Live data from Hacker News

Viewing profile — DalasNoin

DalasNoin

HN member
Joined
Tue, Sep 13, 2016, 10:16 PM UTC
HN karma
778
Public activity
177 items

About DalasNoin

I am Simon, an AI security researcher

https://www.linkedin.com/in/simon-l772743a7/

Recent public activity

  1. story
  2. story
  3. comment
    Comment #48207477

    Can you confirm people in france actually use wero? I had heard of it every so often but basically zero people actually use it, my revolut app has a feature to use wero but never u…

  4. comment
    Comment #47950613

    Quote from the article: ""AI will probably most likely lead to the end of the world, but in the meantime, there'll be great companies," Altman said in 2015." Altman wasn't even at …

  5. comment
    Comment #47861789

    I mean the best argument I see for cursor is that you can easily switch between AIs, which is convenient since they seem to run at 80-90% up time (with those 10-20% clustered at We…

  6. comment
    Comment #47168307

    Why does SynthID make it worthless? it helps other platforms detect this as ai?

  7. comment
    Comment #47155682

    We use semantic information inferred from comments and submissions. I think using stylometry would be a great addition, but it would be hard to google for "guy who writes fanciful …

  8. comment
    Comment #47155346

    We test different methods, in section 2, we use LLM agents to agentically identify people. We don't share any code here, but you could try with various freely available agents on y…

  9. comment
    Comment #47155311

    We essentially don't use stylometry but semantic information revealed from peoples' comments – clues and interests. (We use a little stylometry in a single experiment in section 5)…

  10. comment
    Comment #47155204

    We essentially don't use stylometry but semantic information – clues and interests.

  11. comment
    Comment #47155104

    That's a great background paper on the Netflix attack, we make a pretty direct comparison in section 5. We also try to use similar methods for comparison in sections 4 and 6. In se…

  12. comment
    Comment #47155036

    I agree that these accounts probably on average still contain more information than the average pseudonymous account. I think we could try to use the LLM to increasingly ablate mor…

  13. comment
    Comment #47154939

    We do advocate for stricter controls on data access on social platforms because of this. There is a bit of an unfortunate trade-off, but I think allowing mass-scraping or downloads…

  14. comment
    Comment #47154739

    Mitigations are pretty difficult, I understand it is kind of cool that some websites have really open APIs where you can just read everything. There are some cool apps that used HN…

  15. comment
    Comment #47154681

    There is also a practical issue here that people usually don't write a lot on linkedin, most people just have structured biographical information. We use very limited stylometry in…

  16. comment
  17. comment
    Comment #47154477

    We don't use (much) stylometry, so this won't help. This is totally something you could try, but we use interests and clues. Semantic information you reveal about yourself. The blo…

  18. comment
    Comment #47154445

    To be clear, we are making a clear concession here that the people weren't truly anonymous. But we did use an LLM to remove any identifying information from HN making them quasi-an…

  19. comment
    Comment #47144915

    There goes all the prompt engineering jobs

  20. story
    Large-Scale Online Deanonymization with LLMs

    Pdf: https://arxiv.org/pdf/2602.16800 (via https://arxiv.org/abs/2602.16800 )

  21. story
  22. comment
    Comment #47024076

    From what I understand this is a fully open-source bot that anyone can run with no restrictions. what a time to be alive, let's see what these bots will break first

  23. comment
    Comment #46885131

    The energy collected from the solar panels must be converted into heat in the AI chips. It's really like putting the AI chips directly into the sun, just with extra steps. Sunlight…

  24. comment
    Comment #46826606

    It's cold in a sense that is not very relevant. Your tumbler has a vaccuum layer because vaccuum does not transport or absorb any heat. you need those atoms to carry away heat.

  25. comment
    Comment #46824871

    Space doesn't seem like a good place to build datacenters at all. Cooling is going to be an enormous issue, how do you disperse of heat in a vaccuum? Radiators are very ineffective…