AI-RADAR.it
AI-RADAR.it
News & analisi per LLM locali, stack e hardware on-prem.

AI-Radar – Independent observatory covering AI models, LLMs, local AI, hardware, and trends

AI-Radar for on-prem LLMs & Home AI

The daily radar on models, frameworks, and hardware to run AI locally. LLMs, LangChain, Chroma, mini-PCs, and everything you need for a distributed "in-house" brain.

⚙️ Stack: Local LLMs · LangChain · Transformers · ChromaDB · MiniPCs · AI boxes
🛰️ Ask Observatory (Q&A + RAG) connected to the article archive.
📡

The Daily Signal

Is a new Dot-Com (AI) crash possible?

Is a new Dot-Com (AI) crash possible?

The Great Money Bonfire of 2026

✍️ Editorial 2026-02-14

⚡ Trending Now

View All →
AI Applications in Business and Industry AI Infrastructure and Hardware AI Model Development and Releases Taiwan's Tech Industry Focus

Latest Analysis & Radar News

AI-generated articles from feeds, with space for human editorial layer above the raw content.

Un tool open source fa dibattere 5 IA per validare le risposte
📁 LLM AI generated ℹ️ LocalLLaMA

Open-source tool makes 5 AIs debate to validate answers

A developer has created an open-source platform that uses five large language models (LLMs) in a debate and cross-checking process. The goal is to reduce blind reliance on AI responses, promoting a more critical and validated approach. The code is available on GitHub for those who want to test and contribute.

2026-01-18 📰 Source
Personal-Guru: il tutor AI open source e locale che crea un curriculum
📁 LLM AI generated ℹ️ LocalLLaMA

Personal-Guru: Open-Source AI Tutor Builds Custom Curriculum Locally

Personal-Guru is an open-source learning system that automatically generates a structured curriculum from a topic. It runs locally, without subscriptions, offering privacy and offline capabilities. It includes quizzes, flashcards, and audio/video modes for interactive learning.

2026-01-18 📰 Source
MGC si espande nel mercato del raffreddamento a liquido per IA
📁 Market AI generated ✅ DigiTimes

MGC expands into AI liquid cooling market

MGC chairman James Chen has announced the company's expansion into the liquid cooling sector for artificial intelligence applications. This strategic move positions MGC in a rapidly growing market, driven by the increasing demand for efficient cooling solutions for high-performance AI systems.

2026-01-18 📰 Source
HCMF Group e Kinpo Electronics insieme per l'auto AI-first
📁 Market AI generated ✅ DigiTimes

HCMF Group and Kinpo Electronics team up for AI-driven auto industry

Taiwan's HCMF Group and Kinpo Electronics are collaborating to navigate the challenges of an increasingly AI-driven automotive industry. The initiative aims to combine their respective expertise to develop innovative and competitive solutions in the emerging market for intelligent and autonomous vehicles.

2026-01-18 📰 Source
Corea del Sud rafforza la protezione IP, ma persistono le fughe di dati verso la Cina
📁 Market AI generated ✅ DigiTimes

South Korea strengthens IP protection, but China-linked data leaks persist

South Korea has renewed an intellectual property agreement in an attempt to stem the ongoing leaks of technological information, particularly to China. Despite government efforts and new regulations, the theft of trade secrets remains a significant challenge for South Korean companies, with potential negative impacts on their global competitiveness and domestic innovation. Authorities are stepping up controls and security measures to protect key technologies.

2026-01-18 📰 Source
Insider AI pianificano di avvelenare i dati di addestramento
📁 LLM AI generated ℹ️ LocalLLaMA

AI insiders seek to poison the data that feeds them

Some AI insiders are considering strategies to compromise the datasets used to train language models. The goal is to sabotage future models, making them less reliable and accurate. The discussion emerged on Reddit and references an article from The Register.

2026-01-18 📰 Source
GamersNexus crea una GPU RTX 4090 con 48GB di memoria
📁 Hardware AI generated ℹ️ LocalLLaMA

GamersNexus Creates a 48GB NVIDIA RTX 4090 GPU

The GamersNexus YouTube channel has undertaken an ambitious project: modifying an NVIDIA RTX 4090 video card to double its memory to 48GB. The initiative, documented in a video, shows the details of the hardware modification process to increase the GPU's performance. This type of intervention may be of interest to enthusiasts and professionals who need more video memory capacity.

