OpenAI Responds to Axios Developer Tool Compromise

#1 📝 OpenAI News

Our response to the Axios developer tool compromise - OpenAI describes its response to a compromise involving the Axios developer tool, outlining security actions taken, investigation efforts, and mitigation steps to protect users and systems.

#2 𝕏

DeepLearning.AI: Google launched Lyria 3, an AI music generator that transforms text prompts or images into original 30-second songs.

#3 𝕏

Thariq launched /ultraplan in Claude Code, which auto-generates editable implementation plans you can run on the web or in your terminal. It’s now in preview for all users with CC on the web enabled.

#4 𝕏

Claude launched a beta of Claude for Word, enabling drafting, editing, and revising documents directly from the sidebar with tracked changes and preserved formatting. It’s available now on Team and Enterprise plans.

#5 ▶️

Claude Mythos is too dangerous for public consumption...

Fireship

#6 𝕏

Qwen rolled out Code v0.14.0–0.14.2, adding remote control via Telegram/DingTalk/WeChat, Cron-job scheduling for recurring AI tasks, sub-agent model selection, and follow-up suggestions after each run. It also launched Qwen3.

#7 𝕏

Google AI added Notebooks in Gemini App via NotebookLM for private context retrieval and chat-grounded research, and introduced customizable 2D/3D interactive visualizations in Gemini web chats.

#8 𝕏

LlamaIndex 🦙 LiteParse has gained 4K+ GitHub stars in 3 weeks and can parse ~500 pages in 2 seconds—no GPU or API keys needed, with support for 50+ file formats.

#9 𝕏

Google AI spotlights fun builder projects powered by last week’s open-source Gemma 4 models. Examples include automated meeting summarizers, code-generation assistants, and multilingual chatbots, each detailed with tool integrations, performance stats, and user insights.

#10 𝕏

Lenny Rachitsky demos a clever integration of OpenClaw with his Product Pass (lennysproductpass.com) to bypass token limits, leveraging deployments on Railway, Ampcode, FactoryAI and Warp.dev.

#11 📝 Simon Willison

Voice mode is weaker - Simon observes that OpenAI's voice mode appears to run on an older, weaker model, leading to surprising differences in capability depending on access point. This reflection was inspired by an Andrej Karpathy tweet about how different domains and reward functions drive divergent model improvements.

#12 𝕏

Santiago launched Seedance 2.0, a unified audio-video generation model that simultaneously produces fully synced clips instead of adding audio to silent footage.

#13 𝕏

Santiago warns that exploding demand for data center capacity is far outpacing our ability to bring new facilities online. This growing supply shortfall means something’s got to give.

#14 𝕏

claire vo 🖤 has been helping large companies go AI-native with a custom Claude wrapper, seamless connector integrations, and a skills marketplace—and she stresses the missing ingredient is fun for everyone, execs included.

#15 𝕏

Logan Kilpatrick launched his latest Live voice model, which hit #1 on the Tau Voice Bench, demonstrating production-grade voice AI usability.

#16 𝕏

Guillermo Rauch points out that Claude Code already accounts for roughly 4% of all public GitHub commits, so its reported 30% share of platform deployments is shockingly high.

#17 𝕏

Guillermo Rauch rolled out the new AI docs for Workflow.dev (https://useworkflow.dev/docs/ai) and invited PM builders to try it out and share feedback on their use cases.

#18 𝕏

clem 🤗 points out that as building websites and apps becomes trivial, real competitive edge now lies in training, running, and optimizing AI models.

#19 𝕏

Sebastian Raschka observes that for most friends and family the primary AI touchpoint is Apple Intelligence on new iPhones, while ChatGPT is dismissed over hallucination rumors and they default to Google.

#20 𝕏

AI at Meta announces the upcoming Muse Spark API, fueling developer excitement to experiment with Muse Spark inside their agentic harnesses.

#21 𝕏

Harrison Chase likens agent harnesses to Spark and positions LangSmith as the Databricks of agent abstractions, quoting @bllchmbrs’ analogy of them as stable building blocks.

#22 𝕏

Harrison Chase confirms that long-term memory in AI agents will be a major focus going forward.

#23 𝕏

Qwen launched mobile Qwen Code bots for DingTalk, Telegram and WeChat, letting you send commands like “check the logs for errors in /var/log/app” from your phone and get real-time dev-server responses—no SSH or laptop needed.

#24 𝕏

Logan Kilpatrick announces the new live/audio model has jumped to the top of the TAUbench voice leaderboard, delivering much faster inference speeds than previous generations.

#25 𝕏

Claude launched Claude for Word, which shares context with Claude for Excel and PowerPoint, enabling seamless multi-document workflows in a single conversation.