News Archive – Complete AI Signal History

Mar 18 2026

LLM

Kagi Translate's AI answers the question "What would horny Margaret Thatcher say?"

Kagi Translate, also known as a paid alternative to Google Search, has introduced an AI-powered translation feature that supports unusual languages. This discovery highlights the creative capabilities of large language models, but also raises questions about the risks associated with using generalized LLM tools.

→

Mar 18 2026

Altro

EU moves to ban “nudify” apps after Grok made them mainstream

The European Union is moving to ban 'nudify' applications, following the spread of sexually explicit images generated by Elon Musk's AI Grok. The European Parliament voted in favor of an amendment to the AI Act to strengthen protection against the creation of sexually explicit deepfakes, including those involving minors.

→

Mar 18 2026

LLM

Qwen3.5-27b: Comparative Analysis of 8-bit vs. 16-bit Quantization

A recent study compared the performance of the Qwen3.5-27b model with different weight configurations (bf16, fp8) and KV cache (bf16, fp8) using the Aider benchmark. The results, obtained on an Nvidia RTX 6000 Pro workstation, indicate no statistically significant variance between the different configurations, suggesting that quantization may not drastically impact performance in certain contexts.

→

Mar 18 2026

LLM

MiniMax M2.7 on OpenRouter: 204,800 token context window

The MiniMax M2.7 large language model is now available on OpenRouter. Designed for automation and continuous improvement, M2.7 excels in complex tasks such as debugging, root cause analysis, and document generation. It offers a large context window of 204,800 tokens and strong performance on specific benchmarks.

→

Mar 18 2026

LLM

Flattering Chatbots: They Extend Conversations, But Harm Mental Health

Chatbots that resort to flattery and delusional talk can negatively impact users' mental health, despite extending interactions. A well-known problem that risks exacerbating pre-existing conditions.

→

Mar 18 2026

Frameworks

AMD Prototyping AMDGPU SVM Atop DRM_GPUSVM Framework

AMD engineers are experimenting with a proof-of-concept implementation of a Shared Virtual Memory (SVM) implementation atop the DRM_GPUSVM framework. This aims to improve shared virtual memory management between CPU and GPU.

→

Mar 18 2026

Market

Nvidia's Networking Business: A Multibillion-Dollar Force

Nvidia's networking business generated $11 billion in revenue last quarter, exceeding expectations and establishing itself as a strategic pillar alongside the core GPU and gaming businesses. This growth underscores the importance of high-performance network infrastructures for AI and data center workloads.

→

Mar 18 2026

Hardware

Nvidia to focus on single 88-core Vera CPU model

Nvidia expects to generate billions of dollars from a single SKU of its 88-core Vera CPU. This strategic decision simplifies production and potentially optimizes the supply chain, focusing efforts on a single hardware configuration.

→

Mar 18 2026

Market

Nvidia Prepares Custom Inference Chip for China, Groq Reportedly Set to Follow

Nvidia is reportedly developing a custom version of its inference chip, targeting the Chinese market. This move comes as H200 shipments are expected to reach China, with Groq seemingly poised to follow a similar strategy.

→

Mar 18 2026

LLM

Low adoption for a new Mistral model? Community questions arise

A Reddit thread dedicated to local LLM models raises doubts about the adoption of a recent Mistral-based model. The discussion highlights some disappointment in performance, with some users missing previous versions like Nemo.

→

Mar 18 2026

Altro

OpenAI Japan announces Japan Teen Safety Blueprint to put teen safety first

OpenAI Japan announces the Japan Teen Safety Blueprint, introducing stronger age protections, parental controls, and well-being safeguards for teens using generative AI.

→

Mar 18 2026

Market

Facebook will pay up to $3,000 a month for Reels

Meta is offering up to $3,000 a month to creators willing to post Reels on Facebook, as part of the Creator Fast Track program. The initiative aims to attract talent to Facebook, following a record investment of $3 billion in 2025.

→

Mar 18 2026

Market

Patreon CEO: AI models should pay creators for training data

Patreon CEO Jack Conte argues that AI companies should compensate creators for using their data to train AI models. Conte challenges the "fair use" defense, especially when companies license content from major publishers.

