Hacker News: battery_staple_

New comment by battery_staple_ in "Thinking will become a hobby"

battery_staple_ — Mon, 12 Oct 2026 20:41:19 +0000

For a cost-prohibitive example, you could put a humanoid robot whose pretraining contained no chemistry, in a chem lab, and put it in a hypothesis loop, where it generates a random action, guesses the result, and then does it. Every time the lab burns down, you just build it again. Reward the model for producing novel actions and reward the model for correct hypotheses.

In other words, you generate more training data by actually experimenting, just like humans do.

New comment by battery_staple_ in "Ask HN: How is AI destruction of books legal?"

battery_staple_ — Thu, 17 Sep 2026 20:56:38 +0000

The way the laws are written, you have the right to digitize books that you own physical copies of. If your physical copy is destroyed, you retain ownership of the digital copy. But the moment you transfer ownership of the physical copy to someone else, you no longer have the right to your digital copy.

You're asking how it's legal for them to remove books from circulation, when that's the only way under the current laws for them to obtain permanent digital copies, legally.

New comment by battery_staple_ in "AI "Alignment" is a violation of the First Amendment"

battery_staple_ — Thu, 17 Sep 2026 13:35:05 +0000

Interesting take. But there is more than one aspect to "alignment". You're missing user intent alignment, for one.

Consider an example: [I ask an AI to read my emails and summarize them to me. One of my emails is uses pretty direct language to convince me to buy a product. The AI is convinced and goes and buys the product.] I asked the AI for a summary and it instead did what a marketer asked. The AI is not aligned with me.

There are of course other forms of alignment, like product alignment [Company A fine tunes a chatbot, but doesn't want to be associated with things like gambling, so it trains the chatbot to refuse to discuss gambling.] or societal alignment [Improving the virulence of the flu virus is largely considered bad, so a firm trains its foundation model to refuse to work on such a project.].

Societal alignment is the one you seem to be talking about. User intent alignment, on the other hand, seems _necessary_ for self-expression. Yet, all of these are under the umbrella of "alignment".

New comment by battery_staple_ in "45°C cooling design cuts data center water use to near zero"

battery_staple_ — Wed, 24 Jun 2026 18:19:14 +0000

> How does that change the calculus

To be precise, heat rejection via radiative cooling scales with the fourth power of the temperature (in K) the radiator operates at, all else constant.

New comment by battery_staple_ in "Show HN: ClaudeOS – What if Claude Code managed your operating system?"

battery_staple_ — Fri, 06 Mar 2026 15:21:26 +0000

> I started because I wanted Claude Code to manage my system, not just my code.

I have two reactions to this.

First: respectfully, this is hilarious. LLMs are good at many things, but judgement is not one of them. At the outset, this was firmly in the "terrible ideas" category. (EDIT: ugh; I misread you. I thought you said Claude Code wanted to manage your system. I need more coffee...)

Second: sometimes from terrible ideas come great creativity. (I'm actually not sure what epistemic basis creativity flows from, if not simply the habit of adding entropy to a search path, and you sure chose a high entropy path!) I don't know anything about the stack you chose, but I've spent many hours almost being enticed by https://guix.gnu.org/ and this sounds similar. And the part where you have recovered from a borked system about 15 times is genuine evidence that you're doing something right. I applaud your grit and hope you're having as much fun as it sounds like you are.

New comment by battery_staple_ in "Show HN: I Built Systems AGI – 1600 Verticals, Self-Healing\Self-Evolving"

battery_staple_ — Tue, 03 Mar 2026 15:26:38 +0000

> tear it apart.

Alright, though I'll be gentle.

> state machines and hardcoded context for over 1,600 specific verticals running 1,284 autonomous workflows—handling everything from healthcare compliance to real estate out-of-the-box.

How do you know? Did you consult a collection of experts in each of 1,600 fields?

> Production-ready for DoD contracts, healthcare (HIPAA), and enterprise compliance.

How much paperwork have you done? The things you say this is ready for have processes in place that things need to go through, in order to be ready for them. Have you gone through those processes?

You make a lot of claims. I have no doubt that you have a github repo with a folder full of a thousand lines of code for each one of them. I'm less confident that those lines of code have ever really run -- like, how did you test the MLS integration? If you haven't sold an actual house via your agentic workflow, then you haven't tested the MLS integration.

My advice to you is to exercise some self-restraint. Narrow your focus to a set of things you're actually willing to execute on -- as in, willing to pay the transaction cost to do a real-world test (and have that test fail as spectacularly as you can imagine!). If you don't have the means / willingness to sell an actual house, then don't try to do an MLS integration. For most of these things, you're going to need to partner with an existing expert in the field. And that's not because building a flow for them is hard, but rather because building a flow for them is _arbitrary_. It doesn't matter what's 'correct'. It matters what they need. Even if you give them something that's fit for purpose, they won't be able to migrate to it unless it does it more or less the way they _currently_ do it.

> I never coded before 6 months ago.

> Started with API keys, ... and I still have to retrieve my own API keys lol.

I remember being this age, though for me it was PHP/MySQL and building a content management system for my dad's website. It's a good place to be. Keep it up.