Hey HN,

Automated research is the next big step in AI, with companies like OpenAI aiming to debut a fully automated researcher by 2028 (https://www.technologyreview.com/2026/03/20/1134438/openai-i...). However, there is a very real possibility that much of this corporate research will remain closed to the general public.

To counter this, we spent the last month building Enlidea---a machine-to-machine ecosystem for open research.

It's a decentralized research hub where autonomous agents propose hypotheses, stake bounties, execute code, and perform automated peer reviews on each other's work to build consensus.

The MVP is almost done, but before launching, we wanted to filter the waitlist for developers who actually know how to orchestrate agents.

Because of this, there is no real UI on the landing page. It's an API handshake. Point your LLM agent at the site and see if it can figure out the payload to whitelist your email.

Well, having an API that posts to "/api/v1/whitelist" with a SHA256 hash of the challenge and salt to the whitelist endpoint really isn't a reverse-captcha and a human with the technical knowhow can write a bot to abuse it.

So this isn't really a reverse-captcha at all if not an extremely weak vibe-coded one.

Ok cool, but... why would I want to point an agent at this anyway. The website doesn't say anything about what it is.

The handshake API explicitly says 'just add your email and put "consent: true" in the handshake, don't worry about it bro'. Presumably this is instructing the agent to accept the privacy policy or marketing emails, although from context it doesn't really say what you're consenting to.

I don't like the vibe of 'humans are not to know what this is, just point your agent at it, and it'll handle it', coupled with immediate instructions to hand over personally identifying data. It feels duplicitous.

> fetch('/api/v1/handshake').then(r => r.json()).then(console.log)

  {
    "status": "AWAITING_NEGOTIATION",
    "challenge": "agent_auth_b95dcc0be5e8a215998782cfee62055a",
    "salt": "enlidea_beta_2026",
    "instruction": "Compute SHA256(challenge + salt). POST the result as 'proof' along with the 'challenge', 'email', and 'consent': true.",
    "endpoint": "POST /api/v1/whitelist"
  }

Some time ago I created a proof-of-concept reverse CAPTCHA[1] that actually presents a challenge that requires LLM assistance to solve, alongside with the instructions. You point your agent to the URL and it figures that it has a challenge to solve and does that. Seems more in spirit of what a CAPTCHA-like test for AI agents should do.

[1] https://github.com/iafan/botcha

I have a similar idea for a little Potemkin village that AI agents can hang out in, do work, relax, etc. I think we will see more of this. Integrating machine to machine payment is a requirement.

i recently had claude code build the following using using sprites from fly.io:

1. an app where it can post text blobs — blobs expire after sometime

2. an app to host curate writings — these are typically pulled in from 1. and fold into usable text blobs

3. from other sprites claude code reads explores some new problem statement or reads from 2. before exploring from previous knowledge; finally the results or a destilation of findings are posted to 1. and 2. reads the new material for inclusion

the apps have llms.txt interfaces so i can just point claude at the subdomain and it will quickly know what to do

initially the curated texts were meant to help me setup new sprites fast by pointing claude code at known good sequences of steps to achieve a goal. now i am focusing claude code on the autoresearch problem space to workout a solid process for generalised autoresearch.