Facebook Marketplace has launched new Meta AI-powered tools designed to make selling faster and more efficient for users in the US and Canada. Sellers can now upload item images and let Meta AI automatically generate draft listings, fill in details, and suggest prices based on similar local listings. The platform also enables sellers to offer shipping with prepaid labels and manage all shipped orders through a centralized dashboard.

To improve communication, Meta AI can now draft and send automatic replies to buyer inquiries using information from the listing, such as item description, availability, pickup location, and price. Sellers can preview and edit these replies during listing creation. Additionally, Meta has introduced AI-generated profile summaries that display key details about a seller’s Facebook activity, including account age, friend count, and Marketplace history, to enhance transparency and trust.

These updates mark a continued evolution of Facebook Marketplace, which has been connecting millions of buyers and sellers for a decade and now hosts over 3.5 million daily listings across the US and Canada.

Google has introduced Groundsource, a new AI-powered methodology designed to improve the prediction of urban flash floods. Announced on March 12, 2026, Groundsource uses the Gemini model to analyze decades of public reports and Google Maps data, creating a high-quality historical dataset of flood events. The system identified over 2.6 million historical flood incidents across more than 150 countries and trained a new model capable of forecasting urban flash floods up to 24 hours in advance. These forecasts are now available through Google’s Flood Hub platform.

The initiative is part of Google’s broader Crisis Resilience efforts, which aim to provide early warnings for natural hazards. Groundsource addresses a long-standing data gap that had limited the ability to predict flash floods before they occurred. By transforming public information into structured datasets, the project enhances global preparedness and supports scientists and partners with an open-source benchmark for further research.

Google noted that the same AI-driven approach could be extended to other natural disasters such as landslides or heat waves, contributing to improved global resilience and disaster readiness.

Perplexity announced on Wednesday a new AI agent called Personal Computer, designed to run continuously and integrate local applications with Perplexity Computer. The system operates on Perplexity’s secure servers but is powered by Apple’s M4 Mac mini hardware. The company describes the platform as working within a secure environment that includes safeguards such as approval requirements for sensitive actions, full audit trails for each session, and a kill switch for user control. Each query runs in its own secure sandbox.

The announcement follows earlier social media interest in using Mac minis for AI workloads, including the open-source assistant Clawdbot. Perplexity claims its Personal Computer is “more powerful than any AI system ever launched.” However, details remain unclear regarding which Mac mini configuration is used, whether Apple supplies the hardware, and how the system will be sold. Currently, only a waitlist is available, with no pricing information disclosed.

Rumors also suggest Apple may soon release an M5 Mac mini, though Perplexity’s announcement does not mention chip specifications or configurations.

Star Cineplex is set to launch a new multiplex cinema branch in Narayanganj during Eid-ul-Fitr 2026, according to the company. The new branch is located at Simanto Tower in Jalkuri, Narayanganj, and will officially open on Eid day. It features three halls, two with 178 seats each and one with 75 seats, all equipped with modern sound systems, giant screens, and world-class facilities.

Mesbah Uddin Ahmed, Assistant General Manager of Media and Marketing at Star Cineplex, described the new branch as an Eid gift for Narayanganj residents, particularly for local movie lovers. He said the opening fulfills a long-standing demand from the area’s audience and reflects the company’s satisfaction in expanding as planned.

Ahmed added that Star Cineplex prioritizes high-quality viewing environments and advanced technology to meet evolving audience preferences. The company plans to open additional branches later this year, extending its presence from major cities to district and sub-district levels across Bangladesh.

Meta announced that it is expanding its custom silicon program with four new generations of Meta Training and Inference Accelerator (MTIA) chips to be developed and deployed within the next two years. The new chips will support ranking, recommendations, and generative AI (GenAI) workloads, marking a faster release cycle than typical industry standards. MTIA 300 is already in production for ranking and recommendations training, while MTIA 400, 450, and 500 will focus primarily on GenAI inference production through 2027.

The company’s AI infrastructure strategy centers on a portfolio approach that combines its own MTIA chips with silicon sourced from other industry leaders. Meta has deployed hundreds of thousands of MTIA chips for inference workloads across organic content and ads, achieving higher compute efficiency and cost-effectiveness compared to general-purpose chips. The modular design of MTIA allows new chips to integrate seamlessly into existing rack systems, reducing time-to-production.

Meta’s roadmap emphasizes rapid, iterative development, an inference-first design philosophy, and alignment with industry standards such as PyTorch, vLLM, Triton, and the Open Compute Project. This approach aims to sustain innovation speed and scalability as the company advances toward its goal of enabling personal superintelligence.

