Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Claude discovers a novel enzyme system with CRISPR-like repeats(anthropic.com)
    180comments
  2. Fixing the Portobello Police Station Clock(pointinthecloud.com)
    68comments
  3. How we made claude.ai 3x faster in two weeks(claude.dev)
    13comments
  4. A brief history of Windows scroll bar shortcuts(devblogs.microsoft.com/oldnewthing)
    16comments
  5. Cloud Agents Are Inevitable AI Prisons(normanponte.io)
    17comments
  6. Italian parliament votes for return to nuclear energy(apnews.com)
    160comments
  7. Gemini 3.8 text-to-speech(blog.google)
    94comments
  8. Radicle: Disclosure of Vulnerability in the Network Protocol(radicle.dev)
    29comments
  9. Jev in 25 Lines of Python(nobodywho.ai)
    173comments
  10. GPT-6 Sol and Luna(openai.com)
    814comments
  11. Claude Opus 5.5(anthropic.com)
    1060comments
  12. Claude Code reads AGENTS.md only when telemetry is on [fixed](szypowi.cz)
    234comments
  13. Z80 REPL (2018)(abagames.github.io)
    17comments
  14. Stripe's Knowledge AI Platform(stripe.dev)
    95comments
  15. I don't want the details(michaelheap.com)
    169comments
  16. UK military jamming other nations' satellites to defend itself, BBC told(bbc.com)
    115comments
  17. Show HN: I built a post-mortem debugger for native Windows x64/x86 crashes(forensicdbg.com)
    discuss
  18. Show HN: Conway's Game of Life in boot sector(github.com/0xax)
    1comments
  19. QuestDB (YC S20) Is Hiring a Sales Engineer(questdb.com)
    discuss
  20. Seattle City Council votes to ban surveillance pricing in sale of groceries(consumerreports.org)
    110comments
  21. 28% of job postings on company career sites have been open over 90 days(unlisted.careers)
    217comments
  22. Web-based IBM 1620 emulator and IPL-V from 1963(github.com/pkimpel)
    6comments
  23. Tokens Too Cheap to Meter(jyn.dev)
    150comments
  24. OpenAI GPT–6 Astra breaks Enigma message that has resisted solution since 2005(cryptocellar.org)
    431comments
  25. Transit rewards(waymo.com)
    293comments
  26. Strands Harness(strandsagents.com)
    84comments
  27. GPT-6 Astra has gained the ability to drive a car(drivingbench.com)
    199comments
  28. What California is learning from solar panels built over irrigation canals(kqed.org)
    664comments
  29. How did AMD Ryzen get 50% faster in two years?(lemire.me)
    190comments
  30. Microsoft killed FoxPro in 2007. Anyway, here's FoxPro revived(foxscript.org)
    243comments

PRC-linked influence operations are targeting AI debates in the US

19 pointsby 3mo agoopenai.com
15 comments
3mo agoHN ↗

This shit's happening with every topic now. I really wish we'd just start cutting out massive blocks of internet and phones so our old people are not scammed, so our young are not radicalized and our leaders convinced something crazy is happening any more.

3mo agoHN ↗

America already lost the battle. Multiple administrations on both sides of the aisle have reported China having access to America's lawful intercept wiretap system: https://en.wikipedia.org/wiki/Salt_Typhoon

No amount of authoritarian backpedaling will put the genie back in the bottle. They've got blackmail on our politicians, backdoors in our telecoms, and Chinese apps at #1 on the App Store. America is paying for "freedom" through the nose right now, the will to quit using social media and smartphones has to come from within.

3mo agoHN ↗

By freedom you mean freedom to invade other countries? Then don't be surprised those other countries fight back.

3mo agoHN ↗

You might want to read about before WWII

3mo agoHN ↗

America is paying for "freedom" through the nose

if you ask people on the street i think more than half of them will say they dont care if china gets ahead. even with conservatives who dont like china its more about keeping the american way of life as in "i dont want them to bring communism here", not some global clash of civilizations thing. if they really hated china those apps would never get to #1. its obvious that american citizens want more friendly relations or at least keeping it at the current level instead of tariffs and war mongering.

the only ones who really want a cold war is the corrupt national security industrial complex, think palantir and anthropic and lockheed martin. theres a lot of money to be made from selling overpriced missiles to the air force and forcing americans to use bad quality more expensive ai providers after the chinese or eu based get banned for "security" reasons.

even the fact that "national security" exists as a concept separate from public security means theres a big misalignment in the system. the purpose of a state is to serve and protect its people, not secure its power or even its own existence. as long as the cia, nsa, ice, dea and all the other agencies are allowed to do influence ops on their own citizens america cant be a true democracy. and the same goes for the uk, china, india, germany, really any country thats big and powerful enough that this kind of corruption can start.

3mo agoHN ↗

AI is able to recursively improve its code now, and Anthropic warns we should slow down development before it gets too advanced (reported 2019, and 2026)

PRC-linked influence operations are targeting AI debates

Anthropic and OpenAI holds all the power and will create world utopia

oh btw, we will both IPO after SpaceX

lol, clown world.

3mo agoHN ↗

In this report, we describe two clusters of ChatGPT accounts likely originating from China that we banned after they used our models in support of apparent covert influence operations that promoted narratives in an attempt to manipulate a legitimate debate about American AI and wider tech policies.

Why would people in China choose ChatGPT for this? Surely one of the multiple Chinese LLMs would be perfectly suitable for generating social media posts, no?

I don't doubt that the PRC is up to astroturfing shenanigans (they do have a vested interest in convincing the American public to kneecap America's tech sector), but using American LLMs via American services seems like an odd choice for that.

3mo agoHN ↗

It's entirely possible that the vast majority of those involved do use Chinese LLMs. Some small portion may have used ChatGPT for higher quality posts.

3mo agoHN ↗

Some small portion may have used ChatGPT for higher quality posts.

Deepseek, Kimi, GLM have comparable quality to GPT for this purpose. More likely they are lying about it to discredit real opposition to data center expansion.

3mo agoHN ↗

i think people misunderstand foreign influence with the wrong POV. they dont need to be sided. in democracy, it functions when theres smooth transitions, fact finding, debates

these campaigns are equally about polarization. the goal is to keep people from acting rational and foment irrational behavior.

it should be obvious to people in america and britain that this tactic works.

now fuck data centers, but you cant use "they had POV X" as a deciding factor on anything. aside from lying, theyre an AI company so the bullshit could come from anywhere.

3mo agoHN ↗

“The PRC wants you to regulate us so don’t do it.”

Gotcha.