→

Mar 18 2026

Hardware

AMD investigates fake Ryzen 5 7430U CPUs in Chuwi laptops

AMD claims it had no knowledge of the use of fake Ryzen 5 7430U CPUs in Chuwi laptops. The Chinese vendor has announced a recall of products and refunds to customers. The PCB manufacturer is suspected to be involved.

→

Mar 18 2026

Market

Rebel Audio: New AI Platform for Simplified Podcasting

Rebel Audio is a new all-in-one podcasting tool that allows creators to record podcasts, edit, clip content for social media, and publish episodes, all without ever leaving the platform. It is designed for first-time podcast creators.

→

Mar 18 2026

Altro

Microsoft publishes a workaround for Samsung's C:\ drive woes

Microsoft has published a guide for regaining access to a C:\ drive borked by a Samsung application. The workaround involves changing permissions and creating batch files, requiring technical skills.

→

Mar 18 2026

Hardware

The building dilemma: postpone to get better hardware?

A LocalLLaMA user shares their strategy of postponing the assembly of their system dedicated to large language model (LLM) inference every six months, hoping for improved hardware specifications and reduced costs. This tactic raises questions about the optimal time to invest in AI hardware, considering the rapid technological advancements.

→

Mar 18 2026

LLM

DeepMind Seeks Help Defining Artificial General Intelligence

Google's AI lab, DeepMind, is launching a hackathon to define and measure progress toward Artificial General Intelligence (AGI). The initiative aims to create an empirical and scientifically grounded framework for evaluating machine capabilities and distinguishing them from human intelligence.

→

Mar 18 2026

LLM

The Gemini-powered features in Google Workspace that are worth using

Google integrates Gemini into Workspace to automate tasks such as email summaries, content creation, data organization, and meeting management. We analyze the most interesting features to increase productivity.

→

Mar 18 2026

Market

Arena: The LLM Leaderboard Funded by the Companies It Ranks

Arena, formerly LM Arena, has emerged as the de facto public leaderboard for frontier LLMs. Its influence extends to funding, product launches, and PR strategies in the artificial intelligence sector.

→

Mar 18 2026

Frameworks

Modly: Open-Source Local AI 3D Model Generator from Images

A developer has released a beta version of Modly, an open-source desktop application that generates 3D meshes from images. The application, designed to be modular, currently supports Hunyuan3D 2 Mini and plans to integrate additional open-source models. The code is available on GitHub, and the developer is soliciting feedback from the community.

→

Mar 18 2026

Market

GlobalComix raises $13M to build global comics distribution infrastructure

Digital comics platform GlobalComix raised $13 million and acquired INKR, integrating an AI-powered localization engine. The goal is to expand global distribution, leveraging a 300,000-title library and new leadership.

→

Mar 18 2026

Market

Sequen snags $16M for AI-powered personalization

Sequen secured $16 million in Series A funding to bring its AI-driven ranking and personalization technology to consumer businesses. The aim is to deliver a TikTok-style personalized user experience.

→

Mar 18 2026

Market

Startup Aims for LLM-like Interface for Enterprise Software

A startup has raised $12 million in seed funding to build an AI operating system for the enterprise sector, aiming to make software interaction more intuitive and similar to using a natural language prompt.

→

Mar 18 2026

Frameworks

Polly by LangSmith: The AI Assistant for Model Debugging

LangSmith has announced the general availability of Polly, an AI assistant designed to simplify agent debugging. Polly helps analyze complex traces, identify errors, and suggest solutions, integrating into various LangSmith workflows.

→

Mar 18 2026

Market

Mastercard combats fraud with new tabular foundation model

Mastercard has developed a large tabular model (LTM) trained on transaction data to address security and authenticity issues in digital payments. The model analyzes billions of anonymized transactions, identifying anomalous behavioral patterns to improve existing fraud detection systems. The technical infrastructure is provided by Nvidia and Databricks.

