In the News

Chinese AI lab DeepSeek's last model release was V3.2 (and V3.2 Speciale) last December . They just dropped the first of their hotly anticipated V4 series in the shape of two preview models, <a href="https://huggingface.co/deepseek-a

Huawei ​Technologies said on Friday ‌its ⁠Ascend ⁠supernode ​based on Ascend 950 ​AI chips ​would fully ⁠support ‌Deepseek's ​V4 ​versions ⁠after the Chinese artificial ​intelligence startup ​launched a preview of ‌its new model.

Chinese artificial intelligence (AI) start-up DeepSeek has launched its latest AI model, the V4.

China’s DeepSeek launched preview versions of its long-awaited V4 model, breaking months of silence from one of the country’s most closely watched AI labs.

Chinese AI startup DeepSeek has released a preview version of its long awaited V4 large language model.

New version is reportedly better optimised for China’s domestic chips

DeepSeek's latest open-source AI models boast up to 1.6 trillion parameters and elite coding skills. Discover how the new Pro and Flash var

DeepSeek has introduced its new DeepSeek-V4 AI models, comprising Pro and Flash versions. The new model claims to compete with ChatGPT and Gemini in many key benchmarks.

DeepSeek has introduced its new DeepSeek-V4 AI models, comprising Pro and Flash versions. The new model claims to compete with ChatGPT and Gemini in many key benchmarks.

When the Chinese start-up DeepSeek published details about one of its artificial intelligence models last year, it sent shock waves through the tech industry. The company said it had built its system by spending far less on computer chips than American rivals like OpenAI and Anthropic. It marked the

A newly identified cyber threat group, UNC6692, is using a clever mix of social engineering and custom malware to infiltrate corporate networks. By impersonating IT helpdesk personnel on Microsoft Teams, these hackers trick employees into downloading a sophisticated malware suite that steals sens

Microsoft is actively investigating a known issue preventing some users from joining Microsoft Teams meetings on Windows devices, following a recent update to the Microsoft Edge browser. The disruption is affecting organizations, including those using NHSmail infrastructure, with reports indicati

Huawe

DeepSeek on Friday launched V4, its highly anticipated new series of open-source AI models with enhanced reasoning and coding capabilities. This is the first new-generation model that DeepSeek releas

The Chinese startup unveiled the V4 Flash and V4 Pro preview series, touting top-tier performance in coding benchmarks and big advancements in reasoning and agentic tasks

DeepSeek has launched a preview of its highly anticipated new model, the V4. According to the Chinese firm, The pro version of V4 outperforms other open-source models in world-knowledge benchmarks, trailing only Google's Gemini-Pro-3.1 which is a closed-source model.

Cybersecurity researchers at Socket have uncovered a major supply chain compromise affecting the Bitwarden CLI. Attackers successfully abused a GitHub Action in Bitwarden’s CI/CD pipeline to inject malicious code into the popular password manager’s npm package. This breach is part of the br

Chinese companies have embraced making their most advanced artificial intelligence models available to all.

Anthropic traced weeks of Claude Code complaints to three product-layer changes: a reasoning-effort downgrade, a caching bug that cleared thinking blocks every turn, and a verbosity prompt that cut coding-eval scores by 3%. All fixes shipped by April 20.

DeepSeek, a Chinese AI firm, is previewing a significant upgrade to its artificial intelligence model. The new V4 model will feature pro and flash versions.

DeepSeek has finally released its much-anticipated next-generation foundational artificial intelligence model, the open-source V4, which it said was competitive with leading US closed-source models from the likes of OpenAI and Google DeepMind. DeepSeek released two versions of the model on Friday, w

Key PointsServiceNow's first-quarter subscription revenue rose an impressive 22%.

DeepSeek rolled out preview versions of a new flagship artificial intelligence model a year after upending Silicon Valley, calling it the most powerful open-source platform in a challenge to rivals from OpenAI to Anthropic PBC.

<a href="https://news.google.com/rss/articles/CBMiqwFBVV95cUxQUVllcUNPUUdaaW1PZlVEblJFME5Ecnl3R1lvQmYwTVRUdUE2dng2cTdDNUFkTmVkRzNIcEk5WTRKdDJEX3BudElma2hGUjFTTDRTR2xNcndid2VzVmx6Y192Z0FmcDMwVm5nY3ZGY0k3VFVJZFM0dk9vYlhnTGNfU2tlbnRRLVJSbGpXS0J5X1k0MUh1LUZZSE54QUlXNzZLdDNyX3ZicVl0QzA?oc=5" target="_bla

A newly identified threat group, UNC6692, has been caught running a sophisticated multistage intrusion campaign that uses Microsoft Teams impersonation, a custom modular malware suite, and cloud infrastructure abuse to deeply penetrate enterprise networks, all without exploiting a single software

OpenAI has launched its GPT-5.5 model, claiming it is its most intuitive AI yet.The AI mdoel comes with improved efficiency, handles complex coding tasks better, and enhances knowledge work, the company claims

An update on recent Claude Code quality reports It turns out the high volume of complaints that Claude Code was providing worse quality results over the past two months was grounded in real problems.

impact.com expands YouTube integration to give brands and agencies a single system for creator campaign management and verified performance tracking.

