeScript 7.0, which is rumored to be built on Go for even higher efficiency. How are you adapting your codebases to stay ahead of these changes?
Cloudflare launched its Gen 13 server fleet powered by AMD EPY0 5th Gen Turin 9965 processors, achieving double the throughput of Gen 12 servers after rewriting its core request handling layer in Rust.
Cloudflare just redefined edge computing performance with their Gen 13 server fleet, powered by AMD’s EPYC Turin 9965 processors. By ditching legacy NGINX stacks and rewriting their request handling layer in Rust, they achieved a remarkable 2x throughput boost while improving power efficiency by 50%. This isn’t just about raw numbers—it’s a testament to how low-level optimizations can transform large-scale systems. As edge computing becomes the backbone of modern applications, such architectural shifts will define the next generation of cloud providers. What performance bottlenecks in your infrastructure could benefit from similar rearchitecting?
Open standards were highlighted as the backbone of modern observability strategies, with 77% of organizations prioritizing them for flexibility and interoperability.
Observability is undergoing a quiet revolution. According to the latest data, 77% of organizations are prioritizing open standards to break free from vendor lock-in and improve data portability. Tools like Prometheus and OpenTelemetry are leading this charge, offering unified metrics, logs, and traces without the complexity of proprietary solutions. For teams tired of juggling disparate monitoring stacks, this shift promises not just efficiency but long-term resilience. How are you balancing open standards with legacy systems in your observability strategy?
Sysdig’s Threat Research Team developed syscall-level detection for AI coding agents like Claude Code and Gemini CLI to monitor for prompt injection and unauthorized access.
AI coding agents are powerful—but are they secure? Sysdig’s research team has uncovered a critical blind spot: these agents, running with full user permissions, can be exploited via prompt injection attacks buried in code comments or dependency files. Their syscall-level detection system, including four new Falco rules, provides real-time monitoring for unauthorized access and credential theft. As AI tools become ubiquitous in development, this work underscores the urgent need for proactive security measures. How can teams balance the agility of AI assistants with robust threat detection?
Google Cloud announced general availability of direct Identity-Aware Proxy integration for Cloud Run, enabling one-click authentication without load balancers.
Google Cloud is simplifying secure access with direct Identity-Aware Proxy (IAP) integration for Cloud Run. Now, teams can enable one-click authentication for serverless workloads without clunky load balancers, all while maintaining compatibility with Domain Restricted Sharing policies. This is a game-changer for teams looking to streamline access control while reducing operational overhead. In an era where zero-trust principles are non-negotiable, such integrations set a new standard for developer experience. How is your team modernizing authentication for serverless architectures?
Aqua Security’s CI/CD pipeline was compromised, leading to malware-laced versions of Trivy being pushed to Docker Hub between March 19-23.
Supply chain security just hit close to home. Aqua Security’s CI/CD pipeline was compromised, resulting in infected Trivy scanner versions being pushed to Docker Hub between March 19-23. Users who pulled these versions risked exposing CI/CD secrets, cloud credentials, and SSH keys. While Docker and Aqua acted swiftly to mitigate the issue, this incident serves as a stark reminder of how far-reaching supply chain attacks can be. It’s time to audit your dependencies and enforce strict key rotation policies. Have you reviewed your supply chain security posture lately?
A hardware bug in certain CPUs was uncovered through RocksDB stress testing, where the RDSEED instruction returned predictable values, breaking assumptions about randomness.
Even hardware isn’t immune to flaws. A recent RocksDB stress test revealed a critical CPU bug where certain processors returned predictable values from the RDSEED instruction, undermining cryptographic assumptions and causing duplicate IDs. This underscores a rarely discussed risk: dependencies at the hardware level can disrupt even the most well-architected systems. As we build increasingly complex distributed systems, validating every layer—including the silicon beneath—is no longer optional. How can engineering teams better account for hardware dependencies in their reliability and security planning?
n8n-MCP is a new Model Context Protocol server that gives AI assistants like Claude direct access to documentation for 1,239 n8n workflow automation nodes.
What if your AI assistant could build workflows for you—accurately and instantly? n8n-MCP is making that a reality by providing AI assistants like Claude direct access to 1,239 n8n workflow nodes. This open-source tool reduces workflow building time from 45 minutes to just 3 minutes in testing, eliminating guesswork and manual configuration. As AI tools become more integrated into daily workflows, such integrations will redefine productivity. How could your team leverage AI to automate repetitive tasks and reduce cognitive load?