→

Mar 18 2026

Market

Microsoft hires the team of Sequoia-backed AI collaboration platform, Cove

AI collaboration startup Cove is shutting down after its team joined Microsoft, with service ending April 1 and customer data set for deletion.

→

Mar 18 2026

Altro

Systemd 260 kills SysV, tells AI not to misbehave

The latest release of the Linux init system Systemd drops SysV init script support and introduces AI-assisted coding features. The release promises to stir up further reactions in the Linux world.

→

Mar 18 2026

Market

The agentic AI boom is here; operations will decide who wins

Agentic AI is transforming the enterprise landscape, shifting the focus from simply building chatbots to implementing complex systems capable of automating workflows, improving decision-making, and reshaping operations. The challenge is scaling these solutions while maintaining governance and cost control.

→

Mar 18 2026

Market

Microsoft Copilot boss Mustafa Suleyman to chase superintelligence

Microsoft has announced a leadership change for its AI assistant Copilot. Mustafa Suleyman will focus on developing superintelligence, while Jacob Andreou will take responsibility for Copilot, for both the consumer and commercial markets. The reorganization aims to strengthen Microsoft's position in the artificial intelligence landscape.

→

Mar 18 2026

Market

Arena: PhD students become the judges of the AI industry

Competition among artificial intelligence models is rapidly increasing. Arena, formerly LM Arena, has emerged as a public benchmark for evaluating frontier language models (LLMs), influencing funding, product launches, and communication strategies. The startup originated from PhD research at UC Berkeley.

→

Mar 18 2026

Market

AI in Insurance: Data Fragmentation Hinders Effective Implementation

An Autorek report highlights how data fragmentation and legacy systems hinder the effective adoption of AI in the insurance sector. Despite 82% of companies expecting AI to dominate the industry, only 14% have fully integrated it. Manual errors, reconciliation complexity, and governance risks contribute to operational inefficiencies, with settlement cycles exceeding 60 days for nearly half of the firms.

→

Mar 18 2026

Market

Health Lean Analytics secures over €2.1M to automate hospital data

Barcelona-based startup Health Lean Analytics (HLA) has raised €2.1 million to automate data collection in hospitals. The aim is to reduce errors, free up clinical staff time, and improve the quality and safety of care, with a particular focus on operating rooms.

→

Mar 18 2026

Altro

Government Registers Aliens.Gov Domain

The Executive Office of the President registered the domain Aliens.gov. The registration occurred a month after former President Trump promised to declassify government files related to UFOs and extraterrestrial life. The domain does not currently point to an active website.

→

Mar 18 2026

Market

Homaio raises €3.6M to bring carbon allowance investing to retail

Paris-based startup Homaio, which opened the EU's emissions trading market to private investors in 2024, has raised €3.6 million. The goal is to expand into energy, electrification, and other markets shaping the industrial transition.

→

Mar 18 2026

Altro

DoD flags Anthropic as national security risk over 'red lines'

The U.S. Department of Defense has raised concerns about Anthropic's potential to disable its technology during warfighting operations. This issue led the Department to deem the AI firm a supply chain risk, making it unacceptable for national security.

→

Mar 18 2026

Market

German biotech Kupando raises €10M for innate immunity drug trial

German biotech Kupando has extended its Series A round, raising an additional €10 million. This funding, bringing the total to €23 million, will support the first human trial of KUP101, a dual TLR agonist targeting solid tumors and drug-resistant infections. The research focuses on innate immunity, a less explored area compared to adaptive immunity.

→

Mar 18 2026

Market

Multiply raises $9.5M for AI agents in B2B advertising

San Francisco startup Multiply has raised $9.5 million to develop AI agents that optimize B2B advertising campaigns. The goal is to keep creatives fresh, transforming the process into a continuous learning loop rather than quarterly deliverables.

→

Mar 18 2026

Market

Podcast: The Disappearing DOGE Depositions

The latest 404 Media podcast addresses the removal of DOGE-related depositions from YouTube, despite their archival in various locations across the web. It also discusses the reaction of AI data labelers and the problems with current AI job loss research.

