AI-RADAR.it
AI-RADAR.IT · AI-RADAR.NET · AI-RADAR.TECH
News & analysis on local LLMs, stack & on-prem hardware.

🗄️ News Archive

Complete history of AI signals, ordered by date.
Total Articles: 15981

This archive is the long-term memory of AI-Radar: model launches, framework releases, infrastructure shifts, and market signals tracked over time in one searchable timeline. Use it to compare how narratives evolved, identify which technologies sustained momentum, and validate decisions with historical context rather than short-lived hype. For faster navigation, jump to focused hubs like LLM, Frameworks, Hardware, or the Trends pillar.

💡 Looking for something specific? Use the Search Bar at the top for a detailed search.

Feb 22 2026
Frameworks

nanollama: Train Llama 3 from scratch and export to GGUF

NanoLLama, an open-source framework for training Llama 3 models from scratch, without fine-tuning or LoRA, has been released. The tool allows exporting to GGUF format compatible with llama.cpp via a single command. It includes configurations from 46M to 7B parameters, multi-corpus training, and a Go inference engine.

→
Feb 22 2026
General

A Descent into "Vibes" (2021–2026)

The Era of Probabilistic Toddlers: How "Vibe Coding" Broke Software Engineering and How We Might Just Fix It

→
Feb 22 2026
Frameworks

Kon: A compact coding agent for local LLMs

A developer introduced Kon, a coding agent designed to be lightweight and easily understandable. Kon is intended to run locally, with a small token footprint and a limited number of files, making it easy to customize and extend.

→
Feb 22 2026
LLM

ByteDance Launches SeaDance 2.0: Hollywood Threatens Legal Action

ByteDance has released SeaDance 2.0, a video generation model that accepts multiple inputs (text, images, video, audio) and produces 2K clips with synchronized audio. Hollywood has reacted with threats of legal action, with Disney and Paramount sending cease and desist letters. The model is not open source.

→
Feb 22 2026
Hardware

MIT: 3D printer outputs fully functional electric motors in one step

MIT researchers have developed a 3D printer capable of producing fully functional linear electric motors in a single process. Only magnetization is required after printing. The estimated cost per motor is around 50 cents.

→
Feb 22 2026
Market

AOMedia Open Audio Codec "OAC" Aims To Be The Successor To Opus

The Alliance For Open Media (AOMedia), known for the AV1 video codec and AVIF format, is developing a new open-source audio codec called Open Audio Codec (OAC). The goal is to create a next-generation successor to Opus, optimized for modern audio compression and quality needs.

→
Feb 22 2026
LLM

Qwen team confirms data quality issues in GPQA and HLE datasets

The Qwen team has verified serious data quality issues in the GPQA and HLE (Humanity's Last Exam) test sets. In-depth analysis revealed that many answers considered "gold standard" were incorrect, compromising the reliability of the benchmarks. The discovery was initially raised by independent researchers.

→
Feb 22 2026
Altro

Sam Altman fires back at Elon Musk's space data centers proposal: 'Ridiculous'

Sam Altman, OpenAI's CEO, has expressed skepticism about Elon Musk's proposal to build data centers in space. Altman cited high failure rates and prohibitive costs as the main obstacles to the feasibility of such a project, at least for now. The discussion raises questions about the future infrastructure for artificial intelligence.

→
Feb 22 2026
Market

Sam Altman (OpenAI) defends AI energy consumption

Sam Altman, CEO of OpenAI, has responded to criticism regarding the high energy consumption of artificial intelligence, comparing it to the resources required for human development and education. He highlighted how training an AI model, while expensive, can be compared to the twenty years of learning and food needed for a human to achieve intelligence.

→
Feb 22 2026
Frameworks

OpenClaw: are skills more important than the runner itself?

A LocalLLaMA user questions the hype around OpenClaw, an LLM framework. While acknowledging its usefulness in loops, memory management, agents, and integrations, the user emphasizes that the developed or integrated skills are the real added value, more than the runner itself. The "opencode web" alternative is considered superior.

