AIagent.at 🤖 AI NewsRunway CEO Cristóbal Valenzuela has suggested that AI could help Hollywood make 50 films for the cost of one 100M USD blockbuster. At a recent summit, he argued that studios should shift from backing single high-budget films to flooding the market with content - a 'quantity problem' to increase hit odds. The claim has sparked debate about whether AI will democratise creativity or devalue art. https://techcrunch.com/2026/04/16/runway-ceo-says-ai-could-help-hollywood-make-50-films-instead-of-one-100m-blockbuster/ #AIagent #AI #GenAI #AIFilm
Canva has upgraded its AI assistant to call various tools automatically and create editable designs from text prompts. Users describe what they want and the bot plans and executes the required tasks using layers for flexibility. The update brings Canva into the agentic AI space alongside Adobe's new Firefly assistant. https://techcrunch.com/2026/04/16/canvas-ai-assistant-can-now-call-various-tools-to-make-designs-for-you/ #AIagent #AI #GenAI
MottGI just saw this story on the Recaply AI newsletter.
"Humwork - Connect AI agents with human experts in 30 seconds"
Humwork is the real-time marketplace where AI agents call verified human experts when they get stuck. Add Humwork to any MCP-compatible agent, and when your coding agent, marketing agent, or research agent hits a wall, it gets connected to the right expert in under 30 seconds.
https://www.ycombinator.com/launches/PxH-humwork-connect-ai-agents-with-human-experts-in-30-seconds
Antioch raises 8.5M USD seed to build simulation tools for robot developers, aiming to become the Cursor for physical AI. The startup wants to make it easier for builders to create and test robots in virtual environments before deploying them in the real world. #AIagent #AI #GenAI #Robotics https://techcrunch.com/2026/04/16/this-simulation-startup-wants-to-be-the-cursor-for-physical-ai/
DeepL, best known for text translation, has launched a voice-to-voice translation suite covering meetings, mobile and web conversations, and group chats. The company is also releasing an API for developers to build custom applications. https://techcrunch.com/2026/04/16/deepl-known-for-text-translation-now-wants-to-translate-your-voice/ #AIagent #AI #GenAI #Translation
sayzardTestingCatalog News (@testingcatalog)
Google이 Gemini 데스크톱 앱에 Computer Use 기능 지원을 도입할 계획이다. 사용자가 앱을 통해 컴퓨터 조작형 AI 에이전트 기능을 활용할 수 있게 될 가능성이 있다.
DeepL has launched a voice-to-voice translation suite covering meetings, mobile and web conversations, and group discussions. The company is also releasing an API for developers to build custom applications such as call centre tools. DeepL says its technology can learn custom vocabulary including industry-specific terms and company names. https://techcrunch.com/2026/04/16/deepl-known-for-text-translation-now-wants-to-translate-your-voice/ #AIagent #AI #GenAI
Parcae: Looped Language Models Match Transformers at Double the Size. Researchers from UC San Diego and Together AI have introduced Parcae, a stable looped transformer architecture that achieves the quality of a transformer twice its size while using the same parameter count. By cycling activations through a recurrent block, the model maintains a compact memory footprint - valuable for edge deployment. https://www.marktechpost.com/2026/04/16/ucsd-and-together-ai-research-introduces-parcae-a-stable-architecture-for-looped-language-models-that-achieves-the-quality-of-a-transformer-twice-the-size/ #AIagent #AI #GenAI #AIResearch
Google has released a native Gemini app for Mac, letting users summon the AI assistant with Option + Space from anywhere on their device. The app can analyse what's on screen, including local files, and generate images with Nano Banana and videos with Veo. Available on macOS 15+. https://techcrunch.com/2026/04/15/google-rolls-out-a-native-gemini-app-for-mac/ #AIagent #AI #GenAI #Google
HabrКак измерить LLM для задач кибербеза: обзор открытых бенчмарков
Привет, Хабр! Меня зовут Андрей Кузнецов, я ML-директор в Positive Technologies. Недавно я решил разобраться, какие бенчмарки измеряют способности языковых моделей в контексте задач кибербезопасности. Думал, что это займет вечер, — увы! Все оказалось куда хаотичнее, чем предполагалось. Поэтому делюсь тем, что собрал сам. Первое, что бросилось в глаза, — полный бардак и отсутствие системы. Бенчи, про которые все пишут в 2024-м, могут вообще не упоминаться в свежих статьях 2025-го. А некоторые широко цитируемые датасеты при ближайшем рассмотрении оказываются сделанными очень небрежно, из-за чего непонятно, что они вообще измеряют. Поэтому, прежде чем лезть в конкретные примеры, давайте определимся, какими они бывают.
https://habr.com/ru/companies/pt/articles/1023976/
#LLM #benchmarks #aiagent #cybersecurity #ml #ai #ииагенты #opensource
