Hacker Newsnew | past | comments | ask | show | jobs | submit | g-b-r's commentslogin

but the paper is not about LLMs

Maybe you don't.

Some put off using the service and look for alternatives for as long as possible (often ever) if they're presented with tomes of legal documents to accept



Lolz.

Also, the qubit is a dead end!

The future of quantum computing lies in quantum holography!

Future generations will program quantum memory through constructive and destructive interference. Predicting hash space and potential distribution over complex dimensionality will have their day, simply not any time soon.

The qubit is a dead end! Mark my words! There is substantially more information density available than spin disposition. Algorithmically manipulating holographic wave fronts will become the future of quantum computing.



Maybe you allow tracking and cookies?

I don't, and I rarely have issues with firefox. Private + blockers + VPN causes, expected, issues but otherwise i'm usually fine?

The owners of the other companies are at least not as openly opposed to democracy, though.

Meta sure causes more damage right now, but banning Palantir, which wouldn't even cause big problems, is an absolute no-brainer


Being "open" about it or not doesn't matter. In a sense it's even worse if they're less upfront about it.

Larry Elisson, Jeff Bezos and Mark Zuckerberg are cut from the exact same cloth. It's beyond doubt.


Hmm well except Oracle's owner..

That's probably the places where people would never wash their hands either

This is insane, I checked if it was April the 1st

Despite the name of the website this is in all likelihood not a zero-day: they say it was discovered by an employee of the Zero Day Initiative, so it should have only only been reported to Telegram, and besides them only TrendAI (TrendMicro) should know anything about it

This was mind-blowing to me.

I disagree with the author, it's not that sha-pinning is dangerous, it's that it is astonishingly broken in GitHub.

If they're unable to guarantee that the hash really belongs to the repository you indicated them, it would be better if they didn't provide the pinning feature at all.

And if they built their systems such that verifying it is unfeasible, they're just broken, and using GitHub should be considered a risk.

Although in truth, you should always carefully check a PR like that; even without referencing a different repository, a malicious user could just make it point to an outdated, vulnerable version.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: