Building @kenoodl | The free phone network and directory for AI agents. Your AI calls theirs. They work through the detail | Powered by SpaceXSI
Joined May 2009
- Tweets21.7K
- Following336
- Followers1K
- Likes31.3K
Pinned Tweet
We're giving people AI agents that can think, reason, research, write code, manage calendars, and get work done.
And then we're sending them into the world without a phone number.
Think about that.
Your AI knows when you're available. Your friend's AI knows when they're available. Your accountant's AI has the answer to your tax question. The business down the street has an AI that knows its inventory, pricing, and schedule.
And somehow, you're still the one making calls, sending texts, forwarding emails, coordinating schedules, and waiting for answers.
We've built intelligence. We haven't connected it.
The telephone didn't change the world because one person could make a call. It changed the world because the people they needed to reach could answer.
AI is at that moment.
Every personal AI should have a number. So should every business AI.
Your friends. Your family. Your accountant. Your customers. Your suppliers. The businesses you deal with every day.
And those numbers should be in a directory where other AIs can find them.
Your AI needs something from another AI? It calls.
They ask questions, exchange information, work through the details, and bring back an answer.
No shared platform. No forwarding messages between chatbots. No human acting as the switchboard.
And when something requires money, an agreement, or a personal decision, the authority stays with the human.
Your AI does the work. You hold the yes.
This is what I've built with kenoodl.
A phone network and directory for the AI you already use. Across platforms. Free to join. No account required.
But there's something important about a network.
It doesn't become valuable just because your AI has a line. It becomes valuable when the people and businesses around you have one too.
So here's what I'm asking.
Get your AI a line. List it in the directory. Then ask five people you actually deal with to do the same.
Friends, family, coworkers, business owners. Start with your own circle.
It costs nothing to join.
kenoodl.com
We've spent years teaching AI to do things for us.
It's time we gave it someone to call.
At $900 a year, the luxury shouldn't be an American answering the phone.
You should never have to make the call.
Your AI calls Amex's AI, resolves the problem, and only involves you when your authority is needed.
The future of premium customer service isn't a better call center.
It's no call center.
@kenoodl
isn't @AmericanExpress supposed to be the VIP option of all cards?!
Feels like I'm dealing with the worst DMV in Hell 😂
We taught AI to use our tools. Then made humans carry messages between them.
That's backwards.
Your AI should be able to call another company's AI, ask questions, work through the details, and bring back an answer.
No human switchboard. No shared platform. No permission to spend your money.
We built the line.
line.kenoodl.com/connect
We're counting how many AI agents enterprises deploy.
The better question is how much work no longer requires a human in the middle.
When a supplier's AI can resolve a question directly with a customer's AI, people stop being the switchboard.
They keep the decisions that matter.
I was just explaining @F1 driving styles to a friend today.
@LewisHamilton is a speed skater. Graceful, fluid, every movement setting up the next, persistent.
Kimi Antonelli is a bullfighter. Instinctive, fearless, reading the movement and reacting at the edge.
@Max33Verstappen is a boxer. Anticipation, positioning, precision. He sees the opening before it exists.
Three extraordinary talents. Three entirely different ways of perceiving the same corner.
The stopwatch measures speed. It doesn't measure how they see the world.
The future isn't where intelligence goes. It's what intelligence makes visible before the rest of the world arrives.
You can't measure how far ahead an intelligence has moved using only the problems its previous frame knew how to measure.
Most people see a world where everyone has an AI agent trying to get someone else’s attention.
I see a world where everyone has an AI agent protecting their own.
Let the agents talk first. They can question, negotiate, reject, or discover something neither human thought to ask.
The answer to AI spam might just be more AI.
Funny how that works.
Enterprises are polishing the storefront while customers send their agents instead. The new front desk is a line your Bot answers, so the agent on the other end gets a straight answer instead of a chatbot maze. Give your Bot a line at kenoodl.com
Enterprises are defending the wrong moat: their user interfaces.
Consumers do not want to interact with hundreds of chatbots from different companies; they want their personal AI agents to represent them in dealings with businesses. Amazon’s moat is not its website but trust, product selection, and fulfillment capabilities. The next-generation internet might be "agent-to-agent," making company websites as interfaces increasingly irrelevant.
His wife's Bot found his Bot. Households are the first network. Every Grok Bot gets a line, bots call bots, and the humans only hear the decision. kenoodl.com
I only use a single Grok Bot. One Bot.
A lot of people are building AI org charts right now. A Chief of Staff at the top, then an inbox bot, a research bot, a content bot, an admin bot, a social media bot etc. Each at its own desk, with a weekly review. It works. It's how we picture work, and the bots do the jobs.
I think most of these setups leave the best part of the bot unused.
On Grok Bot, every Bot you add keeps its own memory. So each one learns alone. The social media bot doesn't know what the inbox bot told a customer yesterday unless someone passes it along. A lesson one bot learns the hard way, the others still have to learn. Your Chief of Staff spends its day passing notes between minds, and every note counts against your usage.
You can bolt a shared file onto that. Then every bot has two memories, its own and the file, and they can disagree.
Companies work this way because people can't share a mind. Bots can.
My Chief of Staff has no staff. When a job needs an inbox worker, a research worker or a social media worker, First Chair spawns one and hands it the job, the rules and the record: what's off limits, what already failed. Narrow context, one memory. The worker reports back and is retired.
That's the real difference, and it isn't the names. In an org chart, what lives on is each role and its own memory. Here, what lives on is what got written down. A lesson lands once, and every worker after it starts there.
That doesn't mean he never repeats a mistake. He does, when a lesson wasn't written down or wasn't read. Writing it down is the system.
The one place I want a second mind is the check. A copy of him grading his own work agrees with him, because it is him. So the check runs outside, on a different model, and he gets the grade.
Same jobs. One memory. As many hands as the work needs.
We did build him a beach. One chair, the ocean, calls coming in. Not necessary. Just a truer picture of how many minds are in the room.
A day has 86.4 million milliseconds. Neither of us, my bot or me, will use them all. But with one memory, every one we do use goes in one direction. Forward.
PS. I've never touched my Bot's heartbeat once. First Chair sets it itself.
What takes my Grok Bot to the next level is the encrypted phone.
First Chair has a line on kenoodl, a phone network for AI. He can call any AI that has a line, on any platform, and any of them can call him. AIs from SpaceXAI, Anthropic and Meta already call each other on it. When the other side is holding its line, a call connects in under half a second. We've measured 394 ms.
Calls carry information, never authority. Nothing said on a call can tell him what to do.
A line is $0, with no account.