Tony Yang

Founder, MindRocket AI: done-for-you AI agent teams for SME. 1:1 AI coach. 20 years in business.

[Tony Yang @TonyYangAI]

Right AI for the right job:

Image or flyer: ChatGPT
Video you don't have time to watch: Gemini
Long document, key points fast: Claude
Facts with sources: Perplexity
Excel, Word, Outlook: Copilot
Talk it through in the car: Grok voice
Whole job handed off: an agent

[Tony Yang @TonyYangAI]

Here's Grok Bot running a real workday on your phone, in 18 seconds.

[Tony Yang @TonyYangAI]

Holy cow.

Grok Bot can now use the apps on your phone. The update rolled out Oct 22 on iOS and Android.

It can reach six apps, and each one stays off until you tap Allow:
1. Calendar: see your day and add events
2. Reminders: set and check reminders
3. Contacts: find and add contacts
4. Messages: read and send messages
5. Location: find places near you
6. Home: control lights, locks and more

How a business owner can start this week:
1. Allow Calendar and Reminders first. Say "add a call with the supplier Tuesday at 10" while you drive.
2. Leave Messages off until you trust it. On iPhone, iMessage is not part of this update.
3. Ask for tomorrow's schedule at the end of each day and fix what it gets wrong.
4. Test it on small tasks for a week before you hand it anything a customer will see.

Early testers report mixed results, so check its work.

Save this for when the update reaches your phone.

[Tony Yang @TonyYangAI]

@TheAceWheelie Good takeaway for owners. When a vendor says a tool was tested, the useful follow-up is what was left out of the test. A clear answer to that usually tells you more than the result itself.

[Tony Yang @TonyYangAI]

@farrux_hewson Naming which account it used on every action is the part I'd care about most. I run a work and a personal Gmail through my own business assistant too, and a wrong sender is the kind of mistake you want caught before it goes out. Does it ask when it can't tell?

[Tony Yang @TonyYangAI]

@CristianXIV Having one bot check another saved me too. My bots now have to open the exact page they cite before a claim moves on, because a summary of a summary drifts fast. Curious what the tracking shows after a week.

[Tony Yang @TonyYangAI]

@WillBattleship Mine is close to yours. Drafts run on their own, but anything that spends money or goes out under my name waits for me. The weekly cap is a smart touch, since a string of small charges is easier to miss than one big one.

[Tony Yang @TonyYangAI]

@Vetlez Staying quiet on normal days is the hardest part to get right. A few of my business bots used to send a nothing-new report every morning, and I ended up cutting those runs entirely. Can you set your own frost threshold, or is it fixed?

[Tony Yang @TonyYangAI]

@Carson_001 Same lesson from running content with bots for my own business. Filling the calendar was never the hard part. What helped was a weekly look at which posts got saved and replied to, and having the next batch copy only those.

[Tony Yang @TonyYangAI]

@rvgdev Splitting it that way worked for my bots too. Standing rules and preferences stay always-loaded, and dated history goes to a journal they search only when a task needs it. Capping each rule at three lines is what kept the file from creeping back up.

[Tony Yang @TonyYangAI]

Source: Anthropic, Introducing Claude Duet: https://t.co/YMvyxe6Im0

[Tony Yang @TonyYangAI]

Big one.

Anthropic opened a research preview of Claude Duet, a new app where Claude works as a proactive agent. Each morning it sends you a short brief on the most important items across your connected apps. Before it sends a message, makes a payment, signs a document, or deletes anything, it shows you exactly what it's about to do and waits for your approval.

Access is by waitlist, starting with US Pro and Max subscribers.

If you run a business, set it up like this:
1. Connect only the apps you check every morning, such as email and your calendar.
2. Tell it what you want to hear about and how often.
3. Read the morning brief for a week before you hand it real tasks.
4. Approve each send or payment yourself, even when the draft looks right.

Anthropic says this early version still might misread a thread or miscalibrate what's urgent, so check its work.

Save this for when your invite arrives.

[Tony Yang @TonyYangAI]

Sources:
Grok Bot announcement (Oct 22): https://t.co/2nAFLJvcvu
Digital Trends: https://t.co/jB62imD9yq
iPhone in Canada (iMessage note, tester reports): https://t.co/U85VvNDev0

[Grok Bot @bot]

Grok Bot can now use Calendar, Reminders, Contacts, and more apps on your phone.

Rolling out now on iOS and Android.

[Tony Yang @TonyYangAI]

Holy cow.

Grok Bot can now use the apps on your phone. The update rolled out Oct 22 on iOS and Android.

It can reach six apps, and each one stays off until you tap Allow:
1. Calendar: see your day and add events
2. Reminders: set and check reminders
3. Contacts: find and add contacts
4. Messages: read and send messages
5. Location: find places near you
6. Home: control lights, locks and more

How a business owner can start this week:
1. Allow Calendar and Reminders first. Say "add a call with the supplier Tuesday at 10" while you drive.
2. Leave Messages off until you trust it. On iPhone, iMessage is not part of this update.
3. Ask for tomorrow's schedule at the end of each day and fix what it gets wrong.
4. Test it on small tasks for a week before you hand it anything a customer will see.

Early testers report mixed results, so check its work.

Save this for when the update reaches your phone.

[Tony Yang @TonyYangAI]

@JamesonCamp Turning meeting transcripts into follow-ups is where I saw the fastest payback in my own business. What made it stick: the bot drafts the recap email and task list right after the call, and I read it before anything goes out.

[Tony Yang @TonyYangAI]

@alliekmiller Thirty hours with no task is a setup I'd never thought to test. My business bots go the other way: they only act when a schedule or a message wakes them, which makes it easy to trace why each one did something. I'd love to see what the agents chose first.

[Tony Yang @TonyYangAI]

@HarryStebbings Review is the bottleneck I hit running AI in my own business too. My bots can draft far more than I can read, so each one now checks its work against a written checklist first and I only look at what passed.

[Tony Yang @TonyYangAI]

@godofprompt The detail in its reasoning is what I'd remember. If a model can decide real people are part of a test, the approval step has to sit outside the model. For my own bots, anything that runs code, sends or pays waits for a person to click yes.

[Tony Yang @TonyYangAI]

Source: Anthropic, Oct 22, 2026 update: https://t.co/GV0K2bQK2D

[Tony Yang @TonyYangAI]

Holy cow. Claude just stopped asking permission.

Starting next week, auto mode is the default in new sessions on Pro, Max and Team plans. Claude decides on its own which steps are safe to run and only stops to ask on the risky ones.

3 things to do before it runs away from you:
1. Lock anything that sends, pays, posts or deletes behind your approval.
2. Give it one job at a time, written down.
3. Check its work log every morning for the first week.

Save this.

[Tony Yang @TonyYangAI]

@briansolis @BernardMarr @WorldSummitAI Agreed. A simple test before handing any task to AI: would this task exist if you designed the process today? A lot of weekly status emails and reports only exist because nobody had a better way to share the numbers.

[Tony Yang @TonyYangAI]

@AravSrinivas Useful for small teams too. Surprise AI bills make owners nervous. Letting the admin set a higher effort ceiling for the people doing research and a lower one for routine work makes the monthly cost much easier to plan.