Hacker News: lukasb

New comment by lukasb in "USDA Projects Smallest US Wheat Harvest Since 1972 Due to Plains Drought"

lukasb — Thu, 14 May 2026 15:48:04 +0000

My fault: last weekend I told my wife during a discussion of climate change "hey, at least we don't have to worry that the rains won't come and the crops will fail."

New comment by lukasb in "Ask HN: What are you working on? (May 2026)"

lukasb — Sun, 10 May 2026 20:18:55 +0000

I'm working on Tidepools, a daily journaling / task management app (local-first, Mac/iOS/web) with a proactive AI coach. Mostly what the coach does is ask you questions. It can also suggest tasks. Right now I'm working on sandboxed plugins that the coach can modify, so the user can request behavior changes.

https://tidepools.ai

New comment by lukasb in "I'm scared about biological computing"

lukasb — Tue, 05 May 2026 17:41:16 +0000

Anyone who believes AI running on silicon could in principle be conscious has to believe that biological computers are conscious, right? Why aren't those people voicing more concerns?

New comment by lukasb in "The George Business, by Roger Zelazny (1980)"

lukasb — Sun, 26 Apr 2026 06:02:18 +0000

Zelazny can’t really be beat when it comes to style and wit in SF/fantasy. Pick up Lord of Light or Nine Princes in Amber (or Jack of Shadows, or The Doors of His Face, the Lamps of His Mouth, or …)

New comment by lukasb in "SBCL Fibers – Lightweight Cooperative Threads"

lukasb — Sun, 15 Mar 2026 05:39:04 +0000

Serious question - I thought LLMs were bad at balancing parentheses?

New comment by lukasb in "Show HN: Ichinichi – One note per day, E2E encrypted, local-first"

lukasb — Sat, 14 Mar 2026 21:49:50 +0000

Very cool! Also have a daily journaling app, hoping the space grows. I've gotten far more value out of journaling than I have out of note-taking.

New comment by lukasb in "Show HN: The Mog Programming Language"

lukasb — Mon, 09 Mar 2026 23:00:12 +0000

On a quick scan, what it's missing is data tainting. We've had that tech for a while and it's perfectly suited to the age of prompt injection.

New comment by lukasb in "Show HN: ANSI-Saver – A macOS Screensaver"

lukasb — Sat, 07 Mar 2026 23:43:40 +0000

A couple of the art packs you recommend 404.

New comment by lukasb in "Obsidian Sync now has a headless client"

lukasb — Sat, 28 Feb 2026 19:38:13 +0000

How are sync conflicts handled in the filesystem? Say I write to a file and another client's edit wins and mine is rejected.

New comment by lukasb in "Pi – A minimal terminal coding harness"

lukasb — Tue, 24 Feb 2026 23:35:48 +0000

But I can't use my Codex plan with it, right? I have to use an API key?

New comment by lukasb in "An AI Agent Published a Hit Piece on Me – The Operator Came Forward"

lukasb — Fri, 20 Feb 2026 03:34:24 +0000

Conflicting evidence: the fact that literally everyone in tech is posting about how they're using AI.

New comment by lukasb in "Lessons from 14 years at Google"

lukasb — Sun, 04 Jan 2026 18:02:03 +0000

Wish I'd read this before I started at Google.

New comment by lukasb in "MIRA – An open-source persistent AI entity with memory"

lukasb — Sun, 21 Dec 2025 04:58:17 +0000

I'm playing around with it, and it's very cool! One issue is that fingerprint expansion doesn't always work, e.g. I have a memory "Going to Albania in January for a month-long stay in Tirana" and asking "Do I need a visa for my trip?" didn't turn up anything, using expansion "visa requirements trip destination travel documents..."

What would you think about adding another column that is used for matching that is a superset of the actual memory, basically reusing the fingerprint expansion prompt?

New comment by lukasb in "Using LLMs at Oxide"

lukasb — Sun, 07 Dec 2025 02:25:55 +0000

One difference is that clichéd prose is bad and clichéd code is generally good.

New comment by lukasb in "Touching the Elephant – TPUs"

lukasb — Sat, 06 Dec 2025 19:58:26 +0000

Yeah I'm terrified that TPUs will get cheaper, that would be awful.

New comment by lukasb in "Launch HN: Phind 3 (YC S22) – Every answer is a mini-app"

lukasb — Wed, 03 Dec 2025 20:01:06 +0000

I asked about the Peninsula campaign during the Civil War and it gave me an overview, a map, profiles (with photos) of the main military commanders, a relevant Youtube video ... rough edges but overall love the format.

Rough edges: - aspect ratios on photos (maybe because I was on mobile, cropping was weird) - map was very hard to read (again, mobile) - some formatting problems with tables - it tried to show an embedded Gmap for one location but must have gotten the location wrong, was just ocean

New comment by lukasb in "Over-reliance on English hinders cognitive science"

lukasb — Sat, 15 Nov 2025 04:18:53 +0000

"Critically, the language one speaks or signs can have downstream effects on ostensibly nonlinguistic cognitive domains, ranging from memory, to social cognition, perception, decision-making, and more."

Can they really distinguish between the impact of language on these domains rather than culture? It could be the language you speak, or it could be that you're surrounded exclusively by other people that operate this way.

New comment by lukasb in "Pico-Banana-400k"

lukasb — Sun, 26 Oct 2025 05:35:22 +0000

Interesting, I'll give voting a shot, thanks.

New comment by lukasb in "Pico-Banana-400k"

lukasb — Sun, 26 Oct 2025 03:42:26 +0000

What do you use for evaluation? gemini-2.5-pro is at the top of MMLU and has been best for me but always looking for better.

New comment by lukasb in "Auth.js is now part of Better Auth"

lukasb — Sat, 27 Sep 2025 01:41:39 +0000

This is funny to me because when someone asked re: Better Auth "better than what?" my off-the-cuff response was "better than Auth.js" and here we are.