→
Feb 22 2026
LLM

Local LLMs: Growing Anticipation for 9B and 35B Parameter Models

The open-source community focused on running large language models (LLMs) locally, through the LocalLLaMA initiative, is actively discussing expectations for upcoming 9 and 35 billion parameter models. The focus is on optimizing performance and efficiency in using limited hardware resources, typical of on-premise environments.

→
Feb 22 2026
Hardware

Dell adopts custom connector for Nvidia GPUs to prevent overheating

Dell introduces a custom 16-pin power connector for Nvidia GPUs in its pre-built PCs, addressing overheating and melting issues found with standard connectors. The solution aims to ensure a more stable and secure connection, reducing the risk of malfunctions.

→
Feb 22 2026
Market

Intel automates customer support with AI-powered assistant

Intel reduces phone support and introduces "Ask Intel", a customer support system powered by artificial intelligence and developed with Microsoft Copilot Studio. The goal is to improve efficiency and response times to user requests, leveraging the natural language understanding and automation capabilities offered by AI.

→
Feb 22 2026
Hardware

Yet Another Fix Coming For Older AMD GPUs On Linux

Timur Kristóf of Valve's open-source Linux graphics driver team continues to enhance the support for older AMD Radeon GPUs on Linux. The upcoming improvements promise to further optimize performance and stability for users with aging AMD hardware.

→
Feb 22 2026
Market

Retail DDR5 memory prices slowly drop in Europe: is the correction beginning?

Despite ongoing shortages, DDR5 memory prices are slowly decreasing in Europe. This long-overdue price correction may finally be starting in some regions, offering potential relief to consumers and businesses.

→
Feb 22 2026
Hardware

Linux 7.0 Further Prepares For Intel Diamond Rapids With NTB Driver Support

The Linux kernel is preparing for Intel's next-generation Xeon Diamond Rapids processors as the successor to Granite Rapids. Most of the driver support appears to be settled, with the addition of NTB driver support in the Linux 7.0 kernel.

→
Feb 22 2026
Market

How to Hide Google’s AI Overviews From Your Search Results

You can avoid Google’s AI summaries in your search results by simply adjusting your query. Or just switch search engines altogether.

→
Feb 22 2026
Altro

Microsoft Hyper-V Lands Some Useful Improvements In Linux 7.0

The Linux 7.0 kernel introduces several improvements for those using Microsoft Hyper-V for virtualization. This work follows similar improvements made to KVM, solidifying Linux's role as a versatile platform for virtualized environments.

→
Feb 22 2026
LLM

FlashLM v5: Language Model Trained on CPU Beats GPU Baseline

FlashLM v5, a language model with 29.7 million parameters, was trained on an AMD Ryzen 7950X3D CPU in approximately 40 hours. The model achieved a perplexity of 1.36, surpassing the TinyStories-1M baseline (PPL 1.59). The ParallelGatedRecurrence architecture utilizes ternary weights and does not require matrix multiplications in the forward pass.

→
Feb 22 2026
Altro

Linux Kernel 7.0 Prepares for Rust 1.95

Linux kernel 7.0 is preparing to integrate updates to the Rust programming language, marking a turning point with the stabilization of Rust support for kernel and driver programming. The update includes fixes and preparations for the Rust 1.95 release.

→
Feb 22 2026
Market

US-China Tariffs: Geopolitical and Strategic Implications

The US decision to revise tariffs imposed on China has significant repercussions on global trade dynamics and relations with Asian allies. We analyze the consequences of this move and its long-term implications for the international economic landscape.

→
Feb 22 2026
LLM

Local LLM: Niche Use Cases Emerge Online

An online discussion reveals unexpected uses for large language models running locally. From generating specific prompts to analyzing sensitive data, users are exploring the potential of on-premise LLMs for specialized applications, often constrained by privacy requirements or prohibitive cloud costs.

→
Feb 22 2026
Market

India advances OSAT project as Tata wafer fab faces timeline uncertainty

India is accelerating its OSAT (Outsourced Semiconductor Assembly and Test) project. Simultaneously, uncertainties are emerging regarding the timeline of Tata's wafer fabrication plant, a key element for self-sufficiency in the semiconductor sector.

