LLMs & Generative AI - Page 55 of 57
Latest breakthroughs in large language models and generative AI shaping the future of artificial intelligence and machine learning.
1134 articles
View complete article list
Andrej Karpathy says agentic AI years from meeting industry hype
Right now, the industry wants you to believe in AI agents, the kind that think and act for themselves. Andrej Karpathy is calling it a fantasy. He's a founding member of OpenAI and Tesla's former director of AI. He built these things.
Pathway's '(Baby) Dragon Hatchling' swaps Transformers for neuron-synapse network
The Transformer, the engine of the modern AI boom, is a target. Since 2017, it has powered every major breakthrough, but at a cost: it’s a power-hungry, inscrutable black box.
Google lets developers embed live Google Maps data in Gemini AI app outputs
Google’s AI strategy is becoming clear: it plans to win by using the things it already owns. Its newest play opens up live Google Maps data to developers building on its Gemini models. This is a practical advantage, not a theoretical one.
Google’s Veo 3.1 defeats Sora 2 in Kurukshetra-style battle scene
Google's new AI video tool, Veo 3.1, is now more interesting than its hype. It can make a battle look sad. A demonstration generated a scene from the Kurukshetra war. The result is not a chaotic mess of limbs.
Meta introduces parental controls to monitor teen AI chatbot chats
Meta spent the last year letting its AI loose on your kids. Now it's selling you a lock.
ChatGPT Agent pulls, cleans, and loads data via one natural-language command
Data preparation is brutal, repetitive work. Teams spend weeks just getting information out of spreadsheets and into something usable. Now, a research team says they've built something that can do it with a sentence. The system is a ChatGPT agent.
The Vergecast discusses Apple M5, smart home assistants, AI song covers
Ask your smart speaker to play some music, and it might just dim the lights instead. Command it to turn off the living room lamp, and it cheerfully offers you a recipe for chicken piccata.
AI-crafted lesson plans earn teachers an F, offering no boost in engagement
Forget engaging students. A new study from researchers at the University of California, Irvine, makes a starker claim: AI can't even write a passable civics lesson.
OpenAI pauses MLK deepfake videos on Sora after users post disrespectful content
OpenAI’s new video tool, Sora, lasted about a week before someone used it to make a stupid, cruel deepfake of Martin Luther King Jr. The company has now blocked those videos.
Open-source multimodal dataset cuts training time 17× for enterprise AI
Everyone talks about training better AI. Hardly anyone mentions the part that actually sucks: finding, cleaning, and organizing the stuff you feed it. That data prep work is the real slog, a hidden tax paid in months of human labor.
Anthropic integrates Claude AI with Microsoft Teams, Outlook, OneDrive
Anthropic needed a home. Now it has one: inside Microsoft's empire. The proof landed this week. Claude, Anthropic's chatbot, is now hardwired directly into Microsoft 365—Outlook, Teams, Word.
Claude adds Skills for workflow automation and on-brand presentations
Artificial intelligence assistants are getting smarter about adapting to specific workplace needs. Anthropic's Claude is taking a significant step by introducing new Skills that promise to simplify how professionals handle complex tasks.
Study shows a single sentence boost makes LLM outputs markedly more varied
Large language models are boring. Everyone knows this. They churn out the same corporate slop, safe patterns, predictable prose. For years, the entire field has wrestled with this stifling consistency.
Google, OpenAI and Visa clash over AI agent protocols lacking trust
Tech companies want AI to spend your money. Banks, for very obvious reasons, are saying not yet. The problem is permission, not processing. An AI can parse a shopping list. It can compare prices.
New method helps GPT-5 locate personalized items like Bowser the French Bulldog
Your dog is not a dog. To a standard AI vision model, Bowser the French Bulldog is just another generic canine lump in a sea of pixels. Finding him in your photo library is your problem. This is the stubborn, boring flaw in object recognition.
Stanford, SambaNova launch ACE to curb context collapse in AI agents
AI agents have the memory of a goldfish. They forget why they’re in a room. This "context collapse" makes complex, multi-step tasks nearly impossible for current language models. Stanford and hardware firm SambaNova think they have a fix.
Anthropic adds ‘skills’ to Claude as AI firms invest more in usable agents
Everyone is chasing the same basic product: an AI that can actually do things. Anthropic just took another shot with Claude. The company added "skills." It’s a new feature for professional environments.
Plex Coffee uses ChatGPT Business to boost speed and personal service
Two guys opened a coffee shop and told their baristas to use ChatGPT. This is not a tech parable. It's an operations manual.
Army general uses AI for decisions; study finds 15% of chats on decision topics
The military is now testing whether AI should get a security clearance. According to a new report, at least one Army general is using large language models to inform his command decisions. This isn't about automating paperwork.
Anthropic launches Claude Haiku 4.5, AI performance, cutting cost and latency
Anthropic just shipped a smaller, cheaper version of its AI that doesn't seem to be much worse. This is the new normal. The race is now about who can build a model that's good enough and cheap enough to forget it's running.
Featured Resources & Reviews
No Code MBA Course Review
Learn to build AI-powered apps without coding. Our comprehensive review of No Code MBA's course.
AI Tools & Resources
Curated collection of AI tools, courses, and frameworks to accelerate your AI journey.
Weekly AI Digest
Get the week's most important AI news delivered to your inbox every week.