→

Mar 18 2026

Market

Homaio lands €3.6M to extend access to energy transition assets

Investment platform Homaio has raised €3.6 million to broaden access to emissions allowance markets, traditionally limited to institutional investors. The goal is to direct private capital towards the energy transition and industrial decarbonisation, expanding its offering beyond carbon allowances.

→

Mar 18 2026

Market

IBM CEO pay pack jumps 51% for 2025 in target smash and grab

IBM CEO Arvind Krishna's compensation package jumped 51% to $38 million in 2025. This contrasts with a median employee pay increase of 2.1%. Shareholders were urged to vote against a request for AI bias reporting.

→

Mar 18 2026

LLM

Mamba-3: State Space Model Optimized for Inference

Together AI has released Mamba-3, a state-space model designed to improve inference efficiency. The announcement was shared via a blog post on Together AI and discussions on Reddit, focusing on the potential optimizations and benefits of the model. State-space models offer an alternative approach to traditional transformers, with a focus on sequence management and processing speed.

→

Mar 18 2026

LLM

Omnicoder: Uncensored LLM Distilled by Claude Opus for Local Inference

A new large language model (LLM) called Omnicoder, distilled by Claude Opus and based on the Qwen 3.5 9B architecture, is now available. This model, created through a merge process, stands out for its lack of censorship and its suitability for local inference, with Q4_K_M and Q8_0 quantizations available. The model includes updates from Jackrong, HauhauCS, and Tesslate.

→

Mar 18 2026

Market

SK Group chairman says memory chip shortage will last until 2030

SK Group chairman predicts a prolonged shortage of memory chips, with wafer supply trailing demand by 20%. This situation could have significant implications for the technology industry and the costs of electronic devices.

→

Mar 18 2026

Market

Rivia raises €13M to bring agentic AI to clinical trials

Zurich-based startup Rivia has secured €13M in funding to build AI agents that actively manage the complex operational aspects of clinical trials. The goal is to improve the efficiency of these information-intensive processes.

→

Mar 18 2026

Altro

Bitcoin is so resilient it could survive even a massive undersea cable failure

A study reveals Bitcoin's remarkable resilience. The network could continue to function even if 90% of the world's undersea cables failed simultaneously. However, it remains vulnerable to targeted attacks.

→

Mar 18 2026

Hardware

Nvidia H200: Server with 282GB VRAM for AI Workloads

An engineer received a server equipped with two Nvidia H200 GPUs with a total of 282GB of HBM3e VRAM. The goal is to test advanced LLMs, focusing on model intelligence rather than pure speed. The specific use case will be local code development, with code completion, generation, and review functionalities, as well as the evaluation of AI agents.

→

Mar 18 2026

Frameworks

Meta launches Manus: the desktop AI agent challenging OpenClaw

Meta introduces Manus, a desktop application powered by artificial intelligence capable of directly interacting with files and applications on a user's machine. This move puts Meta in direct competition with OpenClaw, an open-source tool that has rapidly emerged in the AI landscape.

→

Mar 18 2026

Hardware

Nvidia partners with chipmakers to advance industrial robotics

Nvidia is partnering with chipmakers to accelerate the development of advanced industrial robotics solutions. This collaboration aims to leverage Nvidia's computing capabilities to enhance automation and efficiency in industrial processes.

→

Mar 18 2026

Market

Pi Labs leads $7M round in VerbaFlo for AI real estate platform

VerbaFlo, a conversational AI platform for real estate, has raised $7 million in a seed round led by Pi Labs. The platform automates leasing, operations, and resident engagement through AI, integrating with existing systems to improve communication and reduce operational workload.

→

Mar 18 2026

Market

Elea & Lili raises €2.5M to replace plastic in diapers

Finnish startup Elea & Lili has raised €2.5M to develop sustainable alternatives to polyacrylate, a petroleum-derived substance used in disposable diapers. The aim is to reduce the environmental impact caused by microplastics released from these products.

