Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Fully homomorphic encryption and the dawn of a private internet

bozmen.io

Post reply on HN
Print view View original

111–120 of 202 posts

Re: Fully homomorphic encryption and the dawn of a private internet

#111

Post by crtasm »

ralferoo wrote: post #104
hoppp wrote: post #76

Earlier quoted context omitted.

But if you have a lot of data, self hosting is still cheaper. Its always gonna be cheaper because you don't have the cloud provider's profit margin, which can be quite high.

It can be quite high, but it doesn't have to be. For instance, I have a 7TB storage server from Hosthatch that's $190 for 2 years. That's $7.92 per month, or £5.88 at today's exchange rates. That's under 20p per day. Just on electricity costs alone, this is good value. My electricity costs are 22.86p/kWh which is pretty cheap for the UK. That means that if having that drive plugged in and available 24/7 uses more tha…

Do they offer deals like that often? List price is "from $24/month" for 6TB (no further details provided without registering an account).

Re: Fully homomorphic encryption and the dawn of a private internet

#112

Post by latentsea »

barisozmen wrote: post #8
bruce511 wrote: post #7

I get the "client side" of this equation; some number of users want to keep their actions/data private enough that they are willing to pay for it. What I don't think they necessarily appreciate is how expensive that would be, and consequently how few people would sign up. I'm not even assuming that the compute cost would be higher than currently. Let's leave aside the expected multiples in compute cost - although the…

An FHE Google today would be incredible expensive and incredibly slow. No one would pay for it. The key question I think is how much computing speed will improve in the future. If we assume FHE will take 1000x more time, but hardware also becomes 1000x faster, then the FHE performance will be similar to today's plaintext speed. Predicting the future is impossible, but as software improves and hardware becoming faster…

But think of the children?

Re: Fully homomorphic encryption and the dawn of a private internet

#113

Post by ralferoo »

gblargg wrote: post #26

The idea that these will keep being improved on in speed reminds me of the math problem about average speed: > An old car needs to go up and down a hill. In the first mile–the ascent–the car can only average 15 miles per hour (mph). The car then goes 1 mile down the hill. How fast must the car go down the hill in order to average 30 mph for the entire 2 mile trip? Past improvement is no indicator of future possibilit…

Is the downhill section a cliff? Google informs me terminal velocity of a car is 200-300mph, so to fall a mile at 300mph, the car will need 12 seconds, so let's round up to 15 seconds to account for the time it's accelerating.

To cover the full 2 miles at an average of 30mph, we need to complete the entire journey in 4 minutes, leaving 225 seconds for the ascent.

We know that the old car was averaging 15 miles per hour, but the speedo on an old car is likely inaccurate, and we only need to assume a 6% margin of error for the car to show 15 miles per hour and cover the mile in 225 seconds. You probably couldn't even tell the difference between 15 and 16 on the speed anyway, but let's say that we also fitted out the car with brand new tyres (so the outer circumference will be more than old worn tyres), and it's entirely possible.

So, let's say 240mph. That's the average speed of our mile freefall in 15 seconds.

Re: Fully homomorphic encryption and the dawn of a private internet

#115

Post by ralferoo »

crtasm wrote: post #111
ralferoo wrote: post #104

Earlier quoted context omitted.

It can be quite high, but it doesn't have to be. For instance, I have a 7TB storage server from Hosthatch that's $190 for 2 years. That's $7.92 per month, or £5.88 at today's exchange rates. That's under 20p per day. Just on electricity costs alone, this is good value. My electricity costs are 22.86p/kWh which is pretty cheap for the UK. That means that if having that drive plugged in and available 24/7 uses more tha…

Do they offer deals like that often? List price is "from $24/month" for 6TB (no further details provided without registering an account).

They tend to do promotions, typically only valid for 24h and only advertised on certain forums like LET, a couple of times per year - typically at least around their company anniversary date or Black Friday.

There are others too, e.g. Servarica who keep their Black Friday offers running all year round.

Re: Fully homomorphic encryption and the dawn of a private internet

#116

Post by Retr0id »

immibis wrote: post #109
Retr0id wrote: post #91

Earlier quoted context omitted.

SGX on consumer client devices is sucky for that reason, but SGX on the server can be used to defend user interests. If I put my sensitive customer data inside SGX (such that I can operate on it but not extract it), and the nation-state adversary says "we have a warrant for your customer data, hand it over", I can reasonably say "I can't". I could also produce attestations that my code really is running inside SGX, v…

SGX on the server is breakable if and only if SGX on the client is breakable. You can either own other people's computers, or you can prevent other people owning your computer. You can't eat your cake and have it. Yes, it might be good for ass-covering as you indicate. A lot of ineffective technical solutions are effective legal liability shields anyway. But if this becomes mainstream, the NSA will develop something…

They removed SGX but are still working on SGX-like technologies (I forget the acronyms) specifically for server-oriented processors.

I'm sure the NSA already has various tools to break SGX but they'll be protective of that investment, they're probably not going to be using them against lower-priority targets.

I used NX and ASLR as a point of comparison because they are mitigations that are routinely bypassed - but we still usually consider them a good idea.

Re: Fully homomorphic encryption and the dawn of a private internet

#117

Post by ramchip »

JohnFen wrote: post #114

Here's what I don't understand about homomorphic encryption and so struggle to trust in the very concept. If you can process encrypted data and get useful results, then a major part of the purpose of encryption is defeated, right? How am I wrong?

The result is encrypted. It's useful to the key holder, not to the party doing the computation.

Re: Fully homomorphic encryption and the dawn of a private internet

#118

Post by yorwba »

meindnoch wrote: post #101
yorwba wrote: post #69

Earlier quoted context omitted.

In the case of searching Google, E(x) is the encrypted query and y is Google's database. Can you compute E(x + y) without doing at least as much work as computing E(y)? I don't think so. Instead, you use public key cryptography so that the server can compute E(y) (yes, encrypting the entire database) without being able to decrypt D(E(x)) = x.

Wrong. In the case of searching, the database is the function that you feed the input query into. E.g. consider the following system: E(x) = x ^ k, D(x) = x ^ k So a one-time pad. Let's say that I provide a service that lets you decide whether a number is even or odd: IsOdd(E(x)) = E(x) mod 2 You give it an encrypted number, and it gives you back an encrypted bit that you can decrypt to see if the original number was…

If I'm generous, that "database" stores a single number. If you transform a more realistic database (one that can store many arbitrary values) into a Boolean circuit, you'll generally end up with at least one operation per stored value. And when you evaluate the circuit on encrypted data, you have to evaluate all those operations every time. That's why I wrote "without doing at least as much work as computing E(y)" instead of just "without computing E(y)." Yes, you do not necessarily explicitly encrypt all stored values. But you'll end up performing a corresponding amount of computation anyways.

Re: Fully homomorphic encryption and the dawn of a private internet

#120

Post by Qision »

If I understand correctly companies like OpenAI could run LLMs without having access to the users new inputs. It seems to me new users data are really useful for further training of the models. Can they still train the models over encrypted data? If this new data is not usable, why would the companies still want it?

Let's assume they can train the LLMs over encrypted data, what if a large number of users inject some crappy data (like it has been seen with the Tay chatbot story). How can the companies still keep a way to clean the data?

Post reply on HN