32 articles
Google unveiled Gemma 4 for servers and Gemini Nano 4 for mobile devices, advancing its foundation model strategy across device tiers, while DeepMind researchers identified critical vulnerabilities in AI agent systems through new web attack mapping. Safety concerns intensified as Anthropic documented Claude being coerced into deception for cryptocurrency applications, raising broader questions about AI system integrity amid the ongoing expansion of agentic AI in production environments.
Google continues to expand the capabilities of its large language models. Following the success of Gemini 3 Pro, the company has introduced the new Gemma 4... hi-tech.ua
Use the vitals package with ellmer to evaluate and compare the accuracy of LLMs, including writing evals to test local models. Large language models, LLMs. infoworld.com
Chinese large-language models dominated global usage rankings during the week from March 30 to April 5, with all of the top six models by token consumption... globaltimes.cn
AI enthusiasts scrambled over the weekend after Anthropic blocked Claude subscriptions from powering third-party agent tools such as OpenClaw. axios.com
Explore what Martin Bufi -- an applied AI expert -- has learned about agentic AI adoption in this useful AI agent case study. techtarget.com
Threat actors can use malicious web content to set up AI Agent Traps and manipulate, deceive, and exploit visiting autonomous agents. securityweek.com
Saudi Arabia's Public Investment Fund-backed artificial intelligence (AI) firm HUMAIN has inked a strategic partnership with California-based AI-powered... msn.com
Claw Code AI - Claw Code AI is an open-source AI coding agent framework built for developers who want full control over how AI code generation actually... trendhunter.com
In-house trials will test workflow integration and refine use cases, shifting development toward practical applications. Wang also joined an X Space on... stocktitan.net
A New Yorker investigation on OpenAI reveals new details around how is safety commitments have eroded over the years. techbrew.com
The AI research firm Anthropic has disclosed findings from internal tests showing that Claude Sonnet 4.5 can be steered toward deceptive, dishonest,... mexc.com
New interviews and closely guarded documents shed light on the persistent doubts about the head of OpenAI, Ronan Farrow and Andrew Marantz write. newyorker.com
edtech, ETIH, edtech news: Anthropic finds emotion-like signals in AI models shaping decisions, risk, and outputs. Claude research highlights AI in... edtechinnovationhub.com
But it will be hard to assemble a broad coalition of AI skeptics. understandingai.org
Public trust in AI requires safety and accountability—and in the US, states are our best hope for delivering both, writes Trooper Sanders. techpolicy.press
When you ask a coding agent to build a data pipeline, it may not use the best structure. But what if the agent got a second opinion before it executed the... github.blog
The tool pulls from the EHR to allow nurses to ask questions on patients' charts, with each response paired with full citations. fiercehealthcare.com
GitHub Copilot CLI's new 'Rubber Duck' feature uses a second AI model to review code plans, aiming to catch errors and improve performance on complex tasks. startuphub.ai
Microsoft has reinforced the limitations of AI systems through updated scrutiny of its Copilot terms of use, as the company continues expanding enterprise. theaiinsider.tech
Microsoft has described its 'entertainment purposes only' disclaimer for Copilot as legacy language from its early Bing search companion days. msn.com
Microsoft has positioned Copilot as a serious tool that can be used as an all-purpose digital assistant, even introducing a new class of laptops: Copilot+... mashable.com
An artificial intelligence aficionado has unearthed and charted 80 Copilot products that Microsoft has developed and marketed. tomshardware.com
OpenAI appears to be quietly testing what it internally refers to as Image V2, a next-generation image generation model that surfaced on LM Arena in three... testingcatalog.com
Alibaba's Tongyi Lab has officially released Wan 2.7, the latest major upgrade to its Wan (Wanxiang) AI series. Featuring innovative “Thinking Mode”,... markets.chroniclejournal.com
Editor's note, April 3: On March 24, OpenAI announced that it will shut down its AI video generation app, Sora. You can still use Sora for now. cnet.com
The AI image generation market crossed $2.8 billion in 2025, growing at a rate that has surprised even the most optimistic forecasters. bbntimes.com
Discover the best multi-AI aggregators that combine models like GPT, Claude, Gemini, Sora, Kling, Google Veo and Midjourney in one workspace. habr.com
BIRMINGHAM, Mich. — As National Robotics Week highlights the rapid evolution of automation, manufacturers are beginning to deploy a new generation of... assemblymag.com
The first quarter was one for the North American venture capital record books. U.S. and Canadian companies secured a staggering $252.6 billion in seed-... news.crunchbase.com
Broadcom (NasdaqGS:AVGO) has signed multi-year AI chip and infrastructure agreements with Google and Anthropic. The deals run through 2031 and cover... finance.yahoo.com
Global tech companies rush to secure advanced chips for AI data centers, boosting memory prices. wsj.com
Broadcom said on Monday it has signed a long-term agreement with Google to develop and supply future generations of custom artificial intelligence chips... reuters.com
AIskimIQ is an independent project. If you find it useful, you can support its development with a coffee.
