Meta's new model is as open as Zuckerberg's private school
… In fact, Google routinely releases small open weights models derived from its larger proprietary Gemini models, with its Gemma 4 family being the latest example. …
Search
Anthropic Mythos model can find and exploit 0-days
… Now they have a new Big Bad: an AI model that can generate zero-day vulnerabilities. Anthropic made the model and named it Mythos. …
Claude Opus wrote a Chrome exploit for $2,283
Security Claude Opus wrote a Chrome exploit for $2,283 Pause your Mythos panic because mainstream models anyone can use already pick holes in popular software Anthropic withheld its Mythos bug-finding model from public release due to concerns that it would enable attackers to find and exploit vulne… …
Maude-HCS helps model and validate covert network designs
… Either way, the team hopes next to invert the process developed with Maude-HCS, using it to design an HCS system that can then be tested, fine-tuned, and generated automatically based on the model’s suggestions. …
Microsoft shivs OpenAI with new AI models for speech, images
… The models include: MAI-Transcribe-1, a speech recognition model that delivers "enterprise-grade accuracy across 25 languages at approximately 50 percent lower GPU cost than leading alternatives"; MAI-Voice-1, a speech generation model that can supposedly produce 60 seconds of audio in less than a … …
Google battles Chinese open weights models with Gemma 4
… During inference, a subset of the model's 128 experts, totaling 3.8 billion active parameters, is used to process and generate each token. So long as you can fit the model into your VRAM, it can generate tokens far faster than a dense model of equivalent size. …
PrismML debuts 1-bit LLM in bid to free AI from the cloud
AI + ML PrismML debuts energy-sipping 1-bit LLM in bid to free AI from the cloud Bonsai 8B model is competitive with other 8B models but 14x smaller and 5x more energy efficient PrismML, an AI venture out of Caltech, has released a 1-bit large language model that outperforms weightier models, with … …
JetBrains Air: agentic IDE built on abandoned Fleet
… Developers have asked for the ability to use local models such as Ollama or Qwen. …
Mistral's new agent proofs your code on the cheap
… Mistral also released Mistral Small 4 , designed as an all-in-one model that can handle reasoning, coding, and instruct/chat tasks, so you don't have to switch between specialized models. ® software france development ai and ml ai
Security reserchers tricked Apple Intelligence into cursing
… "Neural Exec uses an optimization algorithm to speed up the process of injecting the kinds of strings that could be execution triggers and would prompt the model to misbehave." While this type of adversarial input could theoretically work on any model, the smaller, on-device model used in Apple Int… …