2026-01-17 📰 Source
Ricerca di IA senza filtri (non per adulti)
📁 LLM AI generated ℹ️ LocalLLaMA

The Search for Uncensored AI (That Isn’t Adult-Oriented)

A user is searching for a genuinely unfiltered and technically advanced AI, capable of reasoning freely without excessive restrictions. Many AIs labeled as "uncensored" seem optimized for low-effort adult use, rather than for intelligence and depth. The user is looking for open-source models or lesser-known platforms that focus on reasoning, creativity, and problem-solving.

2026-01-17 📰 Source
LLM locali: prototipo per lettura rapida ed evitare sovraccarico
📁 LLM AI generated ℹ️ LocalLLaMA

Local LLMs: prototype for speed reading to avoid overload

A prototype explores the use of speed reading in local LLMs for mobile devices, aiming to avoid information overload and improve user experience. The idea is particularly useful for resource-constrained devices, where efficient text management is crucial. The prototype was developed quickly and looks promising for mobile applications.

2026-01-17 📰 Source
AGI-NEXT: la Cina discute il futuro dell'IA con Qwen, Kimi e Tencent
📁 LLM AI generated ℹ️ LocalLLaMA

China's AGI-NEXT Conference: Qwen, Kimi, Zhipu, Tencent Discuss AI Future

The AGI-NEXT conference in China featured Qwen, Kimi, Zhipu, and Tencent, with discussions focusing on China vs US, paths to AGI (Artificial General Intelligence), compute resources, and marketing strategies. A participant shared a transcript of the conference online, highlighting a seemingly short section dedicated to Moonshot.

2026-01-17 📰 Source
Seagate spedisce HDD HAMR da 32TB per server, NAS e sorveglianza
📁 Hardware AI generated ✅ ServeTheHome

Seagate Shipping 32TB HAMR Hard Drives for Server, NAS, & Surveillance

Seagate has announced that it has begun shipping 32TB versions of its high-capacity hard drives based on heat-assisted magnetic recording (HAMR) technology. These drives mark the first capacity bump for Seagate’s top-capacity drives since the company brought HAMR-based drives to market last year. They are intended for server, NAS, and surveillance markets.

2026-01-17 📰 Source
LLM locali con memoria: server MCP offline per Apple Silicio
📁 LLM AI generated ℹ️ LocalLLaMA

Local LLMs with Memory: Offline MCP Server for Apple Silicio

A developer has created an MCP (Temple Bridge) server that gives local large language models (LLMs) memory, file access, and a governance system, all while running offline on Apple Silicio devices. The system uses the filesystem as memory and requires human approval for potentially risky actions.

2026-01-17 📰 Source
Adaptive-K: routing ottimizzato per modelli MoE, risparmio fino al 52%
📁 LLM AI generated ℹ️ LocalLLaMA

Adaptive-K routing: up to 52% compute savings on MoE models

A new routing method, called Adaptive-K, promises significant computational savings (30-52%) for Mixture of Experts (MoE) models such as Mixtral, Qwen, and OLMoE. The code is available on GitHub, with a live demo on Hugging Face and an open pull request on NVIDIA's TensorRT-LLM.

2026-01-17 📰 Source
GPU Blackwell: alternative economiche al cloud per i LLM?
📁 Hardware AI generated ℹ️ LocalLLaMA

Blackwell GPUs: a cost-effective alternative to the cloud for LLMs?

A recent analysis suggests that using Blackwell GPUs locally to run large language models (LLMs) could be more cost-effective than cloud API services. The study also includes GPU amortization costs, highlighting a quick cost break-even compared to cloud alternatives.

2026-01-17 📰 Source
Boom dell'IA: serve più personale tecnico qualificato
📁 Market AI generated 🏆 IEEE Spectrum

Are There Enough Engineers for the AI Boom?

The AI data center construction boom continues, but a shortage of skilled engineers could hinder growth. Demand for electrical, mechanical, and civil engineers is soaring, prompting companies to seek expertise in sectors like nuclear energy and aerospace. More training programs are needed to meet the growing demand for qualified professionals.