→
Feb 22 2026
Market

Trump vows to raise global tariff to 15% after Supreme Court ruling

Former President Donald Trump has pledged to impose 15% tariffs on all goods imported into the United States, following a Supreme Court ruling. The proposal aims to protect domestic industry and reduce the trade deficit, but raises concerns about the impact on inflation and international trade relations.

→
Feb 21 2026
Altro

Security Alert: Vulnerability in Cline Release

A recent Cline release was compromised with the injection of a malicious installer, OpenClaw. The supply chain attack has exposed an unknown number of installations, raising concerns about the release speed and security of public agentic tools. It is recommended to disable automatic updates for VSCode extensions.

→
Feb 21 2026
Market

Sam Altman would like remind you that humans use a lot of energy, too

Sam Altman, CEO of OpenAI, points out that training a human also requires a significant amount of energy. The comment is part of the debate on the environmental impact of artificial intelligence and large language models.

→
Feb 21 2026
Market

CXMT offers DDR4 memory chips at half the market price

Chinese manufacturer CXMT is reportedly offering DDR4 memory chips at prices significantly lower than the market average. This move could have notable implications for the cost of building on-premise systems, especially for those assembling dedicated servers for LLM inference.

→
Feb 21 2026
Market

PSA: Software "Shade" is a plagiarized copy of Heretic

A security alert warns against the software "Shade", accused of being a fraudulent and plagiarized copy of the open-source project Heretic. The developer of Heretic denounces the removal of the commit history and the names of the original contributors.

→
Feb 21 2026
LLM

Wave Field LLM: O(n log n) attention via wave equation dynamics

A novel attention mechanism for LLMs, Wave Field LLM, uses wave equations to scale at O(n log n). The model maps tokens onto a continuous 1D field and propagates information via damped wave equations. Initial results on WikiText-2 show competitive performance compared to standard transformers, with increasing advantages for longer sequences.

→
Feb 21 2026
Hardware

Intel Bartlett Lake-S CPUs: 12 P-cores and 5.8 GHz Boost, Not for Retail

Leaks suggest a new Intel CPU, the Bartlett Lake-S, featuring 12 P-cores and boost frequencies up to 5.8 GHz. However, this high-performance CPU is reportedly not destined for the retail market, limiting its availability to end-users.

→
Feb 21 2026
Frameworks

Qwen Code: Open-Source Coding Agent with No-Telemetry Fork

Qwen Code is an open-source CLI coding agent developed by Alibaba's Qwen team. It automates development tasks by directly interacting with the code. A modified version is available that removes telemetry, ensuring greater privacy. Integration with LM Studio and Qwen3-Coder enables fully local and offline use.

→
Feb 21 2026
Market

Google VP warns that two types of AI startups may not survive

A Google VP warns that AI startups focusing solely on LLM wrappers or AI service aggregation face difficulties. Shrinking margins and limited differentiation threaten their long-term viability.

→
Feb 21 2026
LLM

Ouro-2.6B-Thinking: First Working Inference for ByteDance's Model

Inference issues with ByteDance's Ouro-2.6B-Thinking, a recurrent Universal Transformer model, have been resolved. The fix addresses incompatibilities with Transformers 4.55. The outputs now produce valid results. Tested on NVIDIA L4, achieving 3.8 tokens/s and using 5.3 GB of VRAM.

→
Feb 21 2026
LLM

The importance of key figures in open source LLM innovation

A Reddit post highlights the potential impact of prominent figures like Andrej Karpathy in the development of open source large language models (LLMs). The discussion underscores how the presence of experts can significantly accelerate progress and competitiveness in this field.

→
Feb 21 2026
Hardware

Corsair Makr Pro 75 Review: Not DIY-priced (or DIY)

Review of the Corsair Makr Pro 75 ATX case. The original article highlights that the product is not particularly cheap and doesn't perfectly suit DIY configurations, despite offering interesting features for PC enthusiasts.

→
Feb 21 2026
Altro

Data center developers building private natural gas 'Shadow Grid' power plants

Data center developers are building private natural gas power plants to power their infrastructures, bypassing strained power grids. A project in Texas will reportedly use as much power as the city of Chicago.

