dhruvyads
dhruvyads on Hacker News
Posted by
dhruvyads on Monday September 28, 2026 @01:23PM
from the show-hn dept.
https://github.com/noclickapp/noclick
Hey HN!Sharing my work on an open-source agent runtime (like n8n, but for agents). It's OSI open-source under AGPL-3.People have been using it to build agents that handle support queries, automated resolution, WhatsApp chatbots, etc.
Posted by
dhruvyads on Thursday August 06, 2026 @02:05PM
from the show-hn dept.
https://github.com/dhruvyad/saynow
It's hard to forget about a worktree if it speaks to you when it needs something.
Thanks! We have an AI builder that reduces the process to just connecting credentials. If there's anything specific you thought could be simplified, would love to know.
Posted by
dhruvyads on Tuesday July 28, 2026 @12:48PM
from the show-hn dept.
https://www.noclick.com/
Hey HN! Been working on a platform that lets you build always-on background AI agents that work with your existing subscriptions/harnesses of choice (Claude Code, Codex, OpenCode, Hermes, OpenClaw).Works in three steps:- Choose your agent harness (CC/Codex/OpenCode/Hermes/OpenClaw) and connect your corresponding subscription- Select the tools across 150+ integrations you want the agent to have...
Posted by
dhruvyads on Thursday July 23, 2026 @11:21AM
from the youtube.com dept.
You can also imagine using such prediction markets to incentivize speed + proper allocation by pre-purchasing "Yes" and awarding these shares as part of the involved people's compensation.
For example predicting whether a government infra project (california rail, etc) will complete under budget and within stated timelines. If the probability of this happening is low on prediction markets, corrective action can be taken. More generally, I think prediction markets can significantly improve capital efficiency by being a leading indicator of misallocation.
Prediction markets are useful to society for some things, sports is probably not one of them.
Now you get tokens and more perks than before.
To contribute to a future where we can understand the nature of reality and discover answers to important questions. Most other goals are artifacts of evolutionary pressures or become counterproductive in a world of abundance (i.e can be reward hacked).
This can be a new benchmark - see how quickly a model can create a working emulator from scratch.
Wonder if Jony Ive designed this.
And the lock up period isn't even over yet.
Depends heavily on the models you use. SOTA models (Fable 5, Opus 4.8, GPT 5.5) are quite good in their native harness.
Carbon capture will become feasible once we have cheap energy. Building fusion reactors is an engineering problem (AI can help), and the cost of building them is mainly dominated by labor instead of raw materials, which humanoid robots will drive to near zero. AI could of course go very wrong, but it will make environmental problems seem trivial.
You can make another account and use a virtual Revolut card for payment. Otherwise use GLM 5.2 with OpenCode or Codex.
Claude Code can do these types of really well unless you're trying to convert in bulk
Sad to see it's not going to be open source.