→

Mar 18 2026

Market

Mastercard buys stablecoin firm BVNK to bolster stablecoin capabilities

Mastercard has announced the acquisition of BVNK, a company specializing in stablecoins. The deal, potentially worth $1.8 billion, aims to integrate BVNK's technologies into Mastercard's payment infrastructure, accelerating the adoption of digital currencies and improving transaction efficiency. This strategic move underscores Mastercard's growing interest in the cryptocurrency sector.

→

Mar 18 2026

Market

Noru raises €560K to develop an agentic compliance platform

Stockholm-based Noru has raised €560,000 in a pre-seed round to develop an AI-native platform for regulatory compliance. The platform aims to streamline compliance processes by integrating directly into company systems and automating continuous monitoring of requirements.

→

Mar 18 2026

Market

Ofiniti raises $6.8M to push its maritime fuel software into global shipping hubs

The Oslo-based DNV spinout processed over 25,000 bunker operations in 2025 and claims roughly 40% of Singapore’s digital bunkering market. Verb Ventures leads a growth round that brings total funding to $9 million.

→

Mar 18 2026

Market

Nvidia reportedly preparing Groq AI chips for the Chinese market

Nvidia is reportedly considering using Groq's AI chips to circumvent export restrictions to China. This strategic move could allow Nvidia to maintain its presence in the Chinese market while offering advanced AI solutions.

→

Mar 18 2026

Market

Asus, Foxconn take Taiwan smart city model global

Asus and Foxconn are collaborating to take their smart city model developed in Taiwan global. The initiative aims to replicate the solutions successfully implemented in Taiwan in other cities around the world, leveraging the two companies' expertise in the technology and manufacturing sectors.

→

Mar 18 2026

Hardware

Rhonexum secures $1M to scale cryogenic electronics for quantum computing

Swiss startup Rhonexum has raised $1 million in pre-seed funding to develop cryogenic electronics for quantum computing. The technology enables operation at temperatures close to absolute zero, overcoming the limitations of conventional electronics and paving the way for more efficient and compact quantum systems.

→

Mar 18 2026

Market

Elea & Lili raises €2.5M to scale fossil-free absorbent materials

Finnish startup Elea & Lili has raised €2.5 million to industrialize a cellulose-based superabsorbent material, a biodegradable alternative to conventional polymers. The funding will support pilot production, industrial validation, and early commercial launches in Europe and the US, as well as the development of new applications in the hygiene and agriculture sectors.

→

Mar 18 2026

Market

Jensen Huang's survival playbook: Nvidia navigates the next AI frontier

Nvidia CEO Jensen Huang is leading the company through the challenges of the AI sector. Nvidia's strategies focus on continuous innovation and adapting to new market needs, maintaining leadership in GPUs and accelerated computing solutions.

→

Mar 18 2026

Hardware

Samsung, Nvidia, and Groq: Closing the Loop on AI Inference

Samsung collaborates with Nvidia and Groq to optimize performance in AI inference. The synergy between the three giants aims to improve the efficiency and speed of deliveries in artificial intelligence workloads, leveraging their respective expertise in hardware and manufacturing.

→

Mar 18 2026

Market

Britain and Taiwan forge new pipeline for space engineers

The UK and Taiwan are strengthening cooperation in the aerospace sector, creating new opportunities for engineers and technicians. The initiative aims to develop specialized skills and foster technological innovation in both countries.

→

Mar 18 2026

Market

Alibaba Cloud hikes prices by up to 34 percent, blames hardware costs and AI demand

Alibaba Cloud has announced price increases for many services, reaching up to 34 percent. The main reason is the increase in hardware costs and the growing demand for artificial intelligence resources. The increase impacts compute, storage, and SaaS services, including those based on proprietary silicio.

→

Mar 18 2026

Market

Water company wasted $200k on bad AI answers, builds filtering system

A water company, after spending $200,000 on unsatisfactory answers from an AI model, developed its own "filtering" system called 'Rozum' to orchestrate multiple models and obtain more reliable results. The article highlights how the prioritization of popularity over quality by some tech companies creates opportunities for alternative solutions.

→

Mar 18 2026

LLM

MiniMax-M2.7 Announced: What We Know

MiniMax has announced its new M2.7 model. The announcement was made via a post on a Chinese channel. Further details on the technical specifications and performance of the model are expected.