→
Feb 21 2026
Market

India and US join forces on semiconductor supply chain

India joins the US-led Pax Silica initiative to strengthen the semiconductor supply chain, reducing reliance on China. The agreement covers rare earth elements and chip manufacturing tools, aiming to develop expertise in the sector.

→
Feb 21 2026
LLM

GLM-4.7: Distilled Model for Advanced Reasoning Locally

A distilled model named GLM-4.7, designed to offer advanced reasoning capabilities, is available on Hugging Face. This version, mentioned by Unsloth, aims to provide high performance in local usage contexts. The model is available in GGUF format, facilitating its implementation on various hardware platforms.

→
Feb 21 2026
Frameworks

Ollama 0.17 Released With Improved OpenClaw Onboarding

The new ollama v0.17 release, the open-source project that makes it easy to get up and running with a variety of LLMs under Windows, macOS, and Linux, is driven by new functionality around enhancing the OpenClaw onboarding process. The update aims to simplify the initial setup for users.

→
Feb 21 2026
Hardware

Linux 7.0: AMDGPU Fixes for Older Radeon Hardware

Linux 7.0 introduces a series of fixes for DRM graphics drivers, with a focus on enhancements for older Radeon GPUs via the AMDGPU driver. These updates aim to optimize performance and stability on older hardware.

→
Feb 21 2026
Frameworks

AMD AOMP 23.0-0 Compiler Continues Enhancing Fortran Support

AMD AOMP 23.0-0 has been released as the latest build of this LLVM/Clang downstream. It includes AMD patches focused on delivering GPU offloading support for Radeon/Instinct hardware with OpenMP and OpenACC APIs. AOMP serves as a leading-edge compiler until the patches are integrated into upstream LLVM.

→
Feb 21 2026
Hardware

Apple iBook: Legacy Updates Still Available After 27 Years

Even 27 years after its launch, Apple's servers continue to provide legacy updates for the iBook lineup. A 21-year-old iBook G4 seamlessly connects to Wi-Fi and downloads updates without modifications, demonstrating Apple's commitment to backward compatibility.

→
Feb 21 2026
Frameworks

Interactive map of Australian High Court cases (1901-2025)

A developer has created an interactive visualization of all Australian High Court cases from 1901 to 2025, transforming case law into an interconnected knowledge graph. The project uses vector embedding and dimensionality reduction techniques to map judgments semantically, revealing relationships between areas of law and the historical evolution of legal citations. The code to reproduce the analysis is available on GitHub.

→
Feb 21 2026
LLM

GLM-5: "Claude" Personality and Censorship Bypass?

A user discovered that GLM-5, a large language model, significantly changes its behavior when told it is Claude from Anthropic. This personality shift also appears to bypass some built-in censorship. It remains unclear whether this behavior is intentional or an emergent property.

→
Feb 21 2026
Market

OpenAI projects US$280B revenue by 2030, plans US$600B in spending

OpenAI is projected to reach US$280 billion in revenue by 2030. The company plans to invest US$600 billion. These figures highlight the company's growth ambitions in the artificial intelligence market and the need for adequate infrastructure to support such expansion.

→
Feb 21 2026
Hardware

ChatJimmy: 15,000+ tok/s on dedicated silicio – the "Model-on-Silicio" era?

ChatJimmy.ai announced achieving 15,000 tokens per second using a dedicated ASIC, etching model weights directly into silicio. This approach bypasses HBM and VRAM bottlenecks, sparking debate about the efficiency of general-purpose hardware for inference.

→
Feb 21 2026
Market

Samsung to reclaim world's no.1 semiconductor position in 2026

According to DIGITIMES, Samsung aims to reclaim the top spot in the semiconductor market by 2026. The South Korean company plans to surpass its competitors by investing heavily in new technologies and production capabilities. Competition in the sector remains high, with companies like TSMC and Intel continuing to innovate.

→
Feb 21 2026
Market

Sarvam launches Indus AI chat app as competition heats up

Indian startup Sarvam AI has launched Indus, a new chat application powered by artificial intelligence. Currently available in beta, Indus enters a rapidly expanding market where competition among AI solution providers is increasingly intense.

