sambejk
Re:Homomorphic Encryption for Pseudonymization – privacy enabled data-...
on Friday November 06, 2026 @06:01AM (#50390829)
Fully Homomorphic Encryption (FHE) allows the calculation on encrypted data. Invented in 2009, it was traditionally too slow to be realistic for real use. With new innovations, it has now become possible to run in production.
This blogpost shows a practical application that uses AES encryption in FHE - yes, you are encrypting under encryption - to mask identifiers to enable datasharing.
Re:Living Maps: Bay Area tech since 1960s
on Friday November 06, 2026 @03:35AM (#50389680)
Most of the finance and legal firms are still there. Guess tech companies also need lawyers and investment bankers.
Re:Chomsky and the Two Cultures of Statistical Learning
on Friday November 06, 2026 @03:31AM (#50389662)
Really great post, though a bit old, still relevant. Peter Norvig is one of the better thinkers of our generation.
Have to say that I confronted myself with the fact that I can't read longer posts anymore due to overuse of AI. Guess a detox is required...
Re:After Texas' blue wave, what happens to Elon Musk and SpaceX?
on Friday November 06, 2026 @03:24AM (#50389605)
let's hope for the best; though most people I talk to believe the DoJ is a ghost of its former self and it will take years to rebuild
Re:Google to cut Search access for schools that block new AI tools
on Friday November 06, 2026 @03:22AM (#50389601)
very fair
Re:Ask HN: Can current AI models reverse engineer CUDA drivers?
on Friday November 06, 2026 @03:22AM (#50389596)
what do you mean? most of those decompile projects were not even close the original software; for a driver, this would be completely different. i do believe there is some open-source drivers that are just not that good.
Re:Herdr vs. T3 Code vs. Amp: Software Factory Test
on Friday November 06, 2026 @03:20AM (#50389582)
Pretty cool! Do wonder how this is different than running plain codex or claude code.
Re:Ask HN: Can current AI models reverse engineer CUDA drivers?
on Thursday November 05, 2026 @03:45AM (#50376769)
I think it is unlikely that this would be possible right now. The complexity of the driver and how it interplays with the hardware makes this really difficult. However wondering why you would need this as they are pretty good and free?
Re:Nova beat Fable 5 with Opus 4.8
on Thursday November 05, 2026 @03:41AM (#50376738)
I'm a bit skeptical. This is just a single benchmark with an already capable (&closed-source) model. Would have been cooler if there was a more recent benchmark with an open-source model.
Re:Google to cut Search access for schools that block new AI tools
on Thursday November 05, 2026 @03:38AM (#50376711)
Schools have to choose between giving kids access to AI tooling (which has been proven to be terrible for education) or basic internet search.
Re:After Texas' blue wave, what happens to Elon Musk and SpaceX?
on Thursday November 05, 2026 @03:36AM (#50376699)
I don't expect anything big to change just because there is one election. The entire judicial system is still very much MAGA.
Re:GPT-6 Astra covertly underperforms on safety research
on Thursday November 05, 2026 @03:34AM (#50376680)
It feels like we have overfitted on safety by a long mile. This is how labs that care less about safety (a.k.a. Chinese & Grok) will probably be able to catch up in the future.
Re:Google to cut Search access for schools that block new AI tools
on Thursday November 05, 2026 @03:27AM (#50376623)
Google has long forgotten their old motto: "Don't be evil", but this is a new extreme. Luckily we have things like Duckduckgo
Re:Homomorphically encrypted CIFAR-10 inference in 200ms
on Friday July 17, 2026 @03:03PM (#48951062)
I think calculator is not part of the labels in cifar-10. Pretty normal as it has ‘only’ 90% accuracy. Things could improve quickly though
Re:Belfort Releases Fastest Encrypted Image Classifer
on Friday July 17, 2026 @09:08AM (#48946972)
Very cool to see practical applications of this technology!