Ahrefs Web Analytics offers a free, privacy-first alternative to Google Analytics with real-time tracking and AI tool visitor insights.
Ahrefs has launched a compelling free alternative to Google Analytics with Ahrefs Web Analytics. This tool stands out by prioritizing privacy—no cookies or personal data collection—and delivers real-time insights in under a minute. What’s particularly valuable for modern marketers is its ability to track visitors from AI tools like ChatGPT and Perplexity, bridging the gap between traditional analytics and emerging search behaviors. As privacy regulations tighten, solutions like this will shape the future of data-driven marketing. How are you adapting your measurement strategies to balance transparency and performance?
A 16-month experiment found AI-generated content gains initial traffic but fails to sustain rankings without SEO fundamentals.
New research reveals a harsh truth about AI-generated content: while it can achieve rapid indexing and early traffic spikes, it struggles to maintain visibility without strong SEO fundamentals. In a 16-month experiment, 71% of pages were indexed quickly but only 3% remained in the top 100 rankings long-term. The data underscores that trust signals, unique insights, and authority are non-negotiable even in an AI-driven content landscape. How can teams balance efficiency with the foundational elements that truly drive sustainable growth?
Study shows stacked discounts drive higher engagement and purchase intent than single discounts of equal value.
A recent study highlights a powerful psychological lever in discounting: stacked discounts outperform single discounts by a significant margin. Splitting a 25% discount into multiple parts increased shopping intent by 15.8% and drove 66.3% more views across 9,000+ deals. Customers perceive stacked offers as 33.1% more unique and feel 20.7% smarter when using them. The takeaway? Simple, clear stacking can transform how your audience perceives value—and ultimately, your conversion rates. How might you redesign your discount strategies to leverage this effect?
Surreal’s unhinged marketing approach targets marketers to organically amplify its brand in the cereal market.
Surreal’s rise in the cereal market is a masterclass in breaking through clutter with chaos. By positioning itself between sugary kids' cereals and bland health options, and wrapping its high-protein, low-sugar products in humor, the brand has cultivated a cult following on LinkedIn—targeting marketers as its amplification channel. Their paid media strategy leans into absurdity: fake celebrity billboards, eyebrow-raising collaborations, and anti-marketing aesthetics that defy consumer expectations. In a world of predictable brand messaging, Surreal proves that disruption can be a sustainable growth engine. What’s one unconventional tactic your brand could experiment with to stand out?
Webinar explains why TV can drive up to 35% more sales for small brands hitting growth plateaus.
Small brands often hit a wall with social media alone—so why aren’t more of them turning to TV? A new webinar reveals how adding TV to the mix can unlock up to 35% more sales even with limited budgets. The session covers why TV drives credibility and reach in ways social media can’t, and how to integrate it with channels like CTV, podcasts, and OOH. For brands plateaued in digital-only growth, this could be the missing piece. Have you considered diversifying your media mix beyond digital-first channels?
Guide demonstrates building real-time monitoring tools using Ahrefs Firehose API for brand mentions and competitor tracking.
The Ahrefs Firehose API is a goldmine for real-time marketing insights—but how do you turn raw data into actionable workflows? A new guide walks through building lightweight tools to monitor brand mentions, competitors, and keywords with surgical precision. From daily Slack digests to automated alerts, the system captures signals, processes them, and delivers results where your team works. In an era where speed and relevance define success, these kinds of automations can be a game-changer. What’s one manual process in your workflow that could be transformed with real-time data?
Content strategy fails when writers are expected to handle both execution and strategic planning.
A critical misstep in many content teams? Expecting writers to also own strategy. Writing and strategic planning require entirely different skill sets, yet job titles often conflate them. When strategists are absent, output lacks direction, and teams end up with content that doesn’t align to revenue or audience needs. The solution? Separate roles: hire a strategist to build the foundation, then empower writers to execute. How are you structuring your content teams to ensure both depth and direction?
ChatGPT Ads pilot lacks programmatic buying, targeting, and measurable ROI for advertisers.