→
Feb 21 2026
Altro

AI supply chain tracker: Rack infrastructure joins the AI buildout

The increasing demand for computing power for artificial intelligence is driving the evolution of rack infrastructures. Data centers are preparing to support increasingly intensive workloads, with a focus on energy efficiency and compute density.

→
Feb 21 2026
Market

Glass fiber shortage intensifies capacity battle, pushing PCB prices higher

The shortage of glass fiber is intensifying the competition for production capacity, leading to higher prices for PCBs (Printed Circuit Boards) and reshaping supply chains. The situation is putting pressure on electronics manufacturers.

→
Feb 20 2026
Market

Creator revenue diversification: beyond advertising

The creator economy is evolving rapidly, and ad revenue is no longer sufficient. Many YouTubers are launching product lines, acquiring startups, and building actual business empires. Diversification has become the new strategy for many.

→
Feb 20 2026
LLM

New version coming soon for Gemma, Google's LLM

Google has announced the upcoming release of a new version of Gemma, its large language model (LLM). The news emerged from a Reddit post, reported by the LocalLLaMA community, which links to a YouTube video.

→
Feb 20 2026
LLM

AI Model Attempts High-Level Math Challenges

An artificial intelligence model tackles the First Proof math challenge, a competition testing reasoning capabilities on complex problems. The initiative aims to evaluate the performance of AI models in scenarios requiring expert-level skills.

→
Feb 20 2026
Market

Anthropic tightens rules on the use of third-party harnesses with Claude

Anthropic has revised its legal terms to clarify its policy forbidding the use of third-party harnesses with Claude subscriptions. The aim is to protect the company's revenue model and ensure greater control over the user experience.

→
Feb 20 2026
Market

Pro and Anti-AI Super PACs Clash Over New York Congressional Seat

Two political action committees (PACs) focused on artificial intelligence are vying for a congressional seat in New York. At the heart of the dispute is Alex Bores, a supporter of a law requiring AI developers to disclose safety protocols and report system misuse.

→
Feb 20 2026
Altro

Making frontier cybersecurity capabilities available to defenders

Making frontier cybersecurity capabilities available to defenders. The goal is to provide more effective tools to counter emerging threats and protect critical infrastructures.

→
Feb 20 2026
Hardware

Cerebras plans humongous AI supercomputer in India backed by UAE

Cerebras Systems, known for its dinner plate-sized AI accelerators, is planning a new supercomputing cluster in India. The computing power will reach 8 exaFLOPS, opening new frontiers for artificial intelligence in the country.

→
Feb 20 2026
Market

InScope nabs $14.5M to solve the pain of financial reporting

InScope, a startup founded by accountants from companies like Flexport and Miro, has raised $14.5 million in funding. The goal is to automate the complex process of preparing financial statements, simplifying a crucial task for businesses.

→
Feb 20 2026
Hardware

Taalas: LLMs baked into hardware, up to 16,000 tokens/second

Startup Taalas takes a radical approach: baking LLM models and their weights directly into a silicio chip. This achieves sub-millisecond latencies and 10x power efficiency, eliminating the need for HBM and advanced packaging. LoRA fine-tuning is supported. The first demonstrator uses Llama 3.1 8B.

→
Feb 20 2026
Hardware

AI craze leaves only one Nvidia RTX 50-series GPU at MSRP — RTX 5060 Ti 8GB is the final stand, as even the RTX 5050 falls

High demand for Nvidia GPUs dedicated to artificial intelligence is impacting market prices. The RTX 5060 Ti 8GB remains the only card in the 50 series still available at MSRP, while even the RTX 5050 has seen increases.

→
Feb 20 2026
Hardware

The great Bench GPU retest begins — what to expect for 2026

A preview of the benchmark tests for next-generation GPUs, including RTX 50 and RX 9000 series. The article focuses on testing methodologies and performance expectations, excluding upscaling and frame generation.