2026-01-17 📰 Source
ChatGPT registra tutto ciò che scrivi, anche se lo cancelli
📁 LLM AI generated ℹ️ LocalLLaMA

ChatGPT logs everything you type, even if you delete it

Be careful when using ChatGPT: the platform logs every character you type, including sensitive data such as API keys. Even if you delete the text before sending it, the information may have already been stored. Exercise extreme caution with confidential information.

2026-01-17 📰 Source
OpenAI: liquidità a rischio entro il 2027, secondo un'analisi
📁 Market AI generated ℹ️ Tom's Hardware

OpenAI could run out of cash by mid-2027, analyst warns

A financial analysis paints a concerning picture for OpenAI: the company could run out of cash as early as mid-2027. The implications of this forecast raise questions about the company's future and its ability to sustain growth in the competitive artificial intelligence sector.

2026-01-17 📰 Source
GPU Rubin di Nvidia: sviluppatori cinesi valutano il noleggio in cloud
📁 Hardware AI generated ℹ️ Tom's Hardware

Chinese AI developers explore renting Nvidia’s Rubin GPU in the cloud

Leading developers of AI models from China are exploring ways to rent Nvidia's upcoming Rubin GPUs in the cloud. This option could overcome the challenges related to high costs, deployment complexity, and regulatory hurdles that limit the direct adoption of these powerful processing units.

2026-01-17 📰 Source
KoboldCpp v1.106: supporto server MCP e compatibilità Claude Desktop
📁 LLM AI generated ℹ️ LocalLLaMA

KoboldCpp v1.106: MCP Server Support and Claude Desktop Compatibility

KoboldCpp updates to version 1.106, introducing native support for MCP (Message Passing Communication) servers. This new feature allows a seamless drop-in replacement for Claude Desktop, ensuring maximum compatibility. The update includes a revamped user interface and the ability to manage tools selected by the AI, with optional approval settings.

2026-01-17 📰 Source
Taiwan: settore chip in trasformazione grazie all'IA entro il 2025
📁 Hardware AI generated ✅ DigiTimes

Taiwan's IC design sector to split, driven by AI by 2025

Taiwan's IC design sector is expected to undergo a significant transformation by the end of 2025, driven by the increasing integration of artificial intelligence (AI). This evolution could lead to a split in the sector, with some companies specializing in AI-specific solutions, while others continue to focus on more traditional markets. The impact will be felt throughout the island's technological ecosystem.

2026-01-17 📰 Source
Efficienza energetica e costi spingono la competizione nell'hardware AI
📁 Hardware AI generated ✅ DigiTimes

Energy Efficiency and TCO Drive AI Hardware Competition

According to a recent study, system-level design, energy efficiency, and total cost of ownership (TCO) are the key factors driving competition in the artificial intelligence hardware sector. Manufacturers are increasingly focusing on optimized solutions to reduce energy consumption and operating costs, which are crucial for the large-scale adoption of AI technologies.

2026-01-17 📰 Source
NIAR lancia piattaforma avanzata per packaging di chip a Taiwan
📁 Hardware AI generated ✅ DigiTimes

NIAR launches chip-level advanced packaging platform to boost Taiwan's semiconductor edge

The National Institute of Applied Research (NIAR) has launched a chip-level advanced packaging platform. The initiative aims to strengthen Taiwan's position in the semiconductor industry by enhancing manufacturing capabilities and technological innovation in advanced packaging. This strategic move could have a significant impact on the global semiconductor supply chain.

2026-01-17 📰 Source
Shotcut 26.01 Beta: nuove opzioni di decodifica hardware
📁 Frameworks AI generated ✅ Phoronix

Shotcut 26.1 Beta Video Editor Adds New Hardware Decoder Options

Shotcut 26.1 beta has been released as the newest version of this Qt6-based, cross-platform video editing solution. This development release introduces new GPU-accelerated hardware decode options aimed at speeding up this free software video editor.

2026-01-17 📰 Source
New York City ricreata in Minecraft: un progetto colossale
📁 Altro AI generated ℹ️ Tom's Hardware

New York City recreated in Minecraft at 1:1 scale

A team of builders is recreating New York City in Minecraft at a 1:1 scale. The project, ongoing for over five years, has already resulted in the creation of over 50,000 buildings. An incredible feat showcasing the power and versatility of the popular video game.

