Google Introduces Gemini 3.1 Pro

#1 𝕏

Google AI launched Gemini 3.1 Pro, doubling core reasoning performance to 77.1% on the ARC-AGI-2 benchmark and demoing its ability to code full environments with integrated generative audio and UI controls.

Also covered by: @Cognition, @Simon Willison, @Jeff Dean, @Peter Yang, @There's An AI For That, @Google DeepMind, @Google DeepMind, @Demis Hassabis, @Sundar Pichai, @Sundar Pichai

#2 𝕏

Qwen launched Qwen3.5-Plus on Qoder, bringing faster processing along with advanced coding, reasoning/agent capabilities and multimodal support.

#3 𝕏

Jeff Dean unveiled Lyria, a new generative music model that builds on Nano Banana (+Pro) for images and Veo for video, enabling users to create original tracks from text prompts, images, or video in seconds.

#4 𝕏

Claude launched PowerPoint integration on its Pro plan, adding connectors that pull context from your daily tools straight into your slides.

#5 𝕏

Josh Woodward notes that most small businesses lack the time and money for a traditional photoshoot, so Google Labs launched Pomelli—a free tool that generates pro-grade visual assets in seconds.

#6 𝕏

Aravind Srinivas unveiled Comet iOS, a Safari-grade browser with Perplexity powering every webpage for smooth, AI-assisted browsing, now available for pre-order on the App Store.

#7 𝕏

Aravind Srinivas upgraded Gemini 3 Pro to Gemini 3.1 Pro for all Perplexity Pro and Max users (consumer & enterprise), and it’s now the second most adopted enterprise model after Claude 4.5 Sonnet/Opus.

Also covered by: @Cognition, @Simon Willison, @Jeff Dean, @Peter Yang, @There's An AI For That, @Google DeepMind, @Google DeepMind, @Demis Hassabis, @Sundar Pichai, @Sundar Pichai

#8 𝕏

Guillermo Rauch announced that video is now supported on the Vercel AI Gateway alongside a new `generateVideo` API in the @aisdk, and he’s offering Grok Imagine Video & Image for free through February 25.

#9 𝕏

Cursor rolled out agent sandboxing on macOS, Linux, and Windows, letting AI agents run in a secure, isolated environment and only request permission when they need to step outside it.

#10 in

Guillermo Rauch unveiled a Node.js-optimized WebStream implementation that autonomously delivers up to 14.6× performance gains with 1100/1116 WPT tests passing, and will be upstreamed to Node.js for everyone’s benefit.

#11 📝 Surge AI Blog

EnterpriseBench: CoreCraft – Measuring AI Agents in Chaotic, Enterprise RL Environments - Surge built CoreCraft, a large-scale simulated startup world, to evaluate AI agents on realistic, messy enterprise tasks rather than tiny lab environments. The benchmark aims to push agents from controlled testbeds into chaotic, real-world enterprise scenarios.

#12 𝕏

Sebastian Raschka built Tiny Aya from scratch: a 3.35B-parameter multilingual decoder transformer featuring SwiGLU, Grouped Query Attention, and parallel transformer blocks. Its compact size and broad language support make it ideal for on-device translation.

#13 𝕏

Santiago explains an AI workflow that uses MongoDB’s native vector search to compare incoming requests against the system’s stored experience in a vector database.

#14 𝕏

Teresa Torres breaks down how ShowMe built AI-native digital sales reps as full-fledged teammates using a multi-agent system—detailing its architecture, evaluation metrics, and roadmap for scaling.

#15 𝕏

LlamaIndex 🦙 tested GPT-5.2 at four reasoning levels on complex document parsing and found higher reasoning slowed processing 5× (241s vs 47s) and spiked costs without improving its ~0.79 accuracy. Their LlamaParse Agentic model instead ran 13× faster at 18× lower cost.

#16 📝 PromptLayer Blog

SuperClaude: How Structured Prompts Turn Claude Code into a True Development Partner - Introduces SuperClaude, a community framework that improves consistency and expert-level outputs from AI coding assistants by using structured prompts. It addresses the gap between an LLM's raw potential and reliable performance on complex coding tasks.

#17 𝕏

NVIDIA AI launched its 2026 State of AI in Telecom Report, revealing how AI has become the core growth engine powering telecom operations, networks, and services.

#18 𝕏

Lenny Rachitsky: Claude Code, launched just a year ago, now writes 4% of all GitHub commits. In conversation with @bcherny he argues coding is “largely solved,” previews the next wave of tech roles, and shares a counterintuitive bet alongside practical tips.

#19 𝕏

Google Research with Harvard, Mount Sinai & Boston Children’s applied Earth AI superresolution to fill public‐health data gaps and estimate zip‐code–level measles vaccination rates, pinpointing undervaccinated hotspots linked to outbreaks.

#20 📝 PromptLayer Blog

The Emergence of Agent-First Software Design - Argues there's a paradigm shift from explicit decision-tree programming toward agent-first designs where engineers orchestrate agents rather than hard-code every branch. The post introduces the idea that the role of software engineers is evolving as agents handle more of the decision-making.

#21 📝 Simon Willison

SWE-bench February 2026 leaderboard update - Discusses a fresh run of the SWE-bench leaderboard against current models, explains the 'Bash Only' benchmark and Verified subset, lists top-performing models, and notes observations about fairness and harness differences.

#22 📝 Ampcode Chronicle

The Coding Agent Is Dead - Amp is removing the editor extension (the Coding Agent) as part of a shift in product direction to step into the future. The announcement signals a deliberate end to that integration and a push toward newer approaches.

#23 📝 Mario Zechner

Österreichische Arbeitsagentur veröffentlicht fragwürdigen KI-Chatbot - Heise reports on the Austrian employment agency releasing a questionable AI chatbot, covering concerns about its reliability and implementation. Highlights criticism from experts and users.

#24 𝕏

I’m unable to view external links—could you please paste the full text of the X announcement here? Then I can craft the 1–2-sentence summary per your format.

#25 𝕏

Peter Yang released a full prototyping tutorial for Gemini 3.1 AI Studio, now available as a detailed written guide.