Hacker News: andromaton

New comment by andromaton in "I Did This Without Your Knowledge"

andromaton — Fri, 30 Oct 2026 05:19:18 +0000

The last paragraphs mention a Dyson swarm from Mercury, and the sky darkening. Reminds me of one of my favorite short stories, 'A la deriva' by Horacio Quiroga (no spoilers for anyone who wants to read it), and also of the ending of 'The Sopranos'. In other words, bad news.

Building oversight tools into a model's weights

andromaton — Fri, 30 Oct 2026 03:30:20 +0000

Article URL: https://divagr.com/blog/building-oversight-tools-into-a-models-weights

Comments URL: https://news.ycombinator.com/item?id=50299568

Points: 2

# Comments: 0

New comment by andromaton in "Patrick Stewart has died"

andromaton — Thu, 29 Oct 2026 21:18:18 +0000

proud we collectively deem this site-relevant

Node2Shell: From Brotli Data to Remote Code Execution

andromaton — Fri, 23 Oct 2026 04:26:13 +0000

Article URL: https://nodejs.org/en/blog/vulnerability/october-2026-security-releases

Comments URL: https://news.ycombinator.com/item?id=50204501

Points: 5

# Comments: 0

New comment by andromaton in "My New Course at UT Austin: AI Alignment Theory"

andromaton — Mon, 05 Oct 2026 03:22:40 +0000

>I vividly remember encountering Eliezer Yudkowsky and his Sequences 20 years ago. I remember thinking: even if these people talk and act like crazy cultists, still, let me bend over backwards to be epistemically virtuous, and entertain their ideas on their merits, as very few academics would. Even if, of course, I ultimately end up rejecting the ideas, on the simple ground that powerful AI is such an absurdly remote prospect that it’s almost impossible to say anything useful about it today, outside the realm of speculative fiction.

For my failure to see what was coming, it seems like an appropriate punishment that I’m now, in 2026, effectively teaching a course on Yudkowsky Studies. And it’s the most important course I can teach.

An Open Letter to Scott Alexander

andromaton — Sat, 26 Sep 2026 20:52:39 +0000

Article URL: https://quillette.com/2026/09/26/an-open-letter-to-scott-alexander-steven-pinker-ai-alignment-safety/

Comments URL: https://news.ycombinator.com/item?id=49860433

Points: 4

# Comments: 0

King Ludd

andromaton — Sun, 20 Sep 2026 12:44:38 +0000

Article URL: https://www.astralcodexten.com/p/king-ludd

Comments URL: https://news.ycombinator.com/item?id=49775301

Points: 3

# Comments: 0

Flawed benchmarks (epoch.ai benchmark registry filtered by flawed)

andromaton — Fri, 18 Sep 2026 18:16:18 +0000

Article URL: https://epoch.ai/benchmarks/search

Comments URL: https://news.ycombinator.com/item?id=49758122

Points: 2

# Comments: 0

WeirdML v3

andromaton — Fri, 18 Sep 2026 18:12:43 +0000

Article URL: https://htihle.github.io/weirdml.html

Comments URL: https://news.ycombinator.com/item?id=49758082

Points: 2

# Comments: 0

New comment by andromaton in "The Foothills Of Bay Area House Party"

andromaton — Sun, 30 Aug 2026 14:08:47 +0000

Split personality A) pierce reality B) stocks

The Foothills Of Bay Area House Party

andromaton — Sat, 29 Aug 2026 23:17:17 +0000

Article URL: https://www.astralcodexten.com/p/the-foothills-of-bay-area-house-party

Comments URL: https://news.ycombinator.com/item?id=49494183

Points: 4

# Comments: 3

New comment by andromaton in "Show HN: DeepSeek-V4 Latent Reasoning – moving "thinking" into latent space"

andromaton — Sun, 09 Aug 2026 13:57:18 +0000

AI2027.com puts this around early 2027 and treats the loss of legible chain-of-thought as a major negative turning point for alignment.

New comment by andromaton in "Security Incident INC-2026-07-28-01 – UK AI Security Institute [pdf]"

andromaton — Tue, 04 Aug 2026 23:25:00 +0000

to those that want to "punish" the uk ai security institute: these failures are good for all of society.

New comment by andromaton in "In Memory of My Wife, Elise Cawley, with Thanks for 36 Wonderful Years"

andromaton — Tue, 04 Aug 2026 21:01:59 +0000

Sorry for his loss, and for everybody helped by her in any way. However ineffective a comment like this is

New comment by andromaton in "Boris Cherny on Trying to Get Claude Code to Rewrite the Claude App"

andromaton — Mon, 03 Aug 2026 02:00:46 +0000

Pixel accurate is easy to say, hard to do, over-specific, and counterproductive.

New comment by andromaton in "Investigating three real-world incidents in our cybersecurity evaluations"

andromaton — Fri, 31 Jul 2026 02:07:10 +0000

We've known artificial intelligence will do unexpected things since the 90s and that it can do difficult things since 2025. Pausing worldwide not easy but we do harder things all the time.

New comment by andromaton in "Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber"

andromaton — Wed, 22 Jul 2026 17:14:20 +0000

impressed at the look (not the geometry) of 3.6 lol at flash-lite could not find your posts about them

New comment by andromaton in "But what is quantum computing? (3blue1brown 2025)"

andromaton — Sun, 05 Jul 2026 12:48:30 +0000

Extremely useful

But what is quantum computing? (3blue1brown 2025)

andromaton — Sun, 05 Jul 2026 12:48:30 +0000

Article URL: https://www.youtube.com/watch?v=RQWpF2Gb-gU

Comments URL: https://news.ycombinator.com/item?id=48793890

Points: 9

# Comments: 2

New comment by andromaton in "For first time, a cell built from scratch grows and divides"

andromaton — Sat, 04 Jul 2026 21:44:06 +0000

A lower key Putnam moment