Good Morning! Here's what I have for you in today's newsletter:
OpenAI opens Finances in ChatGPT to Free and Go users in the U.S.
xAI adds an Agent Dashboard to Grok Build for running agents in parallel
ElevenLabs brings its Eleven v4 voice model to ElevenReader
Build your own JARVIS with OpenAI Dots, drafting while you're away
4 new AI tools worth trying today
AI TOOLS
OpenAI opened Finances in ChatGPT to Free and Go users in the U.S. You can connect bank, card, credit, and investment accounts through Plaid and Experian, then ask questions about your own finances.
ChatGPT can spot forgotten subscriptions, flag unfamiliar or duplicate charges, track recurring bills that have gone up, and send a weekly account summary.
You can see what's driving your monthly expenses, build a budget around what you actually spend, and track your credit score and what's affecting it.
Planning tools help you work through debt repayment, estimate an emergency fund, check whether a big purchase fits your budget, and see how concentrated your investments are across accounts.
“Where did my money go this month?” is now a question Free users can answer using their own account activity. A forgotten subscription or a bill that quietly crept up gives you something concrete to act on. Before linking accounts, check what data you're sharing and how ChatGPT stores it.
AI AGENTS
xAI added an Agent Dashboard to Grok Build, its terminal coding agent. Type /dashboard to see all your running agents on one screen and start new tasks from the same view.
You can launch several agents from the dashboard, giving each a different task to work on at the same time.
Press Ctrl+W to run an agent in its own git worktree, with a separate branch and working copy for its changes.
Grok Build also remembers project conventions and earlier decisions across sessions, saving them as markdown notes it can refer to later.
Give one agent a bug fix, another a test suite, and a third a refactor, then follow their progress from one place. Separate worktrees give each agent its own copy of the project, so you can review each set of changes before bringing them together.
AI MODELS
ElevenLabs brought Eleven v4, its most expressive voice model, to ElevenReader. It reads articles, ebooks, and PDFs aloud in your chosen voice, with support for more than 90 languages.
Eleven v4 adjusts tone, pacing, and emotion to suit the text, making longer reads, such as ebooks and research papers, sound more natural.
The model keeps the same speaker identity throughout long recordings, so the voice you hear in chapter one stays consistent through the final chapter.
Supported languages include Cantonese, Mongolian, and Odia, along with the major languages available in earlier versions of ElevenReader.
That growing pile of saved articles and unread PDFs can come along on your next commute or walk. More natural delivery and a consistent voice could make longer listening sessions easier to stick with, whether you're catching up on course material or finishing an ebook.
HOW TO AI
How to Build Your Own JARVIS With OpenAI Dots
OpenAI Dots can check your schedule, sort your inbox, prepare meeting notes, and draft messages while you're away. You get one Dot per account, working for you.
Step 1: Set the rules
Create your Dot on desktop and skip app connections until it has instructions. Tell it to read, summarize, and draft only for the first week, and to ask before sending, posting, accepting invites, or spending money.
Go to Settings, then Personalization, then Permissions, then Custom rules. Set sending and invites to Ask before taking action, and file edits and payments to Hand off to you. Drafting doesn't authorize sending.
Step 2: Connect your apps
Connect only what your first job needs: calendar, mail, and tasks. At launch, your Dot can't have its own email address.
Test access by asking for your next three events, your newest email, and anything it can't reach. For logins, use the private sign-in form.
Step 3: Teach it how you work
Have your Dot ask five questions about your priorities, urgent contacts, hours, formats, and voice. Approve a one-page note, then ask exactly what it saved.
Add two rules: ask for the report format first, and check sent mail before flagging an email.
Step 4: Give it standing jobs
Schedule a weekday 07:30 morning brief with meetings, prep notes, emails needing replies, deadlines, and conflicts, under 250 words. Check it in Scheduled. Send routine updates to ChatGPT, decisions to Slack or your phone.
Step 5: Review, then loosen
Week one is drafts only, week two adds inbox drafts, and week three adds a weekly PDF and one limited action. Each week, have your Dot list its mistakes and suggest three instruction changes. Allow one new action at a time.
To stop the workflow, check Activity and Scheduled.
P.S. You can access all my AI courses, trainings, prompts library, and $500+ worth of AI tool discounts here.
Anthropic added the You Should Know plugin to Claude Code. It checks Claude's output for important details you might overlook. Turn it on with /plugin enable cc-plugin-you-should-know@builtin.
Google announced changes to Gemini model access starting October 9. Free users will be limited to Flash-Lite, and AI Plus will lose Pro. AI Pro and Ultra keep all models, with Deep Think also coming to AI Pro.
Aleph Alpha released Kolibri, a European open-weight model with 78B total parameters and 3.46B active. It supports up to 1M tokens of context and uses an Apache 2.0 license, so teams can run it on their own hardware.
🎵 Suno: launched Speech Beta, an audio model that generates voice and music together as a single track.
🛍️ Shopify: launched Canvas, a visual workspace where merchants design custom online stores with Sidekick, Shopify's AI agent.
🌐 Cohere: released North Small Translate, an open machine translation model that Cohere ranks best in the world under 1T parameters.
🔤 Timeless: released a free variable type family with Grotesk, Sans, Serif, and Text cuts, each with matching italics.
THAT’S IT FOR TODAY
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