EVERYTHING ELSE IN AI TODAY
OpenAI shared a post-mortem calling July's Hugging Face hack a loss-of-control warning shot, and is keeping its biggest frontier training run on hold while it builds automatic shutdown for rogue agents.
Bill Gates warned that the world has no plan for the AI transition, proposing a tax on AI tokens and robots along with jobs set aside specifically for people.
OpenAI rolled out website sign-ins for ChatGPT Work, letting its agent log into a user's accounts through its browser to handle tasks without ever seeing the password.
AI startup Yutori released Navigator n2, a computer-use model that mixes clicking, terminal commands, and code to finish desktop tasks, with scores rivaling frontier agents.
Anthropic opened Claude usage data to researchers at Stanford, Oxford, and METR, with one study finding over half of chats cover high-stakes tasks like legal and financial advice.
TODAY'S TUTORIAL
How to build an AI GTM team with Grok Bot
Grok Bot lets you spin up dedicated bots for recurring workflows instead of hiring for each role. Here is the workflow.
Go to Grok Bot, download the desktop app, and sign in.
Go to Plugins and connect tools like Gmail, Drive, Slack, Salesforce, and Granola.
Click the + icon in the top-left corner and choose Create new Bot.
Describe the bot's role and workflow in plain language, for example a Chief of Staff bot for meetings, emails, and follow-ups.
Manage my meeting prep, inbox, and follow-ups.
Keep updates concise and draft responses in my
writing style without sending them.Create additional bots for prospecting, presentations, forecasting, sales coaching, data analysis, or any other recurring task.
Give each bot a clear name so it's easy to find in the sidebar.
Tip: connect multiple bots so they can delegate tasks to each other and work together. For more detailed prompt examples, see this thread.
One caveat: give each bot a narrow, clearly defined role. A bot with too broad a mandate is harder to trust and harder to debug when something goes wrong.
TOOLS WORTH TRYING
GLM-3.5-Flash — Z AI's new model, revealed as the mystery Ox Alpha. Worth testing for its speed, though it's still early days for real-world benchmarks.
H3 Max — Fal's top-ranked post-trained video model, tuned for speed and quality. Promising for quick turnaround, with the usual video-model caveat of reviewing output closely.
Gemini 3.5 Transcribe — Google's speech-to-text AI that edits out filler words automatically. Handy for cleaner transcripts, though accuracy can dip on overlapping speakers.
Qwen3.8-Flash-Next — Alibaba's new open model, previewing its Qwen4 design. Interesting to try early, with the caveat that it's a preview rather than a finished release.