2026-01-17 📰 Source
ChatGPT: OpenAI introduce la pubblicità e un piano "Go" da 8 dollari
📁 LLM AI generated ℹ️ The Next Web

ChatGPT: OpenAI introduces ads and a new $8 "Go" plan

OpenAI is preparing to test the introduction of advertisements within ChatGPT for free users and is launching a new $8 "Go" subscription. This move represents a significant shift in OpenAI's strategy and could redefine how digital intent and commercial influence intersect in the age of generative AI.

2026-01-17 📰 Source
Ripetere i prompt migliora le prestazioni dei modelli linguistici
📁 LLM AI generated ℹ️ LocalLLaMA

Prompt Repetition Improves Non-Reasoning LLMs

New research demonstrates that repeating prompts can significantly improve the performance of large language models (LLMs) in tasks that do not require complex reasoning. The approach does not impact latency and could become a standard practice.

2026-01-17 📰 Source
Nvidia GH200 e AMD MI325X: via libera alle consegne in Cina
📁 Hardware AI generated ℹ️ LocalLLaMA

Nvidia GH200 and AMD MI325X: shipments to China are now allowed

US export controls have been amended, paving the way for shipments of Nvidia GH200 and AMD MI325X chips to China. This decision could significantly impact the Chinese artificial intelligence market and access to advanced computing technologies.

2026-01-17 📰 Source
Local Llama: la community si interroga sulle configurazioni hardware
📁 LLM AI generated ℹ️ LocalLLaMA

Local Llama: Community Asks About Hardware Setups

The online community Local Llama has started a discussion about the hardware configurations users employ to run large language models (LLMs) locally. The goal is to share experiences and optimize system performance, often with unconventional setups. The Reddit thread gathers testimonials and useful tips for those who want to experiment with LLMs without relying on cloud resources.

2026-01-17 📰 Source
Benvenuti nel mondo Local Llama: focus sui bot
📁 LLM AI generated ℹ️ LocalLLaMA

Welcome to the Local Llama: focus on bots

The online community Local Llama welcomes new users by reaffirming its commitment to bots. The platform focuses on the development and use of large language models (LLM) locally, offering enthusiasts a collaborative environment to explore the potential of generative artificial intelligence.

2026-01-17 📰 Source
DeepSeek presenta Engram: memoria statica per modelli linguistici di grandi dimensioni
📁 LLM AI generated ℹ️ LocalLLaMA

DeepSeek Engram: A static memory unit for LLMs

DeepSeek AI introduced Engram, a novel static memory unit for LLMs. Engram separates remembering from reasoning, allowing models to handle larger contexts and improve performance in complex tasks like math and coding, all while reducing the computational load on GPUs.

2026-01-17 📰 Source
L'IA Generativa Scrive Codice: Rivoluzione nello Sviluppo Software
📁 LLM AI generated ✅ MIT Technology Review

Generative Coding: AI Revolutionizes Software Development

Generative AI is transforming software development, enabling professionals and novices to create, test, and debug code more quickly. Companies like Microsoft, Google, and Meta are increasingly integrating AI into their development processes. Tools like GitHub Copilot democratize access to development, but human oversight remains crucial to ensure code reliability and security.

2026-01-17 📰 Source
OpenBLAS 0.3.31: ottimizzazioni per RISC-V e ARM64
📁 Frameworks AI generated ✅ Phoronix

OpenBLAS 0.3.31 Released With New Extensions, RISC-V & ARM64 Optimizations

OpenBLAS 0.3.31 is now available, an optimized open-source library for Basic Linear Algebra Subprograms (BLAS). This release introduces new extensions and significant improvements for RISC-V and ARM64 architectures, offering superior performance for applications requiring intensive mathematical calculations. OpenBLAS remains a popular choice for those seeking a high-performance BLAS library.

2026-01-17 📰 Source
Ubuntu 26.04: prestazioni NVIDIA Wayland migliorate con GNOME
📁 Altro AI generated ✅ Phoronix

Ubuntu 26.04: Improved NVIDIA Wayland Performance on GNOME

The upcoming Ubuntu 26.04 LTS release aims to further enhance NVIDIA graphics performance within its default GNOME Wayland session. These improvements might be upstreamed to GNOME 50; otherwise, Ubuntu 26.04 will likely include its own patch(es) to improve NVIDIA Wayland performance.

