HN user

JnBrymn

594 karma
Posts83
Comments15
View on HN
www.youtube.com 2mo ago

The Dark Factory: How Twin Sun Automated Their Dev Pipeline [video]

JnBrymn
2pts0
deepwiki.com 3mo ago

Sachin1801/Claude-Code – DeepWiki

JnBrymn
3pts0
muellerberndt.medium.com 3mo ago

Observers Are All You Need: How Observer-Synchronization Creates All of Physics

JnBrymn
2pts0
blog.southparkcommons.com 3mo ago

Startups building Workflow products for other Software Startups – Beware

JnBrymn
16pts0
medium.com 4mo ago

Stop Building AI "Teams." Start Building Software Factories. – Owen Zanzal

JnBrymn
5pts0
twitter.com 4mo ago

Harrison Chase: How Coding Agents Are Reshaping Engineering, Product and Design

JnBrymn
1pts0
twitter.com 4mo ago

The rise of WORKIGN AI research agents: Andrej Karpathy

JnBrymn
4pts0
www.youtube.com 4mo ago

Claude Code vs. Codex (Nate B Jones) [video]

JnBrymn
1pts0
twitter.com 4mo ago

How do AI-forward teams review giant vibe-coded PRs – line by line?

JnBrymn
1pts0
alignment.anthropic.com 4mo ago

The Persona Selection Model: Why AI Assistants Might Behave Like Humans

JnBrymn
2pts0
twitter.com 5mo ago

Ivan Zhao on X: "On Universe, Life, and AI " / X

JnBrymn
1pts1
models.dev 5mo ago

Models.dev – An open-source database of AI models

JnBrymn
24pts1
arcturus-labs.com 5mo ago

Why Spec-Driven Development Breaks at Scale (and How to Fix It) – Arcturus Labs

JnBrymn
2pts0
arcturus-labs.com 6mo ago

Incremental AI Adoption for E-Commerce – Arcturus Labs

JnBrymn
1pts0
softwaredoug.com 6mo ago

Ralph, too, needs a test train split

JnBrymn
3pts0
pattersonconsultingtn.com 6mo ago

What is MLflow? (A good primer.)

JnBrymn
2pts0
arxiv.org 6mo ago

Everything Is Context: Agentic File System Abstraction for Context Engineering

JnBrymn
1pts0
arxiv.org 6mo ago

Language Models as One-Time Teacher for Hierarchical Planning in Text Environs

JnBrymn
2pts0
softwaredoug.com 7mo ago

RAG Isn't a Vector Search Problem

JnBrymn
4pts0
arcturus-labs.com 7mo ago

Programming in English – A Cursor command to write personalized tutorials

JnBrymn
1pts0
poetiq.ai 7mo ago

Poetiq – Traversing the Frontier of Superintelligence

JnBrymn
2pts1
www.cnbc.com 7mo ago

Anthropic reportedly preparing for IPO in race with OpenAI: FT

JnBrymn
6pts1
arcturus-labs.com 7mo ago

Why Spec-Driven Development Breaks at Scale (and How to Fix It) – Arcturus Labs

JnBrymn
1pts0
openai.com 8mo ago

Understanding neural networks through sparse circuits – OpenAI

JnBrymn
1pts0
cookbook.openai.com 8mo ago

GPT-5.1 Prompting Guide

JnBrymn
4pts0
arcturus-labs.com 8mo ago

Context Engineering Requires AI Empathy – Arcturus Labs

JnBrymn
1pts0
ainativedev.io 8mo ago

Spec-Driven Development: things you need to know about specs – AI Native Dev

JnBrymn
2pts0
arxiv.org 8mo ago

Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity

JnBrymn
1pts0
arxiv.org 8mo ago

Can We Fix Social Media? Testing Prosocial Interventions Using Social Simulation

JnBrymn
2pts0
arcturus-labs.com 9mo ago

Fire Yourself First: The E-Myth Approach to Iteratively AI App Development

JnBrymn
2pts0

It was fun writing our book because I SAW you do that. And I had a different approach - I would outline obsessively and hold the whole chapter in my head at once before I started writing. Holding a whole chapter and cross referencing everything with everything else was O(N^2). You're approach for writing one instance of the chapter was linear O(N) but you did it M times... so O(M*N) ... maybe about the same :P

You're absolutely right. This was a post I tossed together quickly just to see what could be done without thinking too much. In retrospect, I think this would be better implemented using Elasticsearch sparse vector fields which allow you to specify the value of every token. Maybe I'l make an update post to try again.

[dead] 1 year ago

Practical advice! So many good products are lost by people that become fixated on unnecessary evals too early. You need to build your eval muscle AS you release product and get real feedback.

RAG is a pain to set up, so I tried something different. Instead of dealing with vector DBs and all that complexity, just let the LLM navigate well-structured docs like a human—exploring outlines and diving into sections. It’s simple, and works great for stuff like technical manuals or llms.txt.

What is AGI? 2 years ago

Feels like it's a dopaminergic response to hearing a word but not knowing what it is. It's a novelty seeking thing. But once the work seems to be well understood, the novelty wears off and the novelty seeking mechanisms in humans quit responding.

If you get tinnitus, spend the first couple of months looking for a solution. Get opinions and second opinions. If that doesn't work, then wait for 6 months and then try again. If that doesn't work then wait for 9 more months then try again. If that doesn't work wait for 12 more months and then try again.

Most likely you'll habituate to the sound (as I have). What was once a living nightmare will become (miraculously) a non-issue. Occasionally, as one person said it even becomes "comfy" - or as I told my wife recently, it's just the sound that the world makes when you're alive. Really, it's not much different from other signal processing affects in other senses. If you close your eyes and concentrate you can see visual noise. It's also there when you open your eyes (floaters for example). But you don't worry about that because it's commonplace. Sound seems different because it's new - but it will become the same type of thing.

An important part of getting better is accepting that (likely...) you're not going to get better. And then, after just being with it for a while and not trying to push it away, you realize that it's not so bad - you get better. Just let life suck for a while. My experience hasn't been pleasant, but I've learned a lot about the nature of life. Fair trade. Wouldn't want to do it again, but glad to have done it.

I once presented with softwaredoug before a room of 400 people on a project that we didn't quite get finished! Yes... massive amounts of stress as we approached the deadline and realized that we were not going to hit it. Nevertheless, b/c of the stress we pushed harder than we normally would have, we learned a lot more than we otherwise would have, and _despite_ not finishing, the talk was one of the better received of my career. (We turned it into a lessons learned talk. We had plenty of great questions.)

How did we deal with the stress? I think we just sat with it. It was always there. Over time you recognize it for what it is, an illusion. That doesn't make it feel much better, but it gives you a bit more control and equanimity. And being able to push through hard situations despite the illusory feelings of dread opens the doors for doing some really interesting things.

In Nashville (and growing elsewhere) we started a community around self-organizing peer-mentorship. Checkout it out http://pennyuniversity.org

It's very rudimentary and grass roots at this point. (I mean.. we're a Slack Team and a Google Forum.) But the idea is that everyone has things they want to learn and thing that they can share. PennyUniversity serves as the community for people that want to make In-Real-Life connections to learn new things. It's self organized, mostly it's individuals asking to learn something and then setting up a coffee chat or a lunch conversation. Occasionally the conversations grow to larger groups (say 10 people). Often the conversations are one-off, but sometimes they form into longer-term mentorship or topical-groups or reading groups. We encourage face-to-face meetings, but occasionally the meetings are online and recorded - like this: https://www.youtube.com/watch?v=Czy7bgDd7Hc

Check it out. I'd love to hear feedback.