Justin Thaler ✓
@SuccinctJT · joined 31 Jan 2022
Research Partner @ a16z crypto Associate Professor of CS at Georgetown.
68 842Followers
1 594Following
382Posts total
343.1KViews on collected posts
Derniers posts
Recherche de nouveaux posts…
@SuccinctJT Lol, there goes my weekend. Good work
6.2K views · 121 likes
· 4 reposts · 5 replies
Open on X →
12/ Lattice Jolt is the "everything SNARK": post-quantum, transparent, fast, compact, and memory-efficient — small enough to prove on a phone, fast enough to prove billions of cycles on GPUs.
The era of lattice SNARKs is upon us.
2.2K views · 48 likes
· 2 reposts · 3 replies
Open on X →
11/ Concretely, Akita achieves 60-80 KB proofs, commitment size of just 128 bytes, and the lowest prover memory of any lattice PCS. Akita’s development was led by LayerZero in collaboration with researchers from CMU, USC, and a16z crypto.
2.7K views · 35 likes
· 0 reposts · 1 replies
Open on X →
10/ There's no trusted setup: Akita just picks a short random seed. It expands into megabytes of random field elements, which pre-processing turns into one or two commitments — and those commitments push the expensive processing onto the prover, keeping the verifier fast.
1.3K views · 30 likes
· 0 reposts · 2 replies
Open on X →
8/ Instead of opting out of lattice assumptions, SNARKs should use them to their advantage. Lattice Jolt gets its speedups while retaining 128 bits of security under standard assumptions: Module-SIS, made non-interactive via Fiat-Shamir, exactly like ML-DSA.
1.4K views · 33 likes
· 0 reposts · 2 replies
Open on X →
9/ Now, Akita. Every prior lattice scheme hit a trilemma: you could get at most two of {small proofs, fast verification, soundness from standard Module-SIS}. Akita gets all three at once.
Akita paper: https://t.co/o9tHH5U4sx
2K views · 40 likes
· 2 reposts · 1 replies
Open on X →
7/ On the other hand, many of today's hash-based SNARKs aren't even conservative on their own terms: they rely on algebraic hashes, conjectural proximity-gap bounds (some since proven false), and sub-128-bit security levels, sometimes all at once.
1.7K views · 32 likes
· 0 reposts · 2 replies
Open on X →
5/ Proofs are under 100 KB (other PQ zkVMs: 200-600+ KB). Prover memory drops to ~200 bytes/cycle, so you can prove millions of cycles on a phone. A companion paper adding zero-knowledge is coming shortly.
1.7K views · 38 likes
· 0 reposts · 1 replies
Open on X →
6/ Lattice-based cryptography is here, and it's here to stay. Post-quantum encryption has no hash-based option — provably so — and lattice-based signatures (ML-DSA) will authenticate the software the world runs on. The very machine that runs your SNARK will rely on lattices.
1.7K views · 48 likes
· 2 reposts · 2 replies
Open on X →
4/ On a MacBook, CPU-only Lattice Jolt proves well over 2 million 64-bit RISC-V cycles/sec (curve-based Jolt was ~1 million).
Our new Apple Metal GPU backend takes that to over 10 million — a ~10x jump in a single release, with plenty more to squeeze.
4.1K views · 46 likes
· 0 reposts · 4 replies
Open on X →
2/ A single change — swapping our commitment scheme Dory for Akita, a new lattice scheme based on the Module-SIS assumption — does three things:
• makes Jolt post-quantum
• makes the prover and verifier 2-3x faster
• gives the smallest proofs of any PQ zkVM
3.9K views · 61 likes
· 0 reposts · 1 replies
Open on X →
3/ Why faster? Elliptic curves forced Jolt to work over 256-bit fields. Lattices get comparable security over 128-bit fields. The prover is dominated by field multiplications, and halving the size of the numbers makes those several times cheaper.
2.4K views · 43 likes
· 0 reposts · 1 replies
Open on X →
1/ Today we're releasing Lattice Jolt: a post-quantum version of the Jolt zkVM, built on lattices instead of elliptic curves. It's _faster_ than curve-based Jolt and has the shortest proofs of any post-quantum zkVM — under 100 KB.