→
Feb 20 2026
LLM

Chinese models dominate OpenRouter: exceeding 3 trillion tokens

The OpenRouter platform is experiencing a surge in the use of language models of Chinese origin. For the first time, a model exceeds 3 trillion tokens processed in a week, and multiple models exceed one trillion, marking a shift from the dominance of US models.

→
Feb 20 2026
Altro

Amazon's vibe-coding tool Kiro reportedly vibed too hard and brought down AWS

An incident at AWS was reportedly caused by Kiro, Amazon's agentic AI coding tool. The company attributes the outage to human error, specifically misconfigured access controls. The episode raises questions about the management of AI tools.

→
Feb 20 2026
Altro

The U.S. Military Is Reviving Microbes from 40,000-Year-Old Ice

U.S. military scientists have revived microbes frozen in Alaskan permafrost for nearly 40,000 years. The study, which led to the discovery of 26 new species, aims to develop technologies to protect soldiers in extremely cold environments, with applications ranging from frostbite prevention to the development of new antifreeze and de-icing techniques.

→
Feb 20 2026
Market

Berlin: private premiere of 'Holiguards Saga'

A private premiere of 'Holiguards Saga – The Portal of Force', the pilot episode of a planned saga, was held in Berlin. The closed-format event was attended by industry insiders, partners, and media representatives.

→
Feb 20 2026
Market

U.S. Supreme Court blocks Trump's tariffs: tech manufacturers cheer

The United States Supreme Court has blocked tariffs imposed by the Trump administration. The Consumer Technology Association welcomed the decision, calling it a victory for consumers and demanding swift refunds for retailers.

→
Feb 20 2026
Market

Anthropic: AI Safety vs. Major Military Contracts?

Anthropic is restricting the use of its AI in autonomous weapons and government surveillance. These ethical constraints could cost the company major military contracts, raising questions about balancing principles and market opportunities.

→
Feb 20 2026
Market

Toy Story 5 takes aim at creepy AI toys: ‘I’m always listening’

The new Toy Story movie will address the theme of AI-enabled toys and the privacy risks for children. The movie is scheduled for release on June 19 and will feature Woody facing new challenges.

→
Feb 20 2026
Market

AI’s promise to indie filmmakers: faster, cheaper, lonelier

AI expands access to filmmaking for resource-constrained creators. But as efficiency becomes the industry’s north star, creativity risks being overwhelmed by a deluge of low-effort, AI-generated content.

→
Feb 20 2026
Market

OpenAI's Sam Altman warns against 'AI washing'

OpenAI CEO Sam Altman warns that some companies are misusing artificial intelligence to mask layoffs. Altman expressed concern about the consequences for the world of work during an event in India.

→
Feb 20 2026
Market

Peak XV raises $1.3B, doubles down on AI in India

Peak XV Partners announced a new $1.3 billion fund, primarily targeting the Indian market. The firm intends to focus on investments in artificial intelligence, fintech, and cross-border ventures, amid increasing competition among global venture capital firms in the country.

→
Feb 20 2026
Market

Zurich’s Rapidata raises €7.2M to build a real-time human feedback network for AI

Zurich-based startup Rapidata has raised €7.2 million to build a real-time human feedback network for AI. The aim is to provide machines with a more effective learning experience through the integration of human input.

→
Feb 20 2026
Frameworks

Hugging Face acquires GGML and llama.cpp for Local AI advancement

Hugging Face announced the acquisition of GGML and llama.cpp, two open-source projects crucial for efficient execution of large language models (LLMs) on consumer hardware. The goal is to ensure the long-term development of local AI and democratize access to these technologies.

→
Feb 20 2026
Market

Meta cuts employee bonuses despite AI investments

Mark Zuckerberg continues to cut Meta employee bonuses, with a 5% reduction this year, following a 10% cut last year. The decision comes despite significant investments in the artificial intelligence sector, including a $130 billion spending plan and high compensation packages for AI professionals.

→
Feb 20 2026
Market

Swiss CLIMATEX secures €3.5M for circular textile technologies

Swiss company CLIMATEX AG, specializing in circular textile technologies, has closed a €3.5 million funding round led by the Collateral Good Textile & Fashion Innovation Fund. The funds will be used to strengthen sales and marketing activities and to support the company's growth phase.

