Europe's AI code urges companies to disclose training data and avoid copyright violations
AI Act fines could reach 7% of revenue for noncompliance
Large language model articles
Open source project is making strides in bringing CUDA to non-Nvidia GPUs
CUDA compatibility with AMD and Intel GPUs isn't a pipe dream anymore
Why it matters: Nvidia introduced CUDA in 2006 as a proprietary API and software layer that eventually became the key to unlocking the immense parallel computing power of GPUs. CUDA plays a major role in fields such as artificial intelligence, scientific computing, and high-performance simulations. But running CUDA code has remained largely locked to Nvidia hardware. Now, an open-source project is working to break that barrier.
Germany asks Apple and Google to ban DeepSeek over illegal data transfers to China
DeepSeek's failure to follow GDPR puts its future in Europe at risk
YouTube adds more AI, but not everyone's happy about it
New AI-generated search carousels, where "quality and accuracy may vary"
Copilot Vision brings Microsoft's screen-watching AI to everyday Windows tasks
The AI that watches everything–and might do something useful
AI flunks logic test: Multiple studies reveal illusion of reasoning
As logical tasks grow more complex, accuracy drops to as low as 4 to 24%
Microsoft fixes first known zero-click attack on an AI agent
The "EchoLeak" security flaw affected Microsoft 365 Copilot
ChatGPT gets crushed at chess by a 1 MHz Atari 2600
Editor's take: Despite being hailed as the next step in the evolution of artificial intelligence, large language models are no smarter than a piece of rotten wood. Every now and then, some odd experiment or test reminds everyone that so-called "intelligent" AI doesn't actually exist if you're living outside a tech company's quarterly report.
OpenAI warns ChatGPT logs will be retained "indefinitely," blames court order
The Internet is forever
FDA claims victory with "Elsa" AI tool launch, but insiders call it half-baked
Why it matters: It was only a matter of time before governments jumped aboard the AI hype train. The US government has a lousy track record when it comes to rolling out tech projects – remember the Affordable Care Act website fiasco? Now, it is launching a new LLM, and early signs suggest the tool is rushed and unfinished.
Microsoft offers free cybersecurity tools to Europe in major initiative
Redmond seeks to avoid further EU antitrust penalties
"Godfather of AI" warns that today's AI systems are becoming strategically dishonest
Yoshua Bengio says labs are ignoring warning signs
Gmail's smart replies will mimic your writing style by analyzing your emails and Google Drive files
Machines talking to machines
In brief: A future in which generative AIs write emails back and forth to each other on our behalf has moved a little closer. Google is improving Gemini's smart replies, making them not only longer, but also more personalized by analyzing your previous emails and Drive files.
Google adds on-device AI to Chrome to catch scams in real time
GenAI also helps Android apps and mobile search combat scam campaigns
Developer ports compact Llama 2 LLM to DOS in weekend hackathon
Unfortunately, the feat can only be accomplished on computers with 32-bit CPUs
Sam Altman says polite ChatGPT users are burning millions of OpenAI dollars
OpenAI CEO says it's money "well spent"
Manners are not ruining the environment: The costs of training and running artificial intelligence model are massive. Even excluding everything but electricity, AI data centers burn through over $100 million a year to process user prompts and model outputs. So, does saying "please" and "thank you" to ChatGPT really cost OpenAI millions? Short answer: probably not.
ChatGPT gets scarily good at guessing photo locations, sparking doxxing concerns
Simple photos could reveal real-world locations
Microsoft's BitNet shows what AI can do with just 400MB and no GPU
BitNet b1.58 2B4T outperforms rivals like Llama, Gemma, and Qwen on common tasks
New approach from DeepMind partitions LLMs to mitigate prompt injection
Secure interpreter tracks data flow to block unsafe actions triggered by manipulated text
OpenAI is quietly working on a social network similar to Twitter, powered by ChatGPT
One goal of OpenAI's prototype is to gather real-time user data
Sophisticated bot uses OpenAI to bypass filters, flooding over 80,000 websites with spam
AkiraBot exploits contact forms and live chat widgets with "personalized" messages
Google AI falls for April Fools' prank story, presents it as real news
According to Google AI, Cwmbran, Wales is the roundabout capital of the world
We are finally beginning to understand how LLMs work: No, they don't simply predict word after word
Anthropic just ran Claude through a brain scanner
The Weird World of AI Hallucinations: When AI Makes Things Up
When an algorithmic system generates information that seems plausible but is actually inaccurate or misleading, computer scientists call it an AI hallucination.
