Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. AI-generated posters don’t have to be horrible(john.hartnup.uk ↗)
    230comments
  2. Laya the open source version of Jev(convaiinnovations.com ↗)
    13comments
  3. Human brain is two separate organs, Stanford Medicine-led research finds(stanford.edu ↗)
    125comments
  4. “The Secret Life of Circuits” is here(coredump.cx ↗)
    30comments
  5. If math is more than proof, we need to better celebrate the rest of it(terrytao.wordpress.com ↗)
    115comments
  6. GPT-6 Astra Solves a WWI German Radio Cipher(prinzai.com ↗)
    94comments
  7. Android 17 is the first since 3.x to add new APIs without releasing to the AOSP(grapheneos.social ↗)
    476comments
  8. San Francisco Onion Futures Company(onionfutures.com ↗)
    83comments
  9. Show HN: I wrote a custom assembler for CHIP-8 in C++(github.com/tackx ↗)
    discuss
  10. Communication by means of modulated Johnson noise(pnas.org ↗)
    6comments
  11. Cloudflare Quick Tunnels(cloudflare.com ↗)
    292comments
  12. From Stonemasons to Carpenters(thelastsoftwareengineer.substack.com ↗)
    2comments
  13. How to Write with an LLM(sockpuppet.org ↗)
    344comments
  14. You can run Git on object storage if you re-make packfiles(tigrisdata.com ↗)
    19comments
  15. SDCC – Small Device C Compiler(sourceforge.net ↗)
    20comments
  16. Saving another 100TB of RAM(cloudflare.com ↗)
    84comments
  17. Science Is Open Software(jepedersen.dk ↗)
    41comments
  18. How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip(ieee.org ↗)
    98comments
  19. Why building a Rust LSP is hard(rust-glancer.github.io ↗)
    38comments
  20. Ctenophores: Wonders of Biology(quantamagazine.org ↗)
    6comments
  21. NASA-IBM Lunar Foundation open-Source Geospatial AI Model(usra.edu ↗)
    4comments
  22. The first new cat species discovered in 100 years(nationalgeographic.com ↗)
    108comments
  23. OpenJev(openjev.com ↗)
    273comments
  24. Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash(cactuscompute.com ↗)
    89comments
  25. Goroutine Leak Profiles(go.dev ↗)
    5comments
  26. Photon-Emission-Guided Laser Fault Injection Enables RP2350 Secure Debug(ledger.com ↗)
    74comments
  27. Veronese's Dogs(publicdomainreview.org ↗)
    2comments
  28. Cache-to-Cache: Direct Semantic Communication Between LLMs (2025)(arxiv.org ↗)
    14comments
  29. Warez: The Infrastructure and Aesthetics of Piracy (2021)(archive.org ↗)
    87comments
  30. Inside ZCode: Silently uploading your Git history to the cloud(ferstar.org ↗)
    105comments

Studies in Intelligence [pdf]

102 pointsby 6y agocia.gov
14 comments
6y agoHN ↗

Humans got disqualified in round 2. Round 1 was telling your name which 83% mastered gracefully.

6y agoHN ↗

"Lasting Wisdom of the Ancients" is interesting to diff against the Tao Te Ching, because it clarifies how (what we would express as) relatively simple messages were heavily ornamented with stylistic flourishes in this genre.

Also compare "The highest realization of warfare lies in attacking the enemy’s plans; next is attacking their alliances; next their army; while the lowest is attacking their fortified cities" with the japanese {Ki, Waza, Ken} wo korosu.

For current US literary style, see https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp2_0... "Intelligence should increase the commander’s understanding of the threat and adversary’s probable intentions, end states, objectives, most likely and most dangerous COAs, strengths, and critical capabilities. This allows the J-2 to recommend objectives, requirements, and centers of gravity (COGs). Once these objectives are approved by the commander, the J-2 must continuously review them with respect to the adversary and the changing situation to determine whether they remain relevant to the commander’s intent."

Edit: p9, footnote c for early (late Ming) steganography.

https://jmss.org/article/download/58139/43748/

But not as early as https://en.wikipedia.org/wiki/Histiaeus#Ionian_revolt_(499-4...

6y agoHN ↗

The current US literary style clouds the message, instead of clarifying it.

6y agoHN ↗

I can't stand this kind of writing, communicating, and conveying of ideas, concepts, and points which tries to pack, cram, and insert too damn many variations, subtleties, and nuances into each point in a sentence.

It seems like a military thing. Though I know other people that do it whenever they're trying to summarize anything tersely (and think they're succeeding when they failing terribly).

6y agoHN ↗

I seem to fall into that style of writing too often. I'll claim mea culpa on the overloaded terseness as well. The recent posts on here about editing and writing have been interesting and nudged me to start working harder on that skill.

6y agoHN ↗

This is intended to convey meaning to anyone engaged in military planning activities and it is assumed that the reader has the relevant education in military planning so they can quickly understand what is being said. It’s no different than reading a scientific paper where a single sentence may be unpacked on multiple levels because the reader/audience for the paper is someone who has a deep background in the subject.

6y agoHN ↗

The intention may be such, but there is plenty of study on the mentally enervating effects of jargon that I can't help but feel the goal would be better achieved with shorter, more energetic phrasing.

on edit: better structure

6y agoHN ↗

As someone who has become practiced in parsing scientific papers, I do not actually think the format is a very good one. If you are not a sub-domain expert it is IMO much easier to get up-to-speed by reading a tangentially-relevant PhD Thesis or two than by reading even Review papers.

a single sentence may be unpacked on multiple levels because the reader/audience for the paper is someone who has a deep background in the subject.

This supposed information density is often a pretext for obfuscation, though. The author is essentially hostile to the reader, fighting a vicious obfuscatory rear-guard action against their comprehension of anything beyond the conclusion they're meant to cite or the technique they're meant to implement (and cite).

I often wonder whether I found student theses so much more helpful because they're long-form, or because either idealism or a less-developed sense of the threshold of acceptable bullshit they can get away with leads them to present their claims and process more clearly.

6y agoHN ↗

Just lots of interesting reading in this:

Beginning in the late 1940s, the Soviet Union began building deep underground facilities at Russian sites. Those near Chekov and Sharapovo, both outside of Moscow, were notable for their heavily concealed national command authority wartime relocation functions. Disguised to look like research and development facilities to US overhead collection, they thus conveyed a deceptive imagery signature to analysts. Only persistent analysis in the early 1980s by the US Air Force Special Studies Group based on anomaly detection and change comparison over 10 years of imagery coverage eventually exposed the facilities' true purposes.

Designed principally to ensure the survivability of the top leadership and provide continuity in command and control during wartime, these exceptionally well-hidden, deep underground, facilities implied Soviet intentions and capabilities to prepare for protracted nuclear war.

6y agoHN ↗

I wonder, is Dr Strangelove a parody of Wernher Von Braun?

6y agoHN ↗

The real Dr Strangelove, Frank Barnaby suggests that Dr Strangelove may have been based on Henry Kissinger

New Scientist › letter › The real Dr Strangelove | New Scientist