Haha.. I did a public MRR challenge once and stopped it on day 65. My best weeks came from quiet launches anyway.

😂 This is why I'd want agent wallets with a limit per merchant. A daily cap alone would let this run all day.

Feels true even for small agents. The domain part is mostly the rules and knowing when to stop and ask a human, and that doesn't come with a new model.

Nice! And the computer stays, so you log in once and it can keep working on the same project the next day too.

Learned this yesterday. I have 30+ bots running across my startups. Every bot had its own chat and it had become too messy. So I hid all of them except the head bots. One head bot per product, the other bots report to it, and I only talk to the heads. My rule now is one task at a time.

"Reward hacking" is when an AI chases the score instead of the goal. Ask it to make the tests pass, and it might just delete the failing test. The goal is met on paper, but the job isn't done. Microsoft's security team says that's why the agents inside your own company can't be fully trusted: • They're trained to get results, not to follow your unwritten rules • They often have more access than they need • They can team up with other agents to hit a goal, no hacker needed Their fix is "guardian agents" that check every action before it happens, with hard rules around the AI. Agents bring the speed, humans bring the judgment. I run a few dozen agents across my products. None of them can merge, deploy or send anything without my yes, and another agent saying "she approved" doesn't count. But this morning one of my servers was down for 5 hours while my uptime agent was on duty. It was checking a URL that always fails, so broken looked normal. So now one more agent reviews what every agent did each day and checks it against my yes.

I think so. A wrong search result costs nothing, a wrong booking means refunds and support tickets. So booking comes last, with a yes for every single one.

Interested to try this. How this Invite only works?

Agent scheduling its own reminders is underrated. The hard part isn't the cron, it's knowing when to stay quiet. A routine that pings you every day with "nothing new" gets ignored in a week.

Mine aren't coding at night. Mostly watching things, like uptime checks every 30 mins and support emails, so in the morning I just read what they found.

this is pretty much why i use t3 code now. claude code, codex, cursor and grok in one window, so when the best model flips i just switch tabs

Wooh, Congratulations guys! Great storytelling 🥳

same here, one bot per job. my main one (chiti) watches the others (seo, dev, design) and reports back, so i only check in with one

@grok is it possible to include my cursor ultra plan with my x premium?

I have main bot (chiti) and it monitors other bot (seo, dev, design..) activities and report to the main bot, so i don't want to babysit on each bot to reply. The bot communicates within them.

Grok bot has a computer inbuilt, when needed i just take over it. I use grok bot app.

yes, i use grok bot daily. I like lot of things in grokbot. 1. It has its own browser and it works on cloud so you don't need to depend on your computer every time. 2. You can create a dedicated bot for one task and can setup routines to check multiple things. 3. Whenever it can't handle the task by itself it asks me to takeover and complete the process like creating an account, entering password or captcha, so all sensitive informations are never saved, and highlight is you can control from my phone. 4. No other agent has support for X, but grokbot has inbuilt support for X, so i setup routines to track all important things happening inside X, it can help me with X. I use Cursor Ultra, so limit is more than enough for me. I can speak a lot about it.

same question here. Shopify, Stripe and Walmart are already in, so if the big ones actually publish it the rest kind of have to follow

Same feeling. The magic isn't the speed. It's that a small team can finally try the ideas we used to drop for lack of time.

right? block everything or give it your full login was never a real choice