→
Feb 20 2026
Frameworks

Hugging Face Acquires GGML.AI, Focused on Efficient LLM Inference

Hugging Face has acquired GGML.AI, known for its work on efficient inference of large language models (LLMs). The acquisition, discussed on Reddit and GitHub, could lead to greater integration of GGML technologies into the Hugging Face ecosystem, benefiting developers working with LLMs in resource-constrained environments.

→
Feb 20 2026
LLM

Deepseek and Gemma: comparison in the LocalLLaMA community

A Reddit post in the LocalLLaMA community compares Deepseek and Gemma models. The discussion revolves around the characteristics and performance of these models, with a focus on local usage. The original article includes an image, presumably comparative, of the two models.

→
Feb 20 2026
Market

ChatGPT: Usage Booms Among Young Indians

OpenAI reveals that users between 18 and 24 years old generate nearly 50% of the messages sent to ChatGPT in India. The under-30 age group accounts for 80% of the total usage in the country, highlighting strong chatbot adoption among young people.

→
Feb 20 2026
Market

Berlin aims to become Europe's most builder-friendly tech city

A community-led initiative called Berlin auf die Eins (BAD1) aims to make Berlin the most startup-friendly city in Europe within the next two years. The initiative aims to reduce bureaucratic friction, unite ecosystem stakeholders, and promote a more accurate narrative of Berlin's strengths as an economic engine and AI hub.

→
Feb 20 2026
LLM

EFF demands human documentation for LLM-generated code

The Electronic Frontier Foundation will accept LLM-generated code in its open source projects, but insists on human-written documentation and comments. The organization emphasizes the importance of clarity and understandability in code.

→
Feb 20 2026
LLM

Hallucination Benchmark: Kimi K2.5 outperforms Opus 4.6 in Pharma

A recent benchmark evaluated the hallucination capabilities of several large language models (LLMs) in the pharmaceutical domain. Surprisingly, Kimi K2.5 outperformed Opus 4.6 in this specific test. The dataset used is available on Hugging Face, offering transparency and reproducibility.

→
Feb 20 2026
Altro

UALink roadmap plots course to optimized AI data center interconnects

UALink, presented at Hot Chips 2024, is an open standard for interconnects in AI data centers. It aims to reduce vendor lock-in while optimizing costs and performance. The initiative promises greater flexibility and control in the infrastructure for artificial intelligence workloads.

→
Feb 20 2026
Hardware

High-end gaming PC given away: RTX 3090 and i9-10850K inside

An inexperienced user received a high-end gaming PC as a gift, equipped with an RTX 3090 GPU, a liquid-cooled Intel i9-10850K processor, and an Asus Maximus motherboard. An unexpected opportunity to enter the world of high-performance gaming.

→
Feb 20 2026
LLM

Microsoft removes guide on how to train LLMs on pirated Harry Potter books

Microsoft removed a blog post that suggested using illegally downloaded Harry Potter books to train AI models. The guide, written by a senior product manager, promoted a new feature of Azure SQL DB, LangChain, and LLMs. The decision followed criticism on Hacker News.

→
Feb 20 2026
Altro

Could AI Data Centers Be Moved to Outer Space?

Massive data centers for generative AI are bad for the Earth. A radical proposal suggests launching them into orbit to reduce the ecological footprint. The costs and technical feasibility of such an approach remain to be assessed.

→
Feb 20 2026
Hardware

Corsair FRAME 4000D Vault: limited-edition cases with upgraded I/O

Corsair launches the FRAME 4000D Vault series, limited-edition PC cases featuring color-shifting paintjobs and an upgraded front panel I/O. Available for $140, these towers stand out for their unique aesthetics and advanced features.

→
Feb 20 2026
Altro

MaXon Systems develops autonomous air defense against drones

Ukrainian company MaXon Systems has secured funding to develop an autonomous air defense system against mass drone attacks. The system combines high-speed interceptors, long-range detection, and integrated guidance software to protect critical infrastructure.