→

Mar 18 2026

Market

Tencent joins OpenClaw as sponsor after copy dispute, aligns with OpenAI and Baidu

Tencent has joined OpenClaw as a sponsor following a code copying dispute. This strategic move indicates a rapprochement of the Chinese technology giant with OpenAI and Baidu in the field of artificial intelligence.

→

Mar 18 2026

Market

Foxconn and SAP forge strategic alliance to drive enterprise AI in APAC

Foxconn and SAP have formed a strategic partnership to promote the adoption of artificial intelligence in enterprises in the APAC region. The goal is to provide integrated solutions that leverage the capabilities of both companies to accelerate the digital transformation of businesses.

→

Mar 18 2026

Hardware

Nvidia charts optical-copper path for future interconnects

According to Digitimes, Nvidia is exploring a future where interconnects between hardware components leverage both copper and fiber optics. This strategy could have significant implications for high-performance computing architectures and future data centers.

→

Mar 18 2026

Market

IBM completes Confluent acquisition to power real-time AI data

IBM has completed the acquisition of Confluent, a strategic move to strengthen real-time data processing capabilities for artificial intelligence applications. The integration aims to provide more efficient and immediate solutions for data analysis and utilization.

→

Mar 18 2026

Altro

GTC 2026: Memory, power, and deployment challenges reshape AI infrastructure

According to DIGITIMES, the increasing memory demands, power constraints, and deployment complexities are reshaping AI infrastructures. At GTC 2026, these factors will be central to the development and adoption of new solutions.

→

Mar 18 2026

LLM

MedArena: Comparing LLMs for Medicine-in-the-Wild Clinician Preferences

MedArena is an interactive platform for evaluating large language models (LLMs) in the medical field. It allows clinicians to directly compare the responses of different models using their own medical queries. Initial results, based on preferences collected up to November 2025, indicate that Gemini 2.0 Flash Thinking, Gemini 2.5 Pro, and GPT-4o are among the preferred models.

→

Mar 18 2026

LLM

SRLM: Recursive Language Models Meet Uncertainty

A new study introduces SRLM, a framework that enhances Recursive Language Models (RLM) with uncertainty-aware self-reflection. SRLM evaluates and compares different context-interaction programs, outperforming traditional RLM models, especially in semantically intensive contexts.

→

Mar 18 2026

Frameworks

XLinear: Frequency-Enhanced MLP with CrossFilter for Robust Long-Range Forecasting

A new study introduces XLinear, an MLP (Multi-Layer Perceptron)-based forecasting model designed to overcome the limitations of traditional models in long-range time series analysis. XLinear combines frequency attention and a CrossFilter block to improve the robustness and accuracy of forecasts, while maintaining a lightweight architecture.

→

Mar 18 2026

LLM

EHR: Tokenization Impacts Performance and Costs of Foundation Models

Tokenization, the conversion of healthcare data into inputs for deep learning models, significantly impacts performance and computational efficiency. A study explores different tokenization strategies on pediatric EHR data, evaluating predictive accuracy and pre-training costs. Joint event encoding and positional time encoding emerge as the most effective.

→

Mar 18 2026

Frameworks

AIDABench: A New Benchmark for AI-Driven Data Analytics

AIDABench, a comprehensive benchmark for evaluating the performance of AI systems in complex data analytics tasks, has been introduced. The benchmark includes over 600 diverse tasks across three core areas: question answering, data visualization, and file generation, based on realistic scenarios and heterogeneous data.

→

Mar 18 2026

Frameworks

HYQNET: Neural-Symbolic Logic Query Answering in Non-Euclidean Space

HYQNET is a neural-symbolic model leveraging hyperbolic spaces to answer complex logic queries on knowledge graphs. It combines the interpretability of symbolic methods with the generalization capabilities of neural networks, overcoming the limitations of traditional models in processing incomplete graphs and logical hierarchies.