OpenAI’s ChatGPT Ads pilot is leaving early advertisers in the dark. Without programmatic buying, minimal targeting, or measurement tools, it’s nearly impossible to demonstrate ROI. Two agency executives confirmed they couldn’t track business outcomes from their spend. As AI platforms rush to monetize, the gap between hype and deliverability remains wide. For performance marketers, this underscores the importance of waiting for mature measurement frameworks before betting big. How do you validate new ad platforms before scaling spend?
OpenAI shut down its Sora AI video app six months after launch due to high GPU resource consumption and competition with Anthropic and Google.
OpenAI's decision to shutter Sora, its AI video app launched just six months ago, signals a strategic pivot away from resource-intensive media generation. This move underscores the harsh reality that not all AI applications are sustainable at scale, especially when competing for compute with text and code-focused models. With Anthropic doubling down on efficiency and Google leveraging its infrastructure advantages, OpenAI's exit validates a focus on core competencies. For businesses exploring AI, this raises a critical question: Are we building for flash or for long-term viability? How will this shift in OpenAI's strategy influence your AI roadmap this year?
OpenAI canceled a $1B Disney partnership and 200-character licensing agreement after investing in video generation.
OpenAI's abrupt termination of its $1B partnership with Disney marks a dramatic turn in the company's AI video ambitions. Just months ago, this deal represented the largest mainstream collaboration in AI history, but the shutdown of Sora has rendered it obsolete. The breakdown highlights the risks of chasing high-profile partnerships without aligning on product-market fit. For industry observers, this serves as a cautionary tale about overcommitting to unproven technologies. How do you balance bold partnerships with the discipline of focusing on scalable, user-driven solutions?
Sam Altman delegated safety and security oversight at OpenAI to focus on data center construction, fundraising, and supply chains.
Sam Altman's decision to hand off safety and security oversight at OpenAI to focus on data centers and fundraising signals a strategic realignment ahead of the company's anticipated IPO. This move prioritizes infrastructure and capital over governance, raising questions about long-term risk management in AI development. As OpenAI scales toward a public listing, the shift underscores the tension between innovation velocity and responsible oversight. How should companies balance ambitious growth targets with the need for robust safety frameworks in AI?
OpenAI is developing a new AI model codenamed 'Spud' as its next major release.
OpenAI's development of 'Spud,' its next major AI model, could define the company's trajectory in the coming year. Following the shutdown of Sora and the cancellation of the Disney deal, all eyes will be on Spud to determine whether OpenAI's strategic pivot succeeds. If Spud delivers breakthrough performance, today's pruning may be forgotten; if not, the company faces existential questions about its future. For professionals in AI, this is a moment to reflect: Are we witnessing the birth of a new era or a correction in hype-driven innovation? What criteria should guide our expectations for the next generation of AI models?
OpenAI scaled back its ChatGPT shopping feature 'Instant Checkout' after low user adoption.
OpenAI's decision to scale back its ChatGPT shopping feature 'Instant Checkout' highlights the challenges of AI monetization beyond core capabilities. Despite the hype around AI-driven commerce, users have shown little appetite for automated purchasing tools. This failure underscores the importance of aligning AI features with genuine user needs rather than chasing trends. For businesses exploring AI integrations, it's a reminder to prioritize utility over novelty. How can we design AI tools that solve real problems rather than creating artificial ones?
The OpenAI Foundation committed over $1B to AI-driven scientific discovery and societal resilience initiatives.
The OpenAI Foundation's pledge of over $1B to AI-driven scientific discovery marks a significant commitment to leveraging AI for societal good. This investment could accelerate breakthroughs in climate modeling, healthcare, and materials science, areas where AI has already shown promise. For the tech community, it underscores the dual role of AI as both a commercial powerhouse and a tool for global challenges. How can we ensure that these investments translate into tangible benefits for society, rather than just economic returns?
Arm released its first in-house chip with Meta as the debut customer.
Arm's launch of its first in-house chip with Meta as its first customer marks a significant inflection point in the semiconductor industry. While the financial terms remain undisclosed, the potential scale of Meta's $135 billion capital expenditure budget suggests this could be transformative for Arm's business model. This development challenges the traditional fabless semiconductor model and indicates a move toward vertical integration in chip design. For companies relying on Arm's architecture, this shift could impact licensing strategies and competitive dynamics. How might this move accelerate the trend toward in-house chip development among major tech companies?
A federal judge called the Pentagon's actions against Anthropic 'troubling,' suggesting they may have been intended to cripple the company.
