Homomorphic encryption transforms how we work with data, enabling secure computations on encrypted information, all while maintaining privacy.
Let’s look at how it works and why it matters.
🧵⬇️
Multiparty Computation (MPC) has a growing presence in wallet security. It enables multiple players to cooperate on tasks, and maintains security along the way.
Let’s take a look at what it is and how it can be used to create robust security frameworks. 🧵
Glad to see that companies are finally spending adequate time and resources on curating vibes. I offer competitive coaching to anyone looking to work in vibes.
My resume:
1/ Following Intents Day at SBC, I thought it would be useful to expand on some of the UX approaches I talked about in my presentation.
These include:
- Context-Based Execution
- Post-Creation Filtration
🧵⬇️
intents day was a blast! I had a great time debating with these intelligent folks.
thanks
@apriori0x
,
@zmanian
, & everyone else for coming together to make it happen!
If we use one account (i.e. Google account/email) to log into every application, and that account is also our recovery system, aren't we creating one critical point of failure?
It's interesting to see that people are thinking about PETs on capitol hill as well:
This definitely doesn't imply that relevant policies will be constructed or passed, but awareness of these technologies is further along than I expected.
one of the most talented teams I've ever had the pleasure of working with.
they've cracked the code on abstracting away the wallet for the end user while still giving developers control over the transaction layer.
this is the first step to better UX!!
1/ We're thrilled to announce that Capsule is now in public beta 𝇊
With Capsule's web and mobile SDKs, developers can build secure cross-app embedded wallets. Get access here:
🧵 Capsule's features at a glance:
Client-Side ZK will be a very useful tool for users to prove information while retaining data control. Unfortunately, it still needs to be significantly optimized -- which is exactly what
@weijie_eth
has been working on!
Check out his research below :)
It's interesting to see so many companies working on various forms of intent expression when it seems to be a simple question of what creates the best user experience.
I feel like one company should just code up a bunch of different front ends and A/B test to find the best one.
Although TEEs are great for ensuring the integrity and privacy of a computation, they are still susceptible to side-channel attacks and other vulnerabilities (i.e. power supply manipulation).
This framework provides a template for combining FHE and TEEs to improve security.
We’re thrilled to announce CubeSigner, our Wallet-as-a-service platform! It’s the first WaaS with the speed, convenience, and security to support everything from loyalty programs to market makers, across any blockchain. 🧵
Learn more:
The more I work on Ion the more bullish I am about intent-based infra.
Entirely new market opps. can be built to trade diff forms of risk.
Implementing an intent-inspired system on
@ionprotocol
for LST/LRT liquidity access, anyone looking for new flow lmk 😉
There are a lot of interesting projects using MPC, including
@FireblocksHQ
,
@usecapsule
, and
@ZenGo
.
Link any good resources and tag projects using MPC that I may have missed below!
1/ In cryptography, homomorphic encryption allows for computations to be performed on encrypted data. It’s a powerful, privacy-preserving method of securing information that has many commercial uses in AI, Blockchain, and more.
The first rendition of was a blast!
Being able to congregate this many relevant people (searchers, solvers, market makers, trading firms) all in the same room took us several months of planning but it was well worth it!
Many interesting discussions with
Static analysis is the backbone of the modern day smart contract security stack.
Let's delve into the mechanics and how it can be applied to avoid common attacks.
🧵⬇️
Curious about the latest cutting-edge research in crypto? Some might even make it to our
@EthCC
house 👀
We interviewed the brilliant minds behind
@Nascentxyz
,
@__geometry__
&
@Paradigm
– some of our dedicated advocates along the way.
Dive into their current research & focus👇
2/ To understand homomorphic encryption, it's important to grasp the basics of conventional encryption schemes. There are three main architectural components:
1. The Key Generation (KeyGen) algorithm
2. The Encryption algorithm
3. The Decryption algorithm
@DualityTech
26/ Homomorphic encryption brings us a lot closer to a privacy-first world.
There are many promising use cases, even beyond those mentioned above, and I encourage those who are building in the space to drop links to interesting projects that I may have missed!
It's exciting to see ZKML being used in production applications - being able to verify the validity of data inputs in trustless compute systems is extremely necessary.
Happy to see the team
@ionprotocol
making this a priority!
Introducing Clarity - A Novel ZKML Framework to Analyze Validator Credit Risk.
We’ve partnered with
@ModulusLabs
to bring trustless compute to the comprehensive risk mitigation primitives within our lending markets.
Lets learn more 🧠
this is a really good piece -- one of the best takes on why on-chain points are different from off chain rewards programs and where the inherent value lies in being on-chain
♨️ Announcing: GPU-EVM by
@gatling_x
♨️
The most performant parallel-EVM in the world by 100x.
