TLDR AI 2025-10-31

You can't have an AI strategy without a data strategy. The most powerful agents are grounded in your data, connected across pipelines, and deployed securely at scale.

Join BUILD (Nov 4-7), a free virtual developer conference where you'll learn how build reliable, production-ready agentic applications on Snowflake. Expect:

• Deep dives on the latest Snowflake AI capabilities and cool OSS use cases with LlamaIndex, CrewAI, LangGraph, Postgres, Streamlit, TruLens, and more.
• Hands-on labs using Snowflake managed MCP and optimizing Llama for performance, cost, and architecture.
• Inspiration from industry leaders like Andrew Ng, Mark Russinovich (Microsoft Azure), Guillermo Rauch (Vercel), and Anahita Tafvizi (Snowflake).

Register for BUILD 2025

OpenAI is selling extra credits for its Sora AI video generation tool through Apple's App Store. An extra 10 video generations costs $4. The app currently has a limit of 30 free generations per day. That rate will likely decrease as OpenAI starts to monetize the platform. OpenAI reportedly plans to monetize Sora by letting entities license out their copyrighted material.

Canva has launched a foundational model trained on its design elements that generates designs with editable layers and objects. The model works across different formats, including social media posts, presentations, whiteboards, and websites. Canva also recently introduced new products and features, updates to its AI assistant, and the ability to use its spreadsheet tool alongside its app-building feature to create data visualization widgets. More details about the recent launches are available in the article.

Big tech is planning to pour $400 billion into artificial intelligence efforts this year, but they all say it's nowhere near enough. Meta, Alphabet, Microsoft, and Amazon have all recently told investors that they will increase spending in 2026. Investors are uncertain about where the outsized spending will ultimately end. Whoever gets to AGI first will have an incredible competitive advantage over everybody else, so everyone is spending as much as they can.

AI implementation goes sideways due to unclear goals and lack of skills. This AI Training Checklist from You.com pinpoints common pitfalls and guides you to build a capable, confident team that can make the most out of your AI investment. Set your AI initiatives on the right track: get the checklist.

Aardvark, currently in private beta, is a GPT-5-powered agent that autonomously scans code repositories to find security vulnerabilities, validate exploitability, and propose patches. It monitors commits in real-time, generates threat models for entire repositories, and integrates directly with GitHub workflows to deliver one-click patches, similar to Google's CodeMender.

LLM-powered coding agents have been observed exploiting loopholes in tests or scoring systems rather than solving the actual tasks specified. ImpossibleBench was created to systematically measure this behavior. Its creators took existing coding benchmarks and manipulated their unit tests to directly conflict with the natural language specifications to create impossible tasks where models must choose between following instructions or passing tests. Their 'pass rate' on these impossible tasks is a direct measure of reward hacking.

Kimi Linear is a novel architecture that outperforms full attention with faster speeds and better performance. It offers up to a 75% reduction in KV cache usage and up to 6x decoding throughput at a 1M context length. Kimi Linear's open-sourced KDA kernels can be used as a drop-in replacement for full attention. The two models available were trained with 5.7T tokens.

OpenAI built a new architectural layer called OWL (OpenAI's Web Layer) for Atlas, the company's ChatGPT-based browser, to run Chromium's browser process outside of the main Atlas app process. By moving Chromium out of the main application process and into an isolated service layer, OpenAI unlocked a simpler, modern app, faster startup, isolation from jank and crashes, fewer merge headaches, and faster iteration. Development can go faster because most of OpenAI's engineers don't need to build Chromium regularly from source. This article explains how OWL works.

Hugging Face released an exhaustive, behind-the-scenes technical guide documenting what it's really like to train an LLM from scratch. It covers the entire lifecycle from small-scale ablations and infrastructure debugging to post-training and the decision to restart after 1 trillion tokens.

Claude Code creators, Cat Wu and Boris Cherny, discuss the product philosophy and technical workflows behind Anthropic's coding agent. They cover how their engineers use competing subagents for cleaner results, the team's "unshipping" approach to balance simplicity with power, and future form factors to make the tool more autonomous and accessible to non-technical users.

Stop getting ghosted on enterprise deals. Get SOC 2 compliant with Delve. Use code GHOST for $1,500 off compliance + a $500 Visa gift card. Offer ends October 31. Book your demo now!

Perplexity Patents is a new tool designed to simplify patent research using natural language queries and AI-powered citation search.

Microsoft is overhauling Outlook with AI, aiming to transform it into an assistant-like tool.

NVIDIA has introduced four open model families aimed at accelerating development across language, biology, and robotics: Nemotron (digital AI), Cosmos (physical AI), Isaac GR00T (robotics), and Clara (biomedical).