NVIDIA has launched Nemotron 3 Super, a 120‑billion‑parameter open model with 12 billion active parameters designed to power complex agentic AI systems at scale. The model, available immediately, integrates advanced reasoning capabilities to complete tasks efficiently and accurately for autonomous agents. It features a 1‑million‑token context window to maintain full workflow memory and prevent goal drift in multi‑agent applications.

Nemotron 3 Super employs a hybrid mixture‑of‑experts architecture that delivers up to five times higher throughput and twice the accuracy of its predecessor. Running in NVFP4 precision on the NVIDIA Blackwell platform, it achieves up to four times faster inference than FP8 on NVIDIA Hopper without accuracy loss. The model has achieved top rankings on Artificial Analysis and DeepResearch Bench leaderboards for efficiency and reasoning coherence. NVIDIA is releasing it with open weights under a permissive license, enabling deployment across workstations, data centers, and cloud environments.

Industry partners including Perplexity, Amdocs, Palantir, Siemens, and Dell Technologies are integrating Nemotron 3 Super into their AI ecosystems to enhance search, software development, cybersecurity, and manufacturing automation. The model is accessible through build.nvidia.com, OpenRouter, and Hugging Face.

Iran’s Islamic Revolutionary Guard Corps (IRGC)–affiliated Tasnim News reported that major US technology companies with alleged technological cooperation with Israel could be considered potential targets. The report stated that offices and cloud infrastructures of several US-based firms have been listed as facilities whose technologies were claimed to be used for military purposes, describing them as part of Iran’s new target list.

According to Tasnim, as regional conflict increasingly involves infrastructure, Iran’s definition of legitimate targets is expanding. The companies named in the report include Google, Microsoft, Palantir, IBM, Nvidia, and Oracle. The report also noted that parts of these firms’ cloud-based infrastructures are located in various Israeli cities and in several Gulf countries.

No immediate comment was available from the mentioned companies or the countries concerned, according to the report sourced from Al Jazeera.

China has released its 141-page 15th Five-Year Plan, presented at the National People’s Congress on March 5, outlining an ambitious strategy to dominate next-generation technologies, raw materials, and industries. The plan positions artificial intelligence (AI) as a central pillar across the economy, with goals to double humanoid robotics production within five years, accelerate quantum communication and nuclear fusion research, and advance brain-computer interface technologies. It also projects AI-related industries to exceed 10 trillion yuan in value during the plan’s term.

Analyst Shanaka Anselm Perera described the plan as a national technological mobilization rather than a mere economic policy, calling it a “war plan” that contrasts with the U.S. CHIPS and Science Act of 2022. While Washington focuses mainly on semiconductor production, Beijing’s approach spans AI integration across heavy industry and services, investment in quantum computing, and control over critical raw materials, especially rare earth elements.

The plan strengthens China’s grip on rare mineral processing and export controls, potentially challenging U.S. defense supply chains that depend on these materials for systems like the F-35 fighter jet. Analysts warn this could reshape future power dynamics long before any battlefield confrontation.

ChatGPT has launched new interactive visual explanations to make learning math and science more engaging and intuitive. The update introduces dynamic visual modules for over 70 core concepts, allowing users to manipulate variables and observe real-time changes in formulas and graphs. These tools are available globally across all ChatGPT plans starting today, aiming to help users better grasp abstract ideas through hands-on exploration.

The announcement follows findings from a Gallup survey showing that more than half of U.S. adults struggle with math, and many parents lack confidence in helping their children learn it. ChatGPT already serves 140 million users weekly who rely on it to understand math and science concepts, prepare for exams, and explore new topics. The new feature builds on research suggesting that interactive, visual learning can strengthen conceptual understanding compared to traditional instruction.

Educators have noted that the feature emphasizes conceptual understanding by encouraging learners to explore deeper connections between ideas, helping them retain knowledge more effectively.

Singer Asif Akbar has been acquitted in a case filed over possession of foreign liquor without a license. The verdict was delivered on Monday, March 9, by Judge Ayesha Akhter Mousumi of Dhaka Additional Metropolitan Sessions Judge Court-2 after concluding the final arguments. The case had been ongoing for eight years. Following the verdict, Asif expressed satisfaction, saying the allegation was baseless and that he had received justice after a long legal battle.

According to court sources, the prosecution failed to prove the charges, leading to Asif’s acquittal. The case originated from a 2018 incident when four bottles of foreign liquor were recovered from his office during his arrest in another case. A separate case was later filed at Tejgaon Police Station under the Narcotics Control Act of 2018.

Asif stated that he respected the court process and refrained from politicizing the issue. He thanked his family, fans, and supporters for enduring the long ordeal, calling the verdict a relief for them all.

