Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. LLMs in Professional Software Engineering (knorpelsenf.me)
    —discuss
  2. Can Life Be Explained by Physics? (Featuring Prof. Brian Cox) [video] (youtube.com)
    —discuss
  3. ETH Computer Science enrolment saw a 15 percent decline (ethz.ch)
    —discuss
  4. Unified U.S. Site Blocking Bill Targets ISPs and DNS Resolvers but Spares VPNs (torrentfreak.com)
    —discuss
  5. Show HN: Free alternative to graphics design giants (scissor.studio)
    —discuss
  6. Are Big Tech bonds crowding out the US Treasury? (ft.com)
    1comments
  7. The Risks of Ignoring API Security During Mobile App Security Testing
    —discuss
  8. How I Use "AI" (2024) (carlini.com)
    —discuss
  9. Show HN: Redthread – autonomous LLM pentesting with proof-of-concept exploits (millenniums.ai)
    —discuss
  10. Pixi adds non-Python deps to Python notebooks (marimo.io)
    —discuss
  11. Show HN: Impressive hand detection AI algorithm (github.com/aryafarkhondeh)
    1comments
  12. The Cybersecurity Siege Economy (dntls.substack.com)
    —discuss
  13. European industry is doing better than you may think (economist.com)
    —discuss
  14. Doing math together in the age of AI (dustingmixon.wordpress.com)
    —discuss
  15. Show HN: Caspian – Desktop app to build and launch small vibe-coded projects (caspian.md)
    —discuss
  16. What if AGI was invented in Ancient Greece? (echohive.ai)
    —discuss
  17. AI companies in race to demonstrate their model most threatening to humanity (thecivilian.co.nz)
    —discuss
  18. 37,500 border drawings: a map of the world as people remember it (habibicode.org)
    3comments
  19. iOS 27's Reworked Stub Islands (codecolor.ist)
    —discuss
  20. Slow Servers (slowservers.net)
    —discuss
  21. Detecting Compromised AI Coding Agents with Jev and Gryph (safedep.io)
    —discuss
  22. Cloud Emulators (github.com/upgundecha)
    2comments
  23. File Notification Attacks (inoti.fyi)
    —discuss
  24. Turbo Vision for Rust (enzolombardi.net)
    —discuss
  25. Researcher discloses vulnerabilities to OnePlus, they respond with legal threats (notebookcheck.net)
    —discuss
  26. Japan Electric Vehicle Association 1980年 [video] (youtube.com)
    —discuss
  27. Maybe don't let Muse run your Facebook Marketplace account (threads.com)
    13comments
  28. The Overhang (oneusefulthing.org)
    —discuss
  29. OpenAI pauses some training amid allegations its rogue agents behaved more (theregister.com)
    —discuss
  30. Show HN: A website for sharing Muse AI invitation codes (museaicode.cc)
    —discuss

Maybe don't let Muse run your Facebook Marketplace account

22 pointsby 49m agothreads.com
13 comments
37m agoHN ↗

tell boomers its intelligent again and again

they will hook up actually important stuff to agents because its the future

???

400 dead

22m agoHN ↗

"that's on me, I'm owning that, I'm sorry, that shouldn't have happened"

33m agoHN ↗

we need a Nitter for Threads

apparently people use Threads. I suppose the same kind of people who connect Muse to Facebook Marketplace.

29m agoHN ↗

I initially found screenshots of this on Bluesky, figured I'd link to the original source given Threads seems to at least allow us to read without logging in.

With that said, in case Threads isn't available for whatever reason, I've uploaded screenshots of everything (I think?) here: https://imgur.com/a/mceW9WF

27m agoHN ↗

It will be chaos if people let AI agents run amok with their accounts.

Many more such cases are to come.

22m agoHN ↗

Bonus score if your Meta AI Agent does something with your account that the Meta AI Moderator deems inappropriate and bans you.

25m agoHN ↗

"that's on me"

Nice touch by the mechanical parrot, to worthlessly owning it.

19m agoHN ↗

Especially considering the user was mad about the agent doing stuff by itself, so it tried to "fix" this by sending an apology to the buyer without explicit approval of the person "running" the agent, seemingly understanding nothing from the conversation.

Wonder what quantization Meta runs these models on, Q4?

22m agoHN ↗

The AI is not remotely ready for this - this is an absolute delusion they are selling.

"By the way don't do this again" <- as if the AI has the ability to ingest and systemically diffuse this.

I think Zuckerberg himself is deeply into the Koolaid, and is likely himself unaware of the limits of this tech.

He's probably surrounded by enablers.

21m agoHN ↗

Fake surprise - it's obvious that would happen, he did it deliberately for clicks and engagement.

19m agoHN ↗

It's plausible, but there's every reason to believe that someone would trust the technology handed to them by a megacorp to do as advetised.

The 'deliberate failure' is on Meta here.

12m agoHN ↗

Or he just wanted to test if it works for clicks and engagements, and he got a result he didn't anticipate.