Gmail and Docs are getting voice-driven AI modes. Gmail Live lets you search and interact with your emails conversationally on mobile. Docs Live lets you create and edit documents by talking to them. Google Keep gets the same treatment. Speak your thoughts, and it organizes
LLMS
-
Gemini 3.5 Flash Released with Improved Performance Benchmarks
By
–
Gemini 3.5 Flash is the new default model. It beats 3.1 Pro across coding, agentic, and multimodal benchmarks while running 4x faster than other frontier models in output tokens per second. Available now in the Gemini app, Search, Antigravity 2.0, and the API. 3.5 Pro is in
-
Key AI Announcements from Google I/O 2026
By
–
Google I/O 2026 just satisfyingly answered every question about where Gemini is heading. New models. A personal AI agent. Compute-based pricing. Background agents that monitor the web for you. Here's everything that matters for AI operators:
-
TransformerLab: An Open-Source OS for AI Research Labs
By
–
The Operating System for Al Research Labs.
— Akshay 🚀 (@akshay_pachaar) 20 mai 2026
TransformerLab orchestrates GPUs across any cloud and runs any training or eval workflow you define
Pre-built templates for LoRA, DPO, GRPO, MMLU, HellaSwag.
Use it from a GUI, CLI, or agent skill.
100% open-source and FREE to use. pic.twitter.com/Y7ZnUpT8T9The Operating System for Al Research Labs. TransformerLab orchestrates GPUs across any cloud and runs any training or eval workflow you define Pre-built templates for LoRA, DPO, GRPO, MMLU, HellaSwag. Use it from a GUI, CLI, or agent skill. 100% open-source and FREE to use.
-
Core components of agentic AI: from thinking to doing
By
–
Agentic AI = systems, not models • Intelligence (LLMs)
• Memory
• Tools
• Feedback loops That’s how AI moves from thinking → to doing Via Giuliano Liguori (
@ingliguori
) #AI #AIAgents #AgenticAI -
Shanghai AI Lab releases SU-01 30B model with advanced reasoning capabilities
By
–
A 30B model just hit gold-medal scores at the world's hardest math contest. Olympiad math and physics are the hardest reasoning tests around. Most gold-medal scores come from massive specialized systems built for one subject. Shanghai AI Lab just released SU-01, a 30B open
-
The Evolving Role of Developers in the Age of AI Agents
By
–
Developers don't get paid just for coding! If your only skill is writing code, an agent can do that for $20 a month. And it will only get cheaper from here. Who gets paid now is the developer who knows what to build, what to skip, and what to fix when an agent silently gets
-
The shift from static AI research to interactive active learning
By
–
IMO deep research has been ~dead since o3 and interactivity was always more impt for active learning and eliciting intention thoughtless prompt -> long ass report nobody reads is inferior to
read -> think -> ask -> read -> think -> ask -
New Research Survey: Code as a Fundamental Medium for AI Agents
By
–
code as agent harness. a 102-page survey from Stanford, Meta, and UIUC on agent harnesses. the paper argues that code is no longer just the thing agents produce. it’s the medium through which they reason, act, and represent their environment. it calls this “code as agent
-
LLMs empower users and reshape professional roles
By
–
here, the point is that llms allow normal people to deal with complex problems, eg. by translating legalese or writing sophisticated emails, and smart people can use llms as an army of interns, but people who work as professional prompt completers are terrified