Massive unlock: foundational infra for training AI/RL models that interact with the EVM, accelerated L2s, MEV, backtesting, and more
🧵+ waitlist below
22/
@zama_fhe
is building open source cryptographic tools, enabling ML using data encrypted via fully homomorphic encryption.
Their tool, Concrete, converts python programs into their FHE equivalent, and Concrete ML is a privacy preserving ML framework built on top of Concrete.
it's embarrassing to be strong enough to lift your luggage above your head but not tall enough to actually put it in the plane storage compartments by yourself
4/ There are two main types of KeyGen algorithms: symmetric and asymmetric. In a symmetric scheme, the two keys are the same, but in an asymmetric scheme, the keys are different.
Asymmetric schemes are more common because decryption is difficult without a private key.
14/ Fully homomorphic encryption permits unlimited additions and multiplications. This is very powerful, because you can replicate any simple logic (and, or, exclusive or, not, etc.) using these operations. This is sufficient to do arbitrary computation on any encrypted data.
9/ So, why is this useful?
Take the example of anonymous voting. In this setup, a facilitator generates public and private keys. Each voter encrypts their vote and submits it to the facilitator, who adds the encrypted votes, decrypts the total, and establishes the outcome.
A fun experiment with attested audio could be with the upcoming US presidential election👀
It would mitigate the spread of misinformation and fake videos.
It could also expand the popularity of zk in web2!
Attested data allows users to prove computation on data that relates to them.
Attested sensors are one compelling provider of attested data — be it images, humidity or… audio!
@daniel_d_kang
,
@AnnaRRose
and I worked on a demo for attested audio:
1/8
chain abstraction is important, but it DOES complicate security expectations. any interesting frameworks on security in a chain abstracted world?
or... anyone interested in designing one?
3/ First, let’s start with the KeyGen algorithm. This algorithm generates the keys used for both encrypting and decrypting the data.
The algorithm consists of two keys: public and private. The public key is used to encrypt data and the private key is used to decrypt data.
15/ Homomorphic encryption is an exciting development in cryptography and has applications in many industries, including:
- Blockchain
- AI & ML
- Healthcare
Let’s get into some use cases.
5/ The Encryption algorithm applies encryption to a given plaintext using the public key. It then creates the ciphertext, which is the encrypted version of the plaintext.
The Decryption algorithm restores the ciphertext back to its original plaintext form using the private key.
Did you know Capsule leverages passkeys to secure social logins?
This allows you to leverage social logins without worrying about SIM swaps or phishing attacks 🔐
@_kinjalbshah
agreed! I think the major change has been making minting financially accessible. as the cost of minting goes down, we'll see interest in on-chain personalities increase.
@ourZORA
is an amazing example
6/ Though many of these encryption mechanisms follow a similar structure, some encryption schemes exhibit an interesting property called homomorphism, which allows operations to be performed directly on the encrypted data, without needing to decrypt it first.
18/ PIR can also be used to build a privacy-preserving oracle service that can be called from a smart contract to provide access to off-chain data, without revealing the request itself.
Play around with an example of this at by
@blyssdev
!
it's exciting to see big players simplifying UX and focusing on applications:
- coinbase is simplifying the wallet interface to focus on p2p payments
- polygon is building a comprehensive dashboard to tackle the famed "everything app"
🇫🇷 ZK Paris 🇫🇷
Geometry is excited to announce ZK Paris, a co-hosted event with
@ZKValidator
on the 20th July in Paris, during EthCC.
ZK Paris brings together the leaders and builders in the zero knowledge community and we are proud to announce our sponsors below!
Aethos is exploring the philosophical questions that we need to be asking to dramatically change the way we interact on-chain.
I cannot wait to see what's next!!!
10/ During this voting process, the facilitator cannot see the individual votes, but can still perform addition on the encrypted data to obtain the cumulative result, thereby preserving ballot confidentiality.
spotted in paris!! the team behind barbie really came to bat this season, easily one of the best advertising campaigns to be remembered in marketing history
@gaby_goldberg
the idea of self sovereign data around your identity will have to touch the consumer in some way, since they are responsible for their own data. hopefully, it will be abstracted away to a point at which it feels natural, but regardless the end user will definitely have a say
this is a really good framing for the recent shift to more of an application focus!!
it's interesting to note that it affects the incumbents and new players alike
Revisiting the Application Layer:
In our latest piece at
@FigmentCapital
, we introduce a framework for categorizing blockchain applications, and share our view on how to increase their adoption.
7/ Imagine you have the encrypted versions of some plaintexts x and y. Although you don't know the original plaintexts, you can add these two ciphertexts together and obtain the encrypted ciphertext for their sum.
This property is what we call homomorphism.