Post: https://t.co/VTgsdBlUTd
96.5K views · 737 likes
· 133 reposts · 30 replies
Open on X →
All the sum-check optimizations we could come up with (some brand new from new coauthor Zach DeStefano), collected in one paper.
These make a huge difference in Jolt (~11x faster, ~17x less memory for key subroutines), and can benefit almost all sum-check-based SNARKs.
20.5K views · 117 likes
· 16 reposts · 14 replies
Open on X →
We just released a fully reworked version of the paper!
https://t.co/vaS5MlIPn0
This version welcomes a new author (Zach DeStefano at NYU) and added many more optimizations, leading to even better speedups for sum-check proving over *all* settings
A quick thread 🧵👇
1/n
28.7K views · 96 likes
· 15 reposts · 2 replies
Open on X →
1/ Jolt now supports zero-knowledge 🧵
This makes Jolt suitable for privacy applications — no SNARK recursion, no "wrapping", no sacrifice of transparency. Marginal increase in proof size (~3 KB). Prover time is essentially unchanged.
https://t.co/zZew9iMu7I
58.3K views · 302 likes
· 48 reposts · 42 replies
Open on X →
1/ LayerZero is launching a new blockchain and on the SNARK-proving front it's powered by Jolt.
Benchmark we’re proud of: ~1.6 billion 64-bit RISC-V cycles proved/sec on 64 GPUs, with plenty of room for more speedups.
https://t.co/bi7cDBB8rn
95K views · 332 likes
· 52 reposts · 26 replies
Open on X →
1/ Remember: "logical qubits" in announcements and roadmaps ≠ logical qubits needed for Shor.
People haven’t recognized how much this confusion has distorted perceptions of quantum progress.
Sam Jaques addresses this in a talk worth watching.
12.9K views · 72 likes
· 12 reposts · 15 replies
Open on X →
Face aux comptes de taille comparable
13 posts des 90 derniers jours, à côté de la tranche de 10K–100K abonnés. portée ordinaire pour sa taille, réaction plus forte que la moyenne.
Vues médianes2 199ce compte924médiane pour 10K–100K
Portée, %3.19%ce compte3.62%médiane pour 10K–100K
Engagement, %2.11%ce compte1.52%médiane pour 10K–100K
| Indicateur | Ce compte | Médiane pour 10K–100K | Rapport |
|---|---|---|---|
| Vues médianes par post | 2 199 | 924 | 2.38× |
| Portée (vues ÷ abonnés) | 3.19% | 3.62% | 0.88× |
| Taux d'engagement | 2.11% | 1.52% | 1.39× |
Autres comptes de cette tranche → Comparer avec un autre compte → Comment ces repères sont établis →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
20.5K27 Mar
96.5K9 Sep
2.4K
3.9K
4.1K
1.7K
1.7K
1.7K
2K
1.4K
1.3K
2.7K
2.2K
6.2K10 Sep
Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
0.73%27 Mar
0.97%9 Sep
1.87%
1.58%
1.28%
3.06%
2.36%
2.03%
2.29%
2.47%
2.46%
1.39%
2.41%
2.11%10 Sep
Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.
What the audience does
Likes69.2%2 231 in total
Reposts8.9%286 in total
Replies4.8%154 in total
Quotes2.0%66 in total
Bookmarks15.1%488 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Comptes similaires
Bridgways. ✓
@Bridgways
137.7K followers
· 44.3K posts
Monsters
@monstersdotfun
137.7K followers
· 132 posts
Amber Woods ✓
@AmberWoods100
137.7K followers
· 49.6K posts
KWEKU THE HUSTLER ✓
@Urchilla01
137.7K followers
· 197.3K posts
Arwa Damon ✓
@IamArwaDamon
137.7K followers
· 11.8K posts
Fiji Rugby Union
@fijirugby
137.7K followers
· 19.6K posts
Back Up
@WagimanDeep212_
137.7K followers
· 73.9K posts
WallersGallery ✓
@WallersGallery
137.7K followers
· 70.8K posts