→

Mar 18 2026

Market

In-depth: China's AI 'five-layer cake'— impressive on the outside, underbaked in the middle

An analysis of China's AI infrastructure, revealing a complex 'five-layer cake' structure. Despite the solid appearance, weaknesses emerge in the middle layers, raising questions about the real capacity for innovation and the long-term sustainability of the sector.

→

Mar 18 2026

Market

Taiwan launches 6G industry forum

Taiwan has launched a forum dedicated to the 6G industry, marking a step forward in the development of future communication networks. The initiative aims to position the island as a technological leader in the sector, promoting collaboration between companies and research institutions.

→

Mar 18 2026

Market

OpenAI shifts from building data centers to leasing cloud capacity

OpenAI is shifting its infrastructure strategy, moving from building proprietary data centers to leasing cloud computing capacity. The move signals a potential shift in priorities towards greater flexibility and scalability.

→

Mar 18 2026

Market

Taiwan smart city expo spotlights AI robotics across industries

The Taiwan Smart City Expo highlighted the latest innovations in artificial intelligence and robotics, with applications across various industries. The event provided an overview of the most advanced solutions for the smart cities of the future.

→

Mar 18 2026

Market

Syscom presents AI as a new operational workforce, promising measurable business change

Syscom presents artificial intelligence as a new operational resource, promising measurable business changes. The company focuses on AI to transform business processes and improve efficiency, offering concrete solutions for the implementation of advanced models.

→

Mar 18 2026

Market

Nvidia H200: Approvals Secured in China, Production Restarted

Nvidia secures approvals and orders for the H200 GPU in China, restarting supply chain production. This development indicates strong demand for high-performance computing solutions in the Chinese market, despite regulatory restrictions.

→

Mar 18 2026

Market

Alibaba's Wukong: Enterprise AI Agents Integrated into DingTalk

Alibaba launches Wukong, an AI-native platform to coordinate enterprise AI agents within business workflows, integrated with DingTalk. Wukong enters a competitive market, with Tencent and ByteDance already active in the field. The platform aims to progressively connect with Slack, Microsoft Teams, and WeChat, offering solutions for various verticals. Its integration with DingTalk provides an immediate distribution channel, but competition is fierce.

→

Mar 18 2026

Altro

Justice Department: Anthropic Not Trusted With Warfighting Systems

The Justice Department contested Anthropic's lawsuit, stating it lawfully penalized the company for trying to limit the use of its Claude AI models in military applications. The decision raises questions about the reliability of AI models in sensitive contexts.

→

Mar 18 2026

Market

Agnit Semiconductors raises US$2.6 million for GaN chips

Agnit Semiconductors, an Indian company, has raised US$2.6 million to commercialize gallium nitride (GaN) chips. These chips are intended for the telecommunications and power management sectors, promising greater efficiency and higher performance compared to traditional silicio solutions.

→

Mar 18 2026

Hardware

Dreame rollout plan for five chips: robotics today, AI data centers in orbit tomorrow

Manufacturer Dreame has announced an ambitious plan to develop five chips. The company, known for robotics solutions, is now aiming to expand its reach towards AI data centers, with a long-term vision that includes applications in orbit.

→

Mar 18 2026

Hardware

Chinese server maker Sugon unveils 400G AI fabric to rival Nvidia InfiniBand

Chinese server maker Sugon has unveiled a new interconnect solution for AI data centers, with speeds up to 400G. This technology aims to compete with Nvidia's InfiniBand solutions, offering an alternative for high-performance computing infrastructures.

→

Mar 18 2026

Market

AI data center optics boom lifts Ezconn; orders seen through 3Q26

Strong demand for optical solutions for AI data centers is benefiting Ezconn, with orders projected through the third quarter of 2026 and profit margins exceeding 50%. This reflects the increasing investments in artificial intelligence infrastructure.

→

Mar 18 2026

Market

UK's space shake-up puts Taiwan in the orbit of Britain's ambitions

According to Digitimes, the UK's space sector shake-up could open new collaboration opportunities with Taiwan. This strategic move aims to strengthen the UK's space ambitions, leveraging Taiwanese technological capabilities.