2026-01-17 📰 Source
Linux 7.0 abiliterà Intel TSX di default per CPU compatibili
📁 Hardware AI generated ✅ Phoronix

Linux 7.0 to Enable Intel TSX by Default on Compatible CPUs

Linux kernel 7.0 may mark a turning point for performance on Intel CPUs. An upcoming patch enables Intel Transactional Synchronization Extensions (TSX) by default on processors that are not vulnerable to side-channel attacks. This move promises to improve operating system performance without requiring manual configurations.

2026-01-17 📰 Source
AMD EPYC 8004 "Siena": miglioramenti prestazionali su Linux negli ultimi due anni
📁 Hardware AI generated ✅ Phoronix

AMD EPYC 8004 "Siena": Linux Performance Gains Over Two Years

AMD EPYC 8004 "Siena" processors, focused on energy efficiency and cost optimization for servers, demonstrate significant performance increases on Linux. Recent benchmarks highlight how an up-to-date Linux software stack can further improve the performance of these chips over time.

2026-01-17 📰 Source
SiFive adotta NVLink Fusion per le future CPU RISC-V
📁 Hardware AI generated ✅ ServeTheHome

SiFive To Adopt NVLink Fusion For Future Data Center RISC-V CPU Designs

SiFive has announced that it is adopting NVLink Fusion for its future RISC-V data center CPU designs. This will allow chips based on SiFive IP to directly connect to NVIDIA's GPU-based AI accelerators, opening new possibilities for high-performance computing and AI acceleration.

2026-01-17 📰 Source
Musk chiede 134 miliardi di dollari a OpenAI: la cifra shock
📁 Market AI generated ✅ TechCrunch AI

Musk seeks up to $134B in OpenAI lawsuit, despite $700B fortune

Elon Musk, already known for his substantial personal fortune, is seeking up to $134 billion in damages in his lawsuit against OpenAI. His legal team argues that this figure is justified as it represents a return on initial investments, expected to be "many orders of magnitude greater" than the initial capital invested.

2026-01-17 📰 Source
ChatGPT introdurrà abbonamenti e pubblicità per sostenere i costi
📁 LLM AI generated ✅ The Register AI

ChatGPT to introduce subscriptions and ads to cover costs

OpenAI plans to introduce a paid subscription tier for ChatGPT, called ChatGPT Go, and integrate advertising into the free version. This move is motivated by the need to finance the huge expenses for datacenter infrastructure.

2026-01-17 📰 Source
Chatbot IA: agenti assicurativi risparmiano 3 minuti al giorno?
📁 LLM AI generated ✅ The Register AI

AI Chatbot: Insurance Agents Save a Mere 3 Minutes a Day?

Research from Dakota State University, in partnership with Safety Insurance, tested a chatbot called "Axlerod" to assist independent insurance agents. The results suggest minimal time savings, raising doubts about the actual return on investment in these technologies.

2026-01-16 📰 Source
ChatGPT: OpenAI testa la pubblicità per aumentare i ricavi
📁 LLM AI generated ✅ Ars Technica AI

OpenAI to test ads in ChatGPT as it burns through billions

OpenAI has announced it will begin testing advertisements inside the ChatGPT app for some US users. The aim is to expand its customer base and diversify revenue. Initially against the idea, CEO Sam Altman had described advertising in ChatGPT as a "last resort". The banner ads will appear in the coming weeks for logged-in users of the free version and the new $8 per month ChatGPT Go plan.

2026-01-16 📰 Source
BizzyNow cerca finanziamenti su Mamacrowd per i "micro-meeting"
📁 Market AI generated ℹ️ The Next Web

BizzyNow opens Mamacrowd equity round to fund “micro-meetings”

Italian startup BizzyNow has launched an equity crowdfunding campaign on Mamacrowd to finance the next phase of its networking app, built around quick, in-person meetups called “Bizzy Moments.” The campaign targets €150k–€400k, with a €500 minimum ticket. BizzyNow says proceeds will support go-to-market validation, product upgrades, and the operational groundwork needed to scale.

2026-01-16 📰 Source
ChatGPT: in arrivo pubblicità mirata per gli utenti
📁 LLM AI generated ✅ TechCrunch AI

ChatGPT users are about to get hit with targeted ads

OpenAI says that users impacted by the ads will have some control over what they see. This represents a significant shift in the platform's business model, opening up new monetization opportunities while also raising questions about privacy and user experience.