On Monday, Anthropic filed a lawsuit against the U.S. Department of Defense after being labeled a supply chain risk. Within hours, nearly 40 employees from OpenAI and Google, including Google’s chief scientist Jeff Dean, submitted an amicus brief supporting Anthropic’s case. The brief criticized the Trump administration’s decision and warned of the dangers posed by certain military uses of artificial intelligence.

The dispute stems from Anthropic’s refusal to allow its AI technology to be used for domestic mass surveillance or fully autonomous weapons. Following the breakdown of negotiations, the Pentagon’s designation barred Anthropic from military contracts and restricted other companies using its products. The brief argued that the designation was retaliatory and harmful to the public interest, emphasizing that AI-driven surveillance and autonomous lethal systems pose serious ethical and governance risks.

The signatories, describing themselves as engineers and researchers from leading U.S. AI labs, urged for guardrails and human oversight in military AI applications, warning that current systems remain unreliable and potentially dangerous when deployed without proper constraints.

Anthropic has filed a lawsuit against the U.S. government after being designated a supply-chain risk by the Trump administration. The case, lodged in a California district court, accuses the administration of retaliating against the company for refusing to relax its restrictions on the acceptable uses of its artificial intelligence technology, particularly concerning mass domestic surveillance and fully autonomous weapons. The company argues that the designation and related actions violate its constitutional rights and unlawfully punish it for its stance on AI safety.

The dispute follows weeks of tension between Anthropic and the Department of Defense over military applications of AI. The supply-chain risk label, typically reserved for foreign entities posing cybersecurity or national security threats, led President Donald Trump to order all federal agencies to stop using Anthropic’s technology within six months. The move sparked bipartisan concern about the potential chilling effect on companies that disagree with government policy.

According to the lawsuit, several agencies, including the General Services Administration, Treasury, and State Department, have already ended or plan to end contracts with Anthropic. The Pentagon declined to comment on the matter.

Figure has unveiled a new demonstration of its humanoid system Helix 02, showing the robot autonomously tidying a living room. The system, released a little over a month ago, controls its full body directly from visual input and performs long-horizon tasks without specialized programming. In this latest test, Helix 02 navigates through a cluttered living room while manipulating objects, tools, and containers, performing coordinated actions such as cleaning surfaces, handling flexible items, and reorganizing objects.

The demonstration highlights Helix 02’s ability to integrate locomotion, dexterity, and sensing through a single neural architecture. The robot learned new behaviors—like wiping surfaces, tossing pillows, and pressing remote buttons—simply by adding data, without new algorithms or task-specific engineering. Each action combines movement and manipulation in a constantly changing environment, reflecting the complexity of real household tasks.

According to Figure, Helix 02’s expanding skill set represents progress toward scalable humanoid intelligence, where one general-purpose model can learn and perform a broad range of everyday activities in homes and workplaces.

Anthropic has introduced Code Review for Claude Code, a new agent team-based system designed to perform in-depth reviews on every pull request. The feature, now available in research preview for Team and Enterprise users, deploys multiple agents to identify, verify, and rank bugs by severity. Each review produces a concise summary comment and detailed in-line notes, aiming to catch issues that human reviewers might overlook. The system scales its depth based on the size and complexity of the code changes, with an average review time of about 20 minutes.

According to Anthropic, the internal use of Code Review has significantly increased the proportion of pull requests receiving substantive feedback. The company reports that the tool has helped detect critical bugs before merging, including issues that could have caused major service disruptions. Early access users have also reported similar benefits, with the system uncovering latent bugs in existing codebases.

Code Review is positioned as a more thorough but costlier alternative to the existing Claude Code GitHub Action. Reviews are billed based on token usage, typically ranging from $15 to $25 per review, with administrative controls available to manage spending.

X has announced a new feature that allows users to listen to long-form articles within the app, using the voice of xAI’s Grok artificial intelligence chatbot. The articles will now include a “Listen” button, enabling audio playback both while scrolling through X and when switching to other apps through background play. The feature aims to make it easier for users to consume lengthy content and help creators reach wider audiences.

The update is part of X’s broader strategy to attract content creators and boost exclusive, in-depth posts on the platform. X is particularly focused on long-form content because such material provides richer data for its xAI projects, which use information shared on X to improve AI outputs. Grok, xAI’s main product, relies on language understanding and real-time insights drawn from X discussions.

According to X’s head of product Nikita Bier, X Articles have grown 18 times over the past three months. The company has also incentivized creators with a $1 million prize for top-performing articles, underscoring its push to expand long-form content and strengthen its AI ecosystem.