Good Morning! Here's what I have for you in today's newsletter:

xAI launched email for Grok Bot, giving each Bot its own inbox at mail.grokbot.com to sign up for services, contact businesses, and arrange meetings.

Booking enquiries, meeting arrangements, and account sign-ups can now go through your Bot’s mailbox, leaving yours out of the back-and-forth. Sending limits and checks before outgoing messages haven’t been detailed yet, so start with simple tasks you can easily review.

OpenAI introduced composer predictions in Codex, a beta for Pro users that drafts your next message using the conversation so far and the way you usually write.

“Run the tests”, “Fix that error”, “Try again”. Those little follow-ups add up during a long coding session. Codex can now have one drafted when you’re ready to respond, cutting down on repetitive typing. Access is limited to Pro users for now.

Alibaba released Qwen-Image-2.1-Turbo, a faster version of Qwen-Image-2.1 that creates and edits images in eight denoising steps, with open weights and a hosted API.

When you’re testing dozens of image ideas, the wait between versions adds up. Eight-step generation promises a quicker turnaround on the same GPU, making it easier to try another variation. Before using it commercially, check Turbo’s license: the base model’s weights came with a research license.

Let Claude define your criteria and build the checks so Jev sorts your work.

Jev returns choices, scores, and probabilities rather than written responses. Use it for quick judgments you repeat often, like whether an email needs a reply today. You'll need Claude Code, a Mac with Terminal, and a TypeSafe account.

Create a TypeSafe account and API key, then enter the key in Terminal yourself and keep it out of chat. Install the official skill with 

Input costs $0.042 per million tokens, with output free. New accounts receive $5 in signup credit.

Ask Claude to interview you about one decision, such as which emails need a reply today. Give it three emails you would answer, three you would ignore, and a couple that were hard to call. Approve the written definition once it names specific criteria.

Have Claude turn the definition into separate checks, each with a yes, a no, an example, and a borderline case. Code handles exact rules, like a sender in your contacts, and combines the results. Ask for a range around 50 percent that goes to a maybe pile.

Export 12 emails you already reviewed, with your call in a my_call column. Run the checks without that column, then compare Jev's decisions with yours and read each disagreement before changing anything.

Fix the check or weight behind each disagreement and rerun the sample. Then test 8 fresh emails you also labeled, and run the full folder only after those results satisfy you.

P.S. You can access all my AI courses, trainings, prompts library, and $500+ worth of AI tool discounts here.

Anthropic opened Claude Code Projects to every Pro and Max user on the waitlist. Split work from one conversation into parallel cloud threads that keep running after you close your laptop, with a four-minute walkthrough to get started.

ElevenLabs added Brand Kit to ElevenCreative. Save your logo, fonts, and colors once, then apply them to future creations—so your next product launch campaign comes back looking like your brand.

OpenAI shared progress on ChatGPT for Teens, with protections enabled automatically for accounts identified as under 18. It also previewed College Planner for tracking US college application deadlines and financial-aid steps.

🦞 OpenClaw: v2026.10.1 makes installed skills easier to find, puts MCP Apps beside chat, adds full backups to external storage or Cloudflare R2, and simplifies Mac setup.

🧩 Nous Portal: Try MISSINGNO., a model built for coding and agent reasoning that uses fewer tokens. Available free for a limited time inside Hermes Agent.

🌐 Scale AI: Released AgentEnv as open source, giving developers access to the framework it uses to build reinforcement learning environments for training agents.

👁️ Gemini Live: Guided Vision describes the camera view aloud and helps users adjust the framing in real time. Built with the blind and low-vision community.

Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!

See you tomorrow :)

- Dr. Alvaro Cintas

✓ Full archive of premium guides with ready-to-use prompts

✓ Structured AI courses (step-by-step, start-to-finish)

✓ Every upcoming premium tutorial

✓ $500+ worth of AI tools