ernsheong

Hmmm I tried it, but it didn't detect my open herdr sessions in the added machine. I don't see much benefit over herdr in Terminal/Ghostty, and I need to reconfigure my agents to make it work too.

This isn't a direct solution, but here is a HN mirror with dark mode support: https://hnmirror.com/

Use it if you like it ;)

I'm not sure that it needs to be this complicated. But yes, I am interested in the problem of hard guardrails between multiple agents, and it is not an easy problem to solve.

Lost me at a custom captcha.

Make them real sorry by not trying out their post-AI pivot.

Ah, the moral high ground of those who are supposedly not vibe-coding themselves... speak for yourself

I think if Cloudflare wants to have a fighting chance it should at least give its projects a different domain and a different brand. Nobody wants to pile in more failure modes into dash.cloudflare.com. This is probably the reason all their developer products are failing to take off.

This stuff is really complicated. Just trying to build a harness coordinating multiple instances of vanilla pi has been a bit of a nightmare. I'm not sure if the huge added complexity is worth it but kudos for trying and labelling as experimental.

So it seems this started because of DHH. Let's judge the technology by the technology, else we are just taking a holier-than-thou approach whereby we think we are better than others because we don't commit certain sins, which is a real problem because when we look inwardly and honestly we shall see that we are guilty of many of the same sins. Let's judge the technology for the technology itself and leave personal ideology out of it. My 2 cents (which will inevitably be downvoted)

I had a similar idea but was skeptical of developers' willingness to pay for a locked-in tool... all the best

The AI ship has already sailed. You won't lose your job to AI, but to another person using AI better than you. Do you really want to spend hours writing boring one-off shell and test scripts? Even a selective use of AI is better, or at least a personal policy of reviewing and understanding every line of code. Not some blanket no AI policy...

We have to manage it better, introduce guardrails, adversarial agents, better harnesses, etc.

All valid points but a personal boycott is not the solution.

What's more interesting is his "No AI" policy

https://dbushell.com/ai/

They should resurrect the Commonwealth into a stronger union of sorts instead of just playing sports every four years.

Mino 2.5 has been my workhorse for coder and tester agents (the ones planner agents delegate tasks to)

Hey but that's exactly as per design. It is the customer's responsibility to store stuff elsewhere as DR backup, not AWS.

This is potentially huge and can crash the Big Two's stock prices or block their IPOs completely.

Incoming business slop

sounds like you might need to beef up your harness first, and run multi-agent verification loops

Lol the page doesn't even scroll

There's still 3.1 Pro though, as ancient as it sounds now

Flash is the new Pro, try it first

I built my harness in pi within herdr, I cloned (zipped and downloaded) 0xRichardH/pi-herdr-subagents and went from there, and used pi to build itself, adding gate workflow state control, provider fallbacks (I use many token plans), subagent skill injection, etc.

It is highly custom to my needs and wants, and I think every developer needs to do this. I only talk to my planner which plans, and it subs out to designer, oracle, coders, testers, and reviewers, etc. It is thus highly optimized for correctness. You can TDD or no TDD. You can fast track small changes. I tweak my harness dozens of times as I encounter new edge cases (esp when I switch models and encounter models not as good at following instructions).

As you can start to see, it is better to own the harness because nobody can build something custom that 100% fits your needs or development philosophy.

Yeah but engineering is still subject to failure modes, AI can't circumvent that short of formal proofing (which nobody really wants to get into)

It seems like Shopify has fallen into a trap thinking that more complexity costs nothing because of AI. In the age of AI we have to embrace the same principle as before that complexity needs to be tamed, not multiplied. Even as humans struggled with complexity, from what I see now AI struggles very much the same. Hence I don't think this will age well.

Wow, citing we need to figure out how to pay out bills and rejecting PR based on that is such a douche move. The fact is Tailwind core is an MIT-licensed open source project.

Toyota was right all along.

"We couldn't continue with that account. Try again."

Is it geo-restricted to the US for now?

How are you handling subagents? Subtabs?

Hey there, nice one. Care to share a bit what is under the hood? What are the challenges of local AI? It is RAG-ging? What small models do you use?