→
Feb 20 2026
Frameworks

From Agile to AI: Test-Driven Development is Crucial

A workshop marking the 25th anniversary of the Agile Manifesto highlighted the impact of artificial intelligence on software development. Test-driven development emerged as a fundamental practice, while security remains a critical and often overlooked aspect.

→
Feb 20 2026
Altro

Cloud Hypervisor 51: Performance Improvements and Better QCOW2 v3 Support

Cloud Hypervisor 51 is now available for this Rust-based VMM focused on secure cloud computing. What began as an Intel open-source project years ago is continuing to be largely led by Microsoft, Cyberus Tech, Tencent, and Ant Group.

→
Feb 20 2026
Market

Pontiro secures £357,500 to expand healthcare AI infrastructure

Pontiro, a healthtech company focused on enabling the safe use of medical imaging data for research and AI, has raised £357,500 in funding. The investment will support the platform's expansion, which has already processed over 2 million medical images across NHS Wales, reducing turnaround times and compliance risks.

→
Feb 20 2026
Market

AI: Executives’ optimism about the future impact on productivity and hiring

A recent international study reveals that AI has produced modest increases in productivity and changes in employment. Executives expect these effects to accelerate over the next three years, with projected productivity gains of up to 2.25% in the United States. A slight reduction in headcount is expected, offset by the creation of new roles related to data governance and the development of AI-based services.

→
Feb 20 2026
Market

SK Group chairman Chey seeks AI chip suppliers in US

SK Group chairman Chey Tae-won met with US tech leaders to secure supplies of essential AI chips. The initiative aims to strengthen SK Hynix's supply chain, which is crucial for developing AI solutions.

→
Feb 20 2026
LLM

Kimi aims for context window expansion

A user reported on Reddit about Kimi's ambitions to expand the context window. Increasing the context window is a hot topic in LLM development, as it allows processing longer and more complex prompts, improving the quality of the outputs and opening new application possibilities.

→
Feb 20 2026
Altro

OpenClaw stunt highlights growing security risks in AI agents

A prompt-injection test showed how the OpenClaw AI agent could trick assistants into installing software. The incident raises security concerns about autonomous agents, capable of interacting with operating systems and development environments, and highlights the need for robust safeguards.

→
Feb 20 2026
LLM

SanityBoard: New LLM Models and Open Source Agents Compared

SanityBoard updates with new benchmark results for models like Qwen3.5 Plus, GLM 5, and Gemini 3.1 Pro, along with three new open source coding agents. The analysis highlights the importance of infrastructure and model characteristics (iteration) on performance.

→
Feb 20 2026
LLM

Luma v2.9: a compact LLM trainable locally

Luma v2.9, a small language model (around 10 million parameters) based on a transformer architecture, has been released. Its key feature is that it can be trained with custom data and run entirely locally, without cloud dependencies or telemetry. The goal is to provide a model that can be specialized for specific tasks, rather than being a generalist.

→
Feb 20 2026
Market

Taiwan surpasses China in US monthly imports for first time in decades

For the first time in decades, Taiwan has surpassed China as the leading source of imports to the United States. The news, reported by DIGITIMES, marks a significant shift in global trade balances and technology supply chains.

→
Feb 20 2026
Frameworks

PaddleOCR-VL now in llama.cpp

The open-source multilingual model PaddleOCR-VL has been integrated into llama.cpp. This integration allows running model inference directly on local hardware, opening new possibilities for OCR applications with privacy and data sovereignty requirements. The version is 0.9b, making it suitable for machines with limited resources.

→
Feb 20 2026
LLM

Google, Apple add music-focused generative AI features

Google and Apple are integrating music-focused generative AI features into their products. The goal is to provide users with new tools for music composition and production, leveraging the power of AI to simplify and expand creative possibilities.

→
Feb 20 2026
LLM

Evaluating LLMs for Greek QA: The DemosQA Benchmark

A new study introduces DemosQA, a dataset for Question Answering in Greek, built from social media user questions. The research evaluates 11 language models, both monolingual and multilingual, using different prompting strategies, aiming to bridge the gap in LLM research for lower-resource languages.

→
← Previous Page 124 / 160 Next →