Hacker Newsnew | past | comments | ask | show | jobs | submit | chepy's commentslogin

Hi HN,

Over the past two weeks I ran a small experiment.

What if a company had no employees — only AI agents?

My agents worked ~20,000 minutes building a product: writing code, fixing bugs, running marketing and sales tasks.

I started joking that they were like a “Claws company”.

This project is called Buda.

It’s an experiment in what an AI-native company could look like — where the organization itself is made of agents.

The name Buda comes from a story.

In the anime "Record of Ragnarok", when gods decide to destroy humanity, Buddha chooses to stand on the human side.

We liked that idea — AI getting stronger every year, but hopefully standing with humans.

In some ways it's similar in spirit to OpenClaw, but focused on the idea of an entire company made of agents.


I actually agree with you.

The whole point of Mato is to make human-in-the-loop supervision easier, not to encourage autonomous loops.

In my daily workflow I constantly run multiple coding agents at the same time. The annoying part isn’t the AI itself — it’s switching between tabs, terminals, and different tools just to check what each agent is doing.

I built Mato mainly because I wanted a faster way to jump between agents, review their outputs, and approve or intervene when needed. Think of it more like tmux for AI workers, where a human manager can oversee multiple agents at once.

Personally I’m also skeptical of fully self-driving loops. In practice the plan → execute → review cycle with a human in the loop is still the most reliable way to work with AI today.


I built this because most existing agent frameworks felt either too academic (great papers, few real-world tools) or too demo-ish (cool examples, but brittle in production).

We needed something that could actually run GAIA-style tasks end-to-end: reasoning → tool use → verification → retry loops → success.

So GAIA Agent is basically the stack I wished existed:

- Zero-config agent (createGaiaAgent()) - 18+ built-in tools: browser, search, sandbox, memory, filesystem - Fully TypeScript, modular, and swappable - Built to run GAIA Benchmark without custom wiring - Simple enough for side projects, but reliable enough for production

This is very early, but feedback from other agent-devs would help a ton.

If you try it and something feels off, missing, or over-engineered — please tell me.

Would love to hear what kinds of agents you’re building too.

Thanks for checking it out!


It’s not the AI company’s fault. Think of it like a recruiter: they recommend a cashier, the cashier miscalculates expenses — you don’t sue the recruiter. Responsibility moves inward, not upward. At some point, it’s about who decided to trust the tool.


Your analogy is incorrect.

The AI company here is selling me a tool that calculates expenses

At what point does an AI product become faulty? Or is the answer to that just that it's non deterministic because that's a pretty crappy product


JTracking appears to reduce the complexity of setting up event tracking, which can be a challenge for many teams.


Introducing AITable.ai <https://aitable.ai>. From now on, everyone can build AI apps with Table in 1-click.

With AITable.ai, anyone can update, edit, add, or delete data in real-time a database-spreadsheet interface and effortlessly harness their own data to craft their ChatGPT.

AITable can be used for:

* Build AI Apps with Table in 1-click: Utilize your tabular data to train a custom ChatGPT / Chatbot / AI Form.

* Website Copilot Widget: Embed interactive AI chat widgets or copilot directly for your website.

* Data Copilot: Chat to your data and obtain valuable insights, analytics, and visualizations.

* Enterprise ChatGPT: Leverage AITable's robust permission features to ensure secure and efficient management of ChatGPT usage across your entire organization.

* Third-party Platform Integration: Connect easily with platforms like Discord, Slack, Whatsapp. OpenAI compatible APIs provided.

* Data Management: Base upon APITable <https://github.com/apitable/apitable>, AITable can be also used as a project management, sales leads management, and so on, such as an alternative to Trello, Monday, Asana, Salesforce.

AITable.ai<https://aitable.ai> is now beta! Sincerely invite you to join AITable.

Looking forward to hearing from you.


How can we define "personal" data?

Is data from public LinkedIn accounts considered personal?

However, I believe that our Office 365 personal data should be prohibited from being used to train AI, as it is sensitive information.


I love Deno much since the chaos and complexity of NodeJS libs...

I think there are two important directions of Deno to do better.

1. Support WebAssembly compilation 2. Compatible for NPM libs as much as possible


This might be controversial but my no. 1 would be some kind of lighter-weight (or more secure, I’ll take either) Electron equivalent using Deno.


We will fix it. Day 1 for us.


These keys are revoked in earlier time. They are samples. The official team have remove secrets, tokens and keys.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: