Hacker News: lawrenceyan

Fable 5.5: SVG of bike riding a pelican

lawrenceyan — Sat, 17 Oct 2026 20:19:44 +0000

Article URL: https://twitter.com/dummerspast39/status/2111519245733654775

Comments URL: https://news.ycombinator.com/item?id=50134741

Points: 5

# Comments: 0

Decoupled DiLoCo for Resilient Distributed Pre-Training

lawrenceyan — Wed, 30 Sep 2026 20:59:13 +0000

Article URL: https://arxiv.org/abs/2604.21428

Comments URL: https://news.ycombinator.com/item?id=49914359

Points: 1

# Comments: 0

Sharded matrices and how to multiply them (2025)

lawrenceyan — Thu, 24 Sep 2026 02:45:20 +0000

Article URL: https://jax-ml.github.io/scaling-book/sharding/

Comments URL: https://news.ycombinator.com/item?id=49825585

Points: 1

# Comments: 1

A self-evolving agentic system for automated execution of biological protocols

lawrenceyan — Tue, 22 Sep 2026 14:49:13 +0000

Article URL: https://arxiv.org/abs/2606.31763

Comments URL: https://news.ycombinator.com/item?id=49802262

Points: 1

# Comments: 0

Reviving Deserts with Syntropic Agroforestry

lawrenceyan — Mon, 21 Sep 2026 02:08:25 +0000

Article URL: https://wikifarmer.com/library/en/article/reviving-deserts-with-syntropic-agroforestry

Comments URL: https://news.ycombinator.com/item?id=49782264

Points: 3

# Comments: 0

New comment by lawrenceyan in "Gemini 3.8 Flash and 3.8 Flash Cyber"

lawrenceyan — Wed, 09 Sep 2026 23:40:46 +0000

Not sure if you'll see this comment since it's been a week, but how did you find this out? Is 3.5 Pro officially cancelled internally? Are they only working on 4 Pro now?

Why Diamond Transistors Are So Hard [video]

lawrenceyan — Thu, 27 Aug 2026 09:34:25 +0000

Article URL: https://www.youtube.com/watch?v=NLmd5vL0zmk

Comments URL: https://news.ycombinator.com/item?id=49462054

Points: 1

# Comments: 0

A coalescent model reveals deep ancestral structure shared by all modern humans

lawrenceyan — Sat, 15 Aug 2026 02:05:07 +0000

Article URL: https://www.nature.com/articles/s41588-025-02117-1

Comments URL: https://news.ycombinator.com/item?id=49306919

Points: 2

# Comments: 0

New comment by lawrenceyan in "We Are the Last People Who Know How It Works"

lawrenceyan — Wed, 01 Jul 2026 20:13:33 +0000

True

Test Like You Fly

lawrenceyan — Fri, 22 May 2026 17:20:14 +0000

Article URL: https://www.spacex.com/content/starship/test-like-you-fly

Comments URL: https://news.ycombinator.com/item?id=48238732

Points: 1

# Comments: 0

Gut Instinct: Connecting with our intuition

lawrenceyan — Tue, 12 May 2026 04:13:51 +0000

Article URL: https://www.cheltenhamzen.co.uk/writings/gut-instinct

Comments URL: https://news.ycombinator.com/item?id=48104118

Points: 1

# Comments: 0

New comment by lawrenceyan in "Why has there been so little progress on Alzheimer's disease?"

lawrenceyan — Sun, 26 Apr 2026 18:00:03 +0000

I'd argue the reason we’ve spun our wheels on Alzheimer's for decades isn't just about a flawed amyloid model. It’s a symptom of a much larger, dying paradigm in biotech.

For decades, the industry standard was to hunt for targeted small molecules to solve singular biological issues. And to be fair, I don't think this was because of a lack of vision as it was simply the strict limit of our technological capabilities at the time. But attempting to treat a cascading, systemic disease like Alzheimer's with a single targeted molecule is like hoping replacing one pipe will solve the problem when the issue is that the entire plumbing system is corroding.

This fundamental mismatch is exactly why clinical progress has stalled, and I believe the future of treating Alzheimer's will instead closely mirror how our approach to oncology has evolved.

We spent years searching for a universal molecule to cure cancer before we accepted reality. We now know that effective treatment often requires sequencing a tumor's mutanome to develop a highly personalized intervention for the individual. As neurodegenerative systems fail with age, we face that exact same biological complexity. A traditional small molecule is not going to rescue a globally failing network.

My personal take is that we won't see a true breakthrough in Alzheimer's until capital fully rotates out of these legacy, single-target pipelines and into programmable biological systems.

Lazard's Levelized Cost of Energy (2025) [pdf]

lawrenceyan — Fri, 24 Apr 2026 19:42:17 +0000

Article URL: https://www.lazard.com/media/5tlbhyla/lazards-lcoeplus-june-2025-_vf.pdf

Comments URL: https://news.ycombinator.com/item?id=47894892

Points: 3

# Comments: 0

Artemis, Greatness, and Risk

lawrenceyan — Fri, 10 Apr 2026 18:20:23 +0000

Article URL: https://1517.substack.com/p/1-in-30-artemis-greatness-and-risk

Comments URL: https://news.ycombinator.com/item?id=47721795

Points: 3

# Comments: 0

New comment by lawrenceyan in "Marc Andreessen is wrong about introspection"

lawrenceyan — Fri, 03 Apr 2026 18:40:10 +0000

What Andreessen is hinting at, albeit still largely surface level, is 无心 or no mind.

Popular in martial arts and Buddhist philosophy, I think practically what you should take away is that body and mind are fundamentally intertwined.

Introspection is a practice of the mind, specifically cognition centered around portions of the brain like the prefrontal cortex. There’s a lot more to who you are and areas you can hone / cultivate.

The HN crowd is probably overweighted on cognition, and could do with spending more time in other areas: https://www.cheltenhamzen.co.uk/writings/gut-instinct