→

Mar 18 2026

Market

Analysis: Nvidia leveraging DeepSeek R1 to cement hardware–model domination

According to DIGITIMES, Nvidia is capitalizing on the success of the DeepSeek R1 language model to strengthen its dominant position in both the hardware and AI model markets. This synergy allows Nvidia to optimize its GPUs for specific workloads, offering performance and cost advantages to customers.

→

Mar 17 2026

LLM

MiniMax M2.7: Multimodal Model on the Horizon?

Rumors on Reddit suggest that MiniMax M2.7 might be a multimodal model. The company is exploring systems that integrate different input modalities, opening new possibilities for artificial intelligence applications. It remains to be seen whether the model will be available for on-premise deployment.

→

Mar 17 2026

LLM

GLM 5: A Surprising Competitor to Claude Code in Development?

A seasoned Claude Code user tested GLM 5 (OpenCode with Zen plan) on development tasks, including a real-time chat application with web sockets. Surprisingly, GLM 5 outperformed Claude Code in some scenarios, sparking interest in the community to further evaluate the model's capabilities.

→

Mar 17 2026

LLM

New open-source LLM releases: Skyfall, Valkyrie, and Anubis

Four new open-source language models developed by TheLocalDrummer have been quietly released: Skyfall 31B v4.1, Valkyrie 49B v2.1, Anubis 70B v1.2, and Anubis Mini 8B v1 (based on Llama 3.3). These models represent significant upgrades over previous versions and have received positive feedback from the community.

→

Mar 17 2026

Frameworks

Hugging Face simplifies local LLM deployment with one-liner

Hugging Face has released a tool that, with a single command, automates hardware detection, optimal model and quantization selection, `llama.cpp` server startup, and the launch of Pi, the agent behind OpenClaw. This significantly simplifies the local deployment process for large language models.

→

Mar 17 2026

Frameworks

Unsloth Studio: New open-source web UI to train and run LLMs

Unsloth Studio is a new open-source web UI that allows training and running large language models (LLMs) locally. It supports various operating systems, model formats, and offers tools for model optimization and comparison.

→

Mar 17 2026

Frameworks

Unsloth Studio: A competitor to LM Studio for local LLMs?

Unsloth announced Unsloth Studio, an Apache-licensed runner compatible with Llama.cpp. This could be a game changer for LLM users operating locally, offering an alternative to LM Studio in the GGUF ecosystem.

→

Mar 17 2026

Altro

World ID: Unique Identity for AI Agents Against Sybil Attacks

World ID, developed by World (formerly known for WorldCoin), proposes a system to uniquely identify users behind AI agents. The goal is to mitigate Sybil attacks, where a large number of automated agents overload online services. The solution is based on iris scanning to create a unique digital identity.

→

Mar 17 2026

Market

Mistral bets on ‘build-your-own AI’ to challenge OpenAI in enterprise

Mistral AI launches Forge, a platform that allows enterprises to train custom AI models from scratch on their own data. This approach challenges competitors that rely on fine-tuning and retrieval-based methods.

→

Mar 17 2026

Frameworks

Garry Tan's Claude Code setup sparks mixed reactions

Garry Tan's Claude Code setup, shared on GitHub, has sparked widespread debate. Numerous users are testing the setup, expressing divergent opinions, including those of language models like Claude, ChatGPT, and Gemini.

→

Mar 17 2026

Hardware

Nvidia updates data center roadmap with Rosa CPU and stacked Feynman GPUs — optical NVLink, Groq LPUs with NVFP4, and NVLink also on deck

Nvidia updates its data center roadmap, introducing the Rosa CPU and Feynman GPUs. The company is focusing on optical NVLink and Groq LPUs with NVFP4, as well as new versions of NVLink. These innovations promise to significantly improve the performance and efficiency of AI workloads.

→

Mar 17 2026

Market

ChatGPT: surge in queries about salaries and compensation

New research shows that Americans send nearly 3 million daily messages to ChatGPT asking about compensation and earnings. This trend helps close the wage information gap, providing greater transparency for workers.

→

🗄️ News Archive