Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. San Francisco Onion Futures Company(onionfutures.com ↗)
    34comments
  2. Android 17 is the first since 3.x to add new APIs without releasing to the AOSP(grapheneos.social ↗)
    357comments
  3. Typesafe-computer-use drives a Mac toward a goal for 1/50th of a cent per step(github.com/awlevin ↗)
    9comments
  4. SDCC – Small Device C Compiler(sourceforge.net ↗)
    11comments
  5. Science Is Open Software(jepedersen.dk ↗)
    22comments
  6. Cloudflare Quick Tunnels(cloudflare.com ↗)
    272comments
  7. Human brain is two separate organs, Stanford Medicine-led research finds(stanford.edu ↗)
    1comments
  8. How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip(ieee.org ↗)
    74comments
  9. Why building a Rust LSP is hard(rust-glancer.github.io ↗)
    18comments
  10. Saving another 100TB of RAM(cloudflare.com ↗)
    59comments
  11. NASA-IBM Lunar Foundation open-Source Geospatial AI Model(usra.edu ↗)
    discuss
  12. How to Write with an LLM(sockpuppet.org ↗)
    308comments
  13. Xcode 27.1 Beta Release Notes(developer.apple.com ↗)
    81comments
  14. The first new cat species discovered in 100 years(nationalgeographic.com ↗)
    88comments
  15. Goroutine Leak Profiles(go.dev ↗)
    2comments
  16. Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash(cactuscompute.com ↗)
    82comments
  17. You can run Git on object storage if you re-make packfiles(tigrisdata.com ↗)
    1comments
  18. OpenJev(openjev.com ↗)
    255comments
  19. Photon-Emission-Guided Laser Fault Injection Enables RP2350 Secure Debug(ledger.com ↗)
    64comments
  20. The Farnese letter(simonklee.dk ↗)
    6comments
  21. Cache-to-Cache: Direct Semantic Communication Between LLMs (2025)(arxiv.org ↗)
    12comments
  22. Minimal Phone 2(minimalcompany.com ↗)
    205comments
  23. Cyclomatic Complexity in C#(ndepend.com ↗)
    17comments
  24. Claude Code now reads AGENTS.md if there is no Claude.md(claude.com ↗)
    214comments
  25. LispBM is a concurrent Lisp for microcontrollers with message passing(lispbm.com ↗)
    3comments
  26. Inside ZCode: Silently uploading your Git history to the cloud(ferstar.org ↗)
    96comments
  27. Warez: The Infrastructure and Aesthetics of Piracy (2021)(archive.org ↗)
    43comments
  28. Alibaba open-sources AI model that can detect cancer and nearly 150 conditions(scmp.com ↗)
    11comments
  29. How SpaceX streamlined the Raptor engine(construction-physics.com ↗)
    67comments
  30. The Implications of Linguistic Illegibility for LLM Security(arxiv.org ↗)
    26comments

Anthropic CEO admits compromising with authoritarian regimes to get AI funding

10 pointsby 1y agothe-decoder.com
4 comments
1y agoHN ↗

"principles"

No AI company sits forever on their original "principles" when $100B is on the line.

"This is a real downside and I'm not thrilled about it."

He is beyond thrilled.

It will bring Anthropic closer to true "AGI": (The super-cars, pent-houses, mansions and 10+ new family offices for some early employees.)

Like the PayPal Mafia, we will also get: The Anthropic Mafia or something equivalent.

1y agoHN ↗

No AI company sits forever on their original "principles" when $100B is on the line.

Yup. This is why no genAI company can ever be trusted, no matter how trustworthy they may appear to be.

1y agoHN ↗

AI won't be the last frontier. Other mass-capital-intensive ventures and industries will arise. Those who have massive capital are going to grow it because you need to come to them. Then they can spread whatever ideology they have even better.

1y agoHN ↗

I sincerely don't understand the surprise at all. AI is perfect for authoritarian regimes. Typically, they are not as efficient as more traditional democracies because they don't allow the markets to function as freely as in democracies, (even though in democracies there's a lot of corruption as well..) But AI can get around that because it replaces a lot of labour and can do things efficiently without the market.

I'd expect AI to be one of the prime weapons of choice of authoritarian regimes in the future, and no amount of regulation can stop it.

The only thing us normal people will ever get out of AI is increased efficiency at the cost of everyone else competing at an increased efficiency level via the prisoner's dilemma. And the people profiting off that series of prisoner's dilemma games will exactly be companies like Anthropic and dictators or quasi-dictators that can use their existing power to leverage the dominating effect of AI.

The most despicable and dark beauty of this scheme is that we will be completely distracted by the overwhelming information and novelty of it.