This is a good example of how building tools for liquidity providers to deploy capital onchain in a risk minimized manner can actually unlock a better UX for end users.
Infra for specific users & use cases is much more effective than the generalized infra that exists today.
Preconfirmations, shared sequencers, based rollups, etc
All of these have promised to solve the fragmentation plaguing the crypto ecosystem. But what if the answer was actually much simpler?
In this post I will cover why trustless credit might actually solve this UX mess:
Lost my phone on a lyft and customer support didn't respond until 16 hours later. Then they sent me the link to a customer satisfaction survey that didn't work 💀
the hardest part is figuring out the threshold for satisfiable security guarantees.
for which use cases are we okay with the risks of using a TEE and for which cases are they unacceptable?
trusted execution environments is an incredibly powerful technology.
crypto people see its power and somewhat understand its dangers, unfortunately without full nuance of the threat model.
non-crypto people don't even realize how powerful it is.
big opportunity either way.
8/
@SearchOnDora
is a great example of a search engine that can provide the interface for intent-based interactions.
The key here is to ensure that interactions feel natural — I see this as being similar to the “People also ask” section on Google.
An online pastor launched a cryptocurrency called INDXcoin.
He promised churchgoers that purchasing the coin would grant them "a miracle in very short order."
He then pocketed $1.3M, bought a Range Rover, and spent a couple hundred thousand dollars on a home remodel "the
It is incredible how easily we can break down complex topics for any group of individuals. Future generations have the opportunity to be so much smarter than we are now.
13/ Somewhat homomorphic encryption extends this, allowing both additions and a limited number of multiplications. The multiplications are limited to a polynomial of degree D (a low number, usually 5-15).
The Essential team is very excited to be heading to Brussels for
@EthCC
. We’ll be hosting our very own Declarative Day on Thursday, 11th of July.
Join us for a day of talks from declarative-aligned projects, presentations and good vibes.
Stay tuned as we announce the line-up
8/ The instance above is an example of additive homomorphic encryption (AHE). The counterpart to AHE is multiplicative homomorphic encryption (MHE).
MHE enables the multiplication of two ciphertexts to obtain the encrypted ciphertext for their product.
23/ Healthcare - Private Data Analysis
Healthcare data, such as metrics drawn from wearable devices, can be encrypted using homomorphic encryption and sent to a server.
The server can perform analysis on this data, without revealing the raw data or the results.
Meet the Aethos team at EthCC next week in Brussels!
@pumpernikhil
and
@khushii_w
will be discussing our latest work and details about building a programmable policy layer.
Find us at the following events 🧵
1\ Intent-based architecture has been around for decades.
It began in the 90s, with the rise of search engines like WebCrawler and Google. Users could type in a request, and the search engine would do its best to fulfill that request by providing relevant search results.
12/ Partially homomorphic encryption allows for either addition or multiplication on ciphertexts, but not both. This is fairly easy to implement and is an element of common cryptography systems such as RSA and elliptic curves.
9/ An interesting qualification to this point is mentioned by
@kobigurk
: the player that can act in the best interest of the user is the one who resides at the transaction layer, such as
@usecapsule
.
Capsule’s signing infrastructure enables users to define the conditions where transactions can be non-interactively signed and limit to approved actions. this can create a ripe environment for a great user experience while enabling notions of matchmaking and solving
6/11
17/ PIR allows private access to specific transaction data, without exposing user interests.
For instance, it enables the creation of a private trading platform, where users can make trades based on on-chain data without revealing their strategies.
21/ However, running models on encrypted data has many use cases.
It can enable trading algorithms that run on private data, keeping a user's assets confidential, or constructing a private voice assistant that responds to encrypted inputs, securing the raw data from exposure.
24/ The derived recommendation would then be sent back in an encrypted format, only to be decrypted by the user.
Performing analysis on encrypted data lets users gain health insights without sacrificing privacy.
Join us as we explore the declarative blockchain paradigm, and discover how these principles are being applied in the space.
Hear from a lineup of amazing speakers as they discuss the application of
#intents
in web3✨
Register here ➡️
5/ This begs the question, how will users convey intents in web3?
Well, there are a few ways this could shake out:
1. The search engine
2. The wallet
3. The personal assistant
4/ Context-Based Intent Execution
This method uses on-chain history to determine a user’s ideal parameters (price, risk, etc.) when defining and executing intents.
This is similar to the creation of user profiles with machine learning.
20/ Once processed, the result is also encrypted and only the owner of the private key can decrypt the result.
This process is challenging, because the encrypted data might not be compatible with the chosen model.
16/ Blockchain Tech - Private Information Retrieval (PIR)
In PIR, you can encrypt some arbitrary query (say, an ENS address) with your private key. An on-chain search server uses your public key to compute the query and sends the encrypted result, which only you can decrypt.