A federal judge's characterization of the Pentagon's actions against Anthropic as 'troubling' raises serious questions about the use of government influence in the AI sector. The implication that these actions may have been designed to weaken a key player underscores the growing intersection of national security and commercial AI. For policymakers and industry leaders, this case highlights the need for clearer guidelines on government engagement with AI companies. How can we ensure that regulatory and defense actions foster innovation rather than stifle it?
Meta was ordered to pay $375M for violating New Mexico law after failing to safeguard its apps from child predators.
Meta's $375M fine for failing to protect children from predators on its platforms sets a new precedent for corporate accountability in tech. This ruling comes amid growing scrutiny of social media platforms' role in online safety, particularly for vulnerable populations. For companies in the AI and tech sectors, it serves as a stark reminder of the legal and reputational risks of neglecting safety measures. How can organizations balance innovation with the ethical responsibility to protect users, especially the most vulnerable?
Anthropic's Claude Cowork now allows Claude to control Macs autonomously and run tasks remotely via a phone.
Anthropic's expansion of Claude Cowork to enable autonomous Mac control and remote task execution via phone represents a leap forward in AI agent functionality. This capability transforms AI from a passive assistant to an active collaborator, capable of handling complex workflows without direct user intervention. For businesses, this could redefine productivity tools and automation strategies. How will your team integrate autonomous AI agents into daily workflows, and what safeguards will you implement to ensure responsible use?
Researchers discovered that AI-designed polymers were four times tougher than expected due to a previously unknown quantum mechanical effect.
A team of researchers using AI to design new polymers made a stunning discovery: the materials were four times tougher than predicted, revealing a previously unknown quantum mechanical effect. This achievement underscores AI's potential to uncover fundamental scientific insights that even experts might miss. For industries reliant on materials innovation—such as manufacturing, energy, and healthcare—this could accelerate breakthroughs and reduce R&D costs. How can your organization leverage AI to push the boundaries of scientific discovery beyond traditional trial-and-error methods?
Halter, an AI-powered cow collar startup, raised $220M to expand its livestock monitoring technology.
Halter's $220M funding round to expand its AI-powered cow collars highlights the unexpected frontiers of AI adoption. While most headlines focus on consumer or enterprise AI, this investment proves that even traditional industries like agriculture are embracing AI for efficiency and sustainability. For tech innovators, it's a reminder that the most impactful applications of AI may not always be in the spotlight. How can we identify and support AI solutions in overlooked industries that stand to benefit the most from technological transformation?
Apple is testing a standalone Siri app and an 'Ask Siri' feature for deeper AI integration across its platforms.
Apple is quietly redefining the future of digital assistants with its upcoming standalone Siri app and cross-platform 'Ask Siri' feature. This isn't just another UI refresh—it marks Apple's pivot from traditional voice assistants to a systemwide AI agent with conversational text and voice capabilities. With deep integration into iOS 27 and macOS 27, Apple is positioning Siri as an intelligent orchestrator across all platforms. For businesses building on Apple's ecosystem, this means preparing for a new era where digital assistance becomes a seamless, ambient layer of interaction. How will your organization adapt to AI agents that operate as deeply integrated system components rather than standalone tools?
Claude Code's new Auto Mode allows autonomous permission decisions with safety filters.
Anthropic's introduction of Auto Mode for Claude Code represents a major leap in autonomous coding assistance, enabling AI agents to make permission decisions without manual intervention. With a separate classifier model reviewing actions for safety and scope compliance, this addresses critical concerns about uncontrolled agent behavior in development environments. For engineering teams, this could dramatically reduce friction in automated workflows while maintaining security boundaries. As AI systems take on more autonomous roles, the balance between productivity gains and risk mitigation becomes paramount. How can organizations implement these capabilities without compromising control over critical systems?
Meta is shutting down its Metaverse project after investing $80 billion.
Meta's decision to shutter its Metaverse project after a $80 billion investment represents one of the most dramatic strategic pivots in tech history. This move underscores the harsh reality that technological vision must align with genuine user demand and practical applications. While immersive computing remains a compelling long-term opportunity, Meta's experience serves as a cautionary tale about overestimating adoption timelines and underestimating user resistance to headset-based experiences. For companies investing in AR/VR, this highlights the need for clear value propositions beyond novelty. What lessons should the tech industry draw from Meta's Metaverse retreat when evaluating emerging technologies?
Comments