<a href="https://news.google.com/rss/articles/CBMiqwFBVV95cUxOQk5nNGsyeFRKd1Q0X2V4YjE4MG9HY1paWGY2YndLMm4yYkZHaHByU0ExMFZQQWtuRFNjQno5SWUxZnRxZVVYcnRRR19RSnhLUURIcWpwRkFzUy0tdC1oM20yRUJ5TGJOQU9lOENQZ2swd0lwSEJtYkVwQ1hrcjA5TnNzTGJnUjZuNF9MbVBnOHplNmpYOUdWZ3JBU2xTNFJGMV8xWDdmWnpqQWc?oc=5" target="_bla

OpenAI has introduced a version of ChatGPT designed to help clinicians by supporting tasks such as documentation and medical research. The company is making the new ChatGPT for Clinicians available first in the United States and later in additional countries, it said in a Wednesday (April 22) pr

<a href="https://news.google.com/rss/articles/CBMiwwFBVV95cUxNOTJqVG9YNHdwa2FlcGN1WEo0LV9WbUJIYW1DVVh2cklJcXREbFNpNFRSOWxZYmxfWGRqNWFOTW01dlVkWnVZUS1wMWc3RHRLSTNXWjJlUjRjTm1zTFYyVWkyYXdNQ1NvMUhlZ3d5Rkh3MWpqWlRia0stRWZfalZzelJwSHJoNmlFNFl3dmY4MDdOMnotMjdLNWVUVjBRY3JBTnVuVnNfNWk3R2xJQzBlMzk4YS1QWWtNWU

OpenTable's April mandate to become restaurants' primary booking system sparks backlash over exclusivity fears, data burdens, and lost flexibility. Operators demand multi-platform f

Mullvad's new approach addresses an issue with Apple app updates, but you'll need to do more hands-on maintenance.

I'm blown away at just how easy it was

OpenAI launched GPT-5.5, boosting agentic coding and research while advancing its super app that merges ChatGPT, Codex, and a browser. Benchmarks show gains over rivals, now live fo

OpenAI's workspace agents transform ChatGPT into a shared AI workforce for teams, handling cloud-based workflows like feedback triage and sales outreach with tool integrations and a

Anthropic denied “nerfing” Claude Code after a week of user complaints about quality. Samuel Boivin/NurPhoto via Getty Images Anthropic found issues with Claude Code after complaints that the popular tool had gotten worse. The company denied “nerfing” or intentionally degradi

OpenAI has released GPT-5.5, which makes ChatGPT more reliable and useful

@media (max-width: 768px) { .image-container { width: 100% !important; float: none !important; margin: 0 0 1rem 0 !important; } } Image source: The Mo

Anthropic said it found three issues with Claude Code after users complained the AI tool deteriorated.

LlamaIndex have a most excellent open source project called LiteParse , which provides a Node.js CLI tool for extracting text from PDFs. I got a version of LiteParse working entirely in the browser, using most of the same libraries that LitePars

AI agents depend on the quality and accessibility of the data behind them, placing Google Cloud databases at the center of that equation. The agentic data cloud introduced at Google Cloud Next 2026 underscores Google’s push to position its databases as the foundation of AI-driven enterprise archi

The company adds a feature that will reveal what topics teens delve into with AI on Facebook, Instagram and WhatsApp.

Hackers are actively exploiting a critical vulnerability in the Breeze Cache plugin for WordPress that allows uploading arbitrary files on the server without authentication. [...]

The Bittensor price prediction gained fresh weight on April 20 when BitGo launched institutional custody and staking for TAO subnet tokens through a partnership with Yuma, per CoinMarketCap. Bittensor (TAO) trades at $243 with 70% of supply locked in staking. Three days earlier, Worldcoin (WLD) s

Turning messy thoughts into cleanly executed ideas.

OpenAI has launched GPT-5.5, focusing on improving ChatGPT’s ability to handle complex tasks and real-world workflows.

Rising refusal rate from Acceptable Use Classifier leaves customers paying for nothing Anthropic's release last week of Opus 4.7 came with stronger safeguards to prevent misuse. Unfortunately, these safeguards have also managed to thwart legitimate use.…

Software stocks tumbled again after IBM and ServiceNow earnings. But cybersecurity and data stocks may be bargains.

Florida Attorney General James Uthmeier announces the launch of a criminal investigation into OpenAI in connection with the 2025 Florida State University shooting.