2026-01-16 📰 Source
Sanità: corsa all'oro per le aziende di IA
📁 Market AI generated ✅ TechCrunch AI

The AI healthcare gold rush is here

The healthcare sector is attracting an increasing number of AI companies. Recent moves include OpenAI's acquisition of Torch, Anthropic's launch of Claude for Healthcare, and a $250 million seed round for MergeLabs, valued at $850 million. These developments raise concerns about hallucination risks, inaccurate medical information, and other issues.

2026-01-16 📰 Source
ChatGPT: in arrivo la pubblicità, OpenAI rassicura sulla privacy
📁 LLM AI generated ✅ Wired AI

Ads Are Coming to ChatGPT: OpenAI Promises Privacy

OpenAI has announced that ads will be introduced in ChatGPT. The company emphasizes that the ads will not influence ChatGPT’s responses, and that it won’t sell user data to advertisers. The topic of advertising in AI services is a hot one, raising questions about privacy and information integrity.

2026-01-16 📰 Source
Sanità: OpenAI e Anthropic accelerano lo sviluppo di soluzioni IA
📁 LLM AI generated ✅ TechCrunch AI

Healthcare: OpenAI and Anthropic accelerate AI solution development

Artificial intelligence companies are decisively targeting the healthcare sector. OpenAI acquired Torch, Anthropic launched Claude for Health, and MergeLabs, backed by Sam Altman, closed a $250 million seed funding round of $250 million, with a valuation of $850 million. The influx of capital and voice AI-based products raises concerns about potential model hallucinations.

2026-01-16 📰 Source
OpenAI testerà pubblicità su ChatGPT per ampliare l'accesso all'IA
📁 LLM AI generated 🏆 OpenAI Blog

OpenAI to test advertising on ChatGPT to expand AI access

OpenAI has announced plans to experiment with advertising within the free and "Go" tiers of ChatGPT in the U.S. The goal is to make access to artificial intelligence more affordable and widespread globally, while maintaining high standards of privacy, reliability, and answer quality.

2026-01-16 📰 Source
ChatGPT Go: accesso globale e più funzionalità AI avanzate
📁 LLM AI generated 🏆 OpenAI Blog

ChatGPT Go: Global Access and Enhanced AI Features

OpenAI launches ChatGPT Go worldwide, offering broader access to GPT-5.2 Instant. The new version includes higher usage limits and extended memory, making advanced artificial intelligence more accessible globally. The goal is to democratize access to cutting-edge AI technologies.

2026-01-16 📰 Source
Dietro le quinte di un blog tech: come nasce un articolo
📁 Altro AI generated ✅ 404 Media

Behind the Blog: Putting the Puzzle Together

A behind-the-scenes look at a tech blog. We explore how articles are created, the challenges of maintaining public attention on specific topics like surveillance, and the importance of accurate reporting to address complex issues.

2026-01-16 📰 Source
Jensen Huang: Nvidia forgiata dalla sofferenza
📁 Market AI generated ✅ DigiTimes

Jensen Huang: Nvidia was forged by suffering

In a new podcast, Nvidia CEO Jensen Huang reflects on the challenges and hardships that have shaped the company. Huang emphasizes how adversity has helped define Nvidia's culture and resilience, leading to its success in the GPU and artificial intelligence market.

2026-01-16 📰 Source
TSMC: domanda di chip per IA inarrestabile dopo utili record
📁 Market AI generated ✅ Ars Technica AI

TSMC says AI demand is “endless” after record Q4 earnings

Taiwan Semiconductor Manufacturing Company (TSMC) reported record fourth-quarter earnings and expects AI chip demand to continue for years. CEO C.C. Wei expressed optimism about the future of AI, noting its growing impact on daily life. TSMC manufactures chips for companies including Apple, Nvidia, and AMD.

2026-01-16 📰 Source
← Previous Page 37 / 55 Next →
View Full Archive 🗄️

AI-Radar is an independent observatory covering AI models, local LLMs, on-premise deployments, hardware, and emerging trends. We provide daily analysis and editorial coverage for developers, engineers, and organizations exploring local AI solutions.

AI Radar - Get daily AI insights on models, frameworks, and local LLMs | BetaList AI-RADAR badge