blintz

18 ms·

91.
by blintz 4y ago
Yeah, we think private lookup could make distributed retrieval protocols like IPFS, BitTorrent, etc significantly more private. Currently, they are in some ways much worse from a privacy perspective than centralized alternatives, since they
92.
by blintz 4y ago
I think FHE is so powerful that we tend to let the cool possibilities distract from really practical and useful stuff that is possible today. Today, I think just the ability to do private lookups is, while super simple from an academic lens
93.
by blintz 4y ago
Hey, creators here, cool to see people excited about this! We're in the YC W23 batch, happy to answer any questions folks have.
94.
Show HN: Privately resolve “.eth” URLs using homomorphic encryption(sprl.it)
2 points by blintz 4y ago|0 comments
95.
by blintz 4y ago
That’s a fair point, although you can use a debit card for cannabis delivery services too. Maybe a good mitigation for the fraud concern is just using ‘virtual’ or physical prepaid debit cards that have a fixed balance.
96.
by blintz 4y ago
I guess I don’t quite understand why using something like a debit card doesn’t work. The marijuana companies seem to have effectively figured how to take payments with almost no help from the traditional players (credit card companies, bank
97.
by blintz 4y ago
Happy to answer any questions! Yeah, we're taking a narrow application (fetching from a database) and making it really practical. You can privately check a Bitcoin balance at https://btc.usespiral.com or privately read Wiki
98.
by blintz 4y ago
It does - for databases of a reasonable size, this is okay. Roughly 1-50 GB databases are quite practical.
99.
by blintz 4y ago
Yeah, I‘d agree that performance is the crux of the issue for many PETs. I am not terribly confident that over the next 5 years we will see anything like a 10x or 100x improvement in general fully homomorphic encryption. Instead, I think we
100.
by blintz 4y ago
There's no single way to do this kind of verification. Different applications will have different 'roots of trust' for the authenticity of database data. For example, for Wikipedia, we could imagine the Wikipedia site serving
101.
by blintz 4y ago
Yeah, a lot of cool folks are working on doing ML on encrypted data (Zama, Enveil, Duality). It's a really challenging problem, and I think it will take a big, well-funded team to get working well. We're partly interested in PIR b
102.
by blintz 4y ago
Yeah, the challenge always seems to building privacy technologies that people are actually willing to pay for. The one trove of public data that folks seem really eager to access in a metadata-private way seems to be blockchain data. As s
103.
by blintz 4y ago
I think it will be if we're successful! I do think the cool thing about what we're building is you don't really have to understand exactly what homomorphic encryption is to use / benefit from it. We're really just b
104.
by blintz 4y ago
Author here, happy to answer any questions! I know it’s a complicated topic, and I’m still trying to figure out how best to explain it. Also, we’ve applied to YC for W23 to commercialize and expand this tech; wish us luck!
105.
Explained from scratch: private information retrieval and homomorphic encryption(blintzbase.com)
151 points by blintz 4y ago|22 comments
106.
by blintz 4y ago
Hi, creators here. We're applying to YC so we'd love any feedback you have, and happy to answer any questions. We also did a private Wikipedia demo (spiralwiki.com) here a while back. We answer queries privately using homomorphic
107.
Show HN: Private Bitcoin block explorer using homomorphic encryption(btc.usespiral.com)
3 points by blintz 4y ago|1 comments
108.
by blintz 4y ago
The content policy is strikingly puritanical: > "Do not attempt to create, upload, or share images that are not G-rated" https://labs.openai.com/policies/content-policy
109.
by blintz 4y ago
On the public transit point: fun fact, BART (a large SF/Bay Area transit system) trains replace all of the air in each train car every 70 seconds, which is a much higher rate of airflow than most other public spaces. Since they also us
110.
by blintz 4y ago
I am so excited about this news. I understand that some people are pessimistic, and view it as a "giving up" on complete security against nation-states. I think that's the wrong way to analyze the situation. The dream I have
111.
by blintz 4y ago
Yes, all of the mapping of article titles to index happens locally. If the index got too large, we could use a hash to bucket things, but this incurs some waste and would only support exact article title matches. Yes, it is definitely not a
112.
by blintz 4y ago
It's generally multiplicative depth; that is, the number of ciphertext-ciphertext multiplications that feed into each other. The primary way we define and evaluate functions in FHE is as Boolean (or sometimes arithmetic) circuits, so &
113.
by blintz 4y ago
I do indeed use all 6 cores of my server simultaneously, and the caching on the database part of memory is quite effective. If I had a little spare cash I might just purchase some more CPUs :-)
114.
by blintz 4y ago
Yes, the whole database is scanned to answer a single query.
115.
by blintz 4y ago
It would be very cool to federate this! I am particularly interested in applications to DHTs/Kademlia, especially for things like IPFS. Since this has only recently become practical, there is a bit of dearth of resources on it at the m
116.
by blintz 4y ago
Yeah, I mean you generally have to go with trust on first use at some point. You can also do code signing, check hashes, build from source, compare multiple sources, etc. All the standard software supply chain security measures.
117.
by blintz 4y ago
This is PIR. You do have to download the whole database for information-theoretic security, but not for computational security. If you assume the hardness of some problem (in this case, lattices, but it is also possible from RSA, ECC, etc)
118.
by blintz 4y ago
Homomorphic encryption actually causes some interesting asymmetries here. It is very important to have an open source client, since of course a malicious client could trivially leak your desired index. However, the server’s operation is tot
119.
by blintz 4y ago
We never encrypt the database. Only the query is encrypted. The client setup data is only dependent on the client key and the size of the database (not the content). Adding and replacing articles can happen whenever the server wants, and cl
120.
by blintz 4y ago
You would. It’s important for the client to pace it’s requests in a way that does not reveal too much (for example, the client should not just request both chunks of the article at the same time). The best thing to do would probably be to h

More ›