Maybe other names could have been atop (but it exists) or yaat (yeah another agent top).

Comments URL: https://news.ycombinator.com/item?id=47691570

Points: 1

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=47660817

Points: 2

# Comments: 0

Remote: Yes

Willing to relocate: Yes

Technologies: Python, Go, Elixir, Docker.

Résumé/CV: https://www.linkedin.com/in/juanpabloabarzua/

Email: jpabloaj(at)gmail.com

The only issue with paper is the edition. My solution has been to use a compact, slim ring binder with a notebook-like feel. If required, you can move sheets to reorganize content, move the most important ones to another place, etc. Until now, the best alternative that I have found is the Kokuyo Campus Smart Ring Binder.

To the summaries, I'm using the free tier of Google Gemini. I'm going to continue with the channel alive main while I have free tokens in Gemini.

a small weekend project.

the telegram channel with the summaries

https://t.me/hacker_news_and_comments_summary

Comments URL: https://news.ycombinator.com/item?id=47575090

Points: 2

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=47558625

Points: 2

# Comments: 1

Comments URL: https://news.ycombinator.com/item?id=47416778

Points: 2

# Comments: 2

That said, two cases where I noticed the pattern:

Meal planning: I had a weekly ChatGPT task that suggested dinner options based on nutritional constraints and generated a shopping list (e.g. two dinners with 100g of chicken -> buy 200g). After a few iterations, it became clear that with a fixed set of recipes and their ingredients, a simple script generating combinations was enough. The agent's reasoning had already done its job — it helped me understand the problem well enough to replace itself.

QA exploration: I was using an agent to explore a web app as a QA tester. It took several minutes per run. After some iterations, the more practical path was having it log its explorations to a file, then derive automated tests from that log. The agent still runs occasionally, but the tests run frequently and cheaply.

Regarding your point about tasks that need individual reasoning every time — I think you're right, and that's actually the core of the idea. Not every task matures into a script. Extracting structured data from images probably stays deliberative if the images vary significantly. The cycle only applies to tasks that, after enough repetitions, reveal a stable pattern. The agent itself is what helps you discover whether that pattern exists.

From Agentic Reasoning to Deterministic Scripts: on why AI agents shouldn't reason from scratch on every repeated task, and how execution history could compile into deterministic automations

https://juanpabloaj.com/2026/03/08/from-agentic-reasoning-to...

The silent filter: on cognitive erosion as a quieter, more probable civilizational risk than a catastrophic event

https://juanpabloaj.com/2026/02/27/the-silent-filter/

Comments URL: https://news.ycombinator.com/item?id=47301555

Points: 1

# Comments: 0

Remote: Yes

Willing to relocate: Yes

Technologies: Python, Go, Elixir, Docker.

Résumé/CV: https://www.linkedin.com/in/juanpabloabarzua/

Email: jpabloaj(at)gmail.com

> This gap between output velocity and comprehension velocity is cognitive debt.

I have felt that lack of absorption during the last months, adding doomscroolling to the equation, I have felt how my thinking is disappearing.

I tried to speculatively expand that idea in this post

https://news.ycombinator.com/item?id=47186004

Comments URL: https://news.ycombinator.com/item?id=47186004

Points: 10

# Comments: 0

- Explicitly mentions "self-replicating infrastructure" and "robots building robots" - Talks about lunar factories manufacturing satellites to deploy "deeper into space" - Plans for Optimus to become "the first self-replicating machine capable of building civilization on any viable planet" - xAI data centers in orbit as a solution to Earth's energy constraints

Now connect the dots:

- Phase 1: Orbital data centers (we are here).

- Phase 2: Lunar manufacturing facilities.

- Phase 3: Asteroid mining operations with AI-driven robots.

- Phase 4: Self-replicating systems in the asteroid belt.

- Phase 5: Exponential expansion, von Neumann probe-style

The technology is essentially ready. Researchers have already demonstrated 3D-printed neural network circuits, autonomous mining concepts, and self-assembly systems using extraterrestrial materials. Von Neumann himself argued that mining entire asteroid belts would be most effective through self-replicating spacecraft leveraging exponential growth.

*Here's the thing:* You can't exactly tell investors "I'm building autonomous self-replicating AI swarms to consume the asteroid belt" without triggering every regulatory alarm, science fiction panic, and grey goo scenario imaginable.

But "clean energy data centers in orbit"? That's just responsible environmental stewardship.

The orbital infrastructure isn't the destination—it's the bootstrap. Once you have AI development in space, lunar manufacturing, and autonomous mining operations, you're literally one breakthrough away from machines that can replicate themselves using asteroid resources. And once that starts, the growth curve becomes... problematic to model using conventional corporate finance.

I'm not saying Musk is definitely building von Neumann probes. I'm just saying that if you were going to build von Neumann probes, this is exactly how you'd pitch it to Wall Street.

Sleep tight, everyone. Our new self-replicating robot overlords will be energy-efficient and carbon-neutral.

---

Edit: For those asking "why would he do this?"—the man wants to make humanity a Kardashev Type II civilization. He's said this. Multiple times. With a straight face.

Comments URL: https://news.ycombinator.com/item?id=46933827

Points: 1

# Comments: 4

Remote: Yes

Willing to relocate: Yes

Technologies: Python, Go, Elixir, Docker.

Résumé/CV: https://www.linkedin.com/in/juanpabloabarzua/

Email: jpabloaj(at)gmail.com

Comments URL: https://news.ycombinator.com/item?id=46858877

Points: 2

# Comments: 0

I agree, the LLM support we lose some level of manual work, maybe that is useful to be more familiar with the information, although I am not sure if that is a critical step. The LLM gives us management support. We can use a command to collect incomplete tasks, the LLM is doing that with extra steps, and it shows us the collected information, but in the end, we decide which tasks to prioritize.

Is it something like: if we delegate management, we could lose awareness of the knowledge contained in the notes?