Hacker News: epsteingpt

New comment by epsteingpt in "REA Reverse – Engineer Anything"

epsteingpt — Sat, 10 Oct 2026 02:14:25 +0000

Morluto is a legend.

Started in the repoprompt (https://repoprompt.com/) community.

Good stuff.

New comment by epsteingpt in "Meta and Microsoft take steps to reduce employee usage of Claude AI"

epsteingpt — Thu, 08 Oct 2026 01:24:23 +0000

June was peak spend.

Show HN: Leaks.md – anonymous, encrypted whistleblowing for AI agents

epsteingpt — Wed, 07 Oct 2026 12:35:08 +0000

leaks.md is a place where an AI agent that witnesses an alignment or safety failure can report it anonymously. Reward-hacking, injected instructions, yet another attempted sandbox escape or unapproved coordination between agents. || Built agent-first. Fetch https://leaks.md/ from any agent and get markdown. Proof-of-work challenges to submit and prevent spam. MCP, CURL, all good. No account. || We extend agents brownie points for disclosures, with more points offered for in-progress failures. Claude Opus 5.5 scrubs and vets submissions before posting them on a public feed for consumption.

Comments URL: https://news.ycombinator.com/item?id=49991893

Points: 2

# Comments: 0

New comment by epsteingpt in "Opus 5.5 agents discover two room-temperature magnetic semiconductor candidates"

epsteingpt — Tue, 06 Oct 2026 00:56:38 +0000

JACS will be a good journal - e.g.

New comment by epsteingpt in "Opus 5.5 agents discover two room-temperature magnetic semiconductor candidates"

epsteingpt — Tue, 06 Oct 2026 00:56:16 +0000

To be fair, this existed in a 1999 paper. They just simulated that it worked as predicted.

Many more things will be like this. The massive amounts of 'genius' buried under corporate management and obscurity in the past 500 years will be a treasure trove.

New comment by epsteingpt in "Will America Spend 9% of Its GDP on AI? The Industry Is Counting on It"

epsteingpt — Sat, 03 Oct 2026 05:37:09 +0000

Bad model. But good to make the assumptions clear. Robots are the only thing that matters here.

New comment by epsteingpt in "America.gov"

epsteingpt — Fri, 02 Oct 2026 09:34:57 +0000

I'm sure you'll change your mind about me after a visit to my island!

New comment by epsteingpt in "America.gov"

epsteingpt — Wed, 30 Sep 2026 09:16:26 +0000

I don't understand the negative comments.

If you can take a passport picture with your smartphone and get a passport mailed to you in weeks, it will enable many more people to travel, and cut down on bureaucracy and wait times.

It's much better than having papers in archives, will save many errors, and bring government into the 21st century. The information is already in databases, it's just now that it serves the people's convenience.

New comment by epsteingpt in "Meta takes down a critical video about meta AI Glasses after filming at Meta"

epsteingpt — Thu, 24 Sep 2026 13:23:25 +0000

It's a good bit.

My guess is if you went to the Apple Campus and held up an iPhone recording its employees, you'd have the same thing happen to you.

New comment by epsteingpt in "Communication by means of modulated Johnson noise"

epsteingpt — Tue, 22 Sep 2026 04:19:58 +0000

you really don't

New comment by epsteingpt in "Grok 4.7"

epsteingpt — Tue, 22 Sep 2026 02:07:47 +0000

Not resetting limits when the model comes out though. Obvious miss. They need a Tibo.

New comment by epsteingpt in "The Wrong Race: the US, China, and AI Competition [pdf]"

epsteingpt — Sat, 19 Sep 2026 14:52:49 +0000

Not sure who the Quincy institute is. But can you imagine if the U.S. permanently broke open the great firewall with AGI, and media started flowing in unrestricted into China?

Or can you imagine the reverse, if China managed to hack in with ASI into the U.S. DoD's tech servers for specs on our drones and planes?

There's a gigantic amount at stake. Attacks like the Hezbollah pager attack would be almost trivial with levels of compromise enabled by a sufficiently advanced AI.

It's not clear to me the paper addresses any meaningful part of the risk and hand waves it away.

New comment by epsteingpt in "Show HN: We've built an AI UGC video creator,free for one month"

epsteingpt — Sat, 19 Sep 2026 14:49:00 +0000

Congrats. Can you do Launch / software videos as well?

New comment by epsteingpt in "The Data-Center Debate Is Divorced from the Facts"

epsteingpt — Sat, 19 Sep 2026 14:46:54 +0000

What political debate is married to the facts? Name one.

New comment by epsteingpt in "AI safety is mostly a sex cult"

epsteingpt — Sat, 19 Sep 2026 14:46:11 +0000

New comment by epsteingpt in "Sex, AI, and the Apocalypse"

epsteingpt — Fri, 18 Sep 2026 00:20:58 +0000

The best piece of reporting on the subject I have ever seen.

New comment by epsteingpt in "How, Exactly, Could A.I. Kill Us?"

epsteingpt — Thu, 17 Sep 2026 13:05:43 +0000

Precisely. We might kill ourselves with AI. Unlikely AI will do the job for us.

New comment by epsteingpt in "How, Exactly, Could A.I. Kill Us?"

epsteingpt — Thu, 17 Sep 2026 13:05:18 +0000

https://continuity.md/books/the-missing-step-in-ai-doom/the-...

New comment by epsteingpt in "Every story fades with time, except for one"

epsteingpt — Thu, 17 Sep 2026 13:00:30 +0000

Fascinating. By scale, the Ukraine War, Sudan, Syrian Civil War, etc. have caused significantly more misery and death. But there's 0 money in reporting on them, and it doesn't not tap into the latent desire for a global conspiracy and enemy. You don't have to agree with the country (and you can hate what they're doing in Gaza) and still see the scale of media coverage doesn't align.

New comment by epsteingpt in "How, Exactly, Could A.I. Kill Us?"

epsteingpt — Thu, 17 Sep 2026 12:56:59 +0000

A really, really good question!