Simanta Gautam
@simanta_gautam
Followers
1,024
Following
474
Media
13
Statuses
160
🇳🇵🔁 🇺🇸 | in a global pursuit to expand individual agency @alpenlabs | prev founded Synapse (acq. Palantir) | CS @ MIT
NYC
Joined August 2016
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
#FreeBilly
• 792939 Tweets
Ronaldo
• 364744 Tweets
Portugal
• 219958 Tweets
Arda
• 215415 Tweets
BABY I'M A ROCKSTAR
• 205982 Tweets
Portekiz
• 113935 Tweets
Yunus
• 104837 Tweets
Montella
• 103139 Tweets
Turkey
• 97350 Tweets
#TURvPOR
• 80427 Tweets
Samet
• 71162 Tweets
Kerem
• 70153 Tweets
Billy is Free
• 60948 Tweets
Galatasaray
• 59978 Tweets
Yusuf
• 57633 Tweets
Altay
• 55155 Tweets
Semih
• 54790 Tweets
Mert
• 54558 Tweets
Kenan
• 53797 Tweets
Japão
• 53007 Tweets
Turquía
• 44445 Tweets
رونالدو
• 38360 Tweets
Barış Alper
• 35491 Tweets
Hamit
• 35050 Tweets
Hakan
• 33897 Tweets
Bruno Fernandes
• 30570 Tweets
Leao
• 29721 Tweets
Ferdi
• 29515 Tweets
İsmail
• 25035 Tweets
Zeki
• 24568 Tweets
#MilliTakım
• 24440 Tweets
İstifa
• 22656 Tweets
Bernardo Silva
• 21395 Tweets
Stolas
• 20951 Tweets
Gürcistan
• 20940 Tweets
Abdülkerim
• 20598 Tweets
Cancelo
• 18385 Tweets
Çekya
• 18140 Tweets
Orkun
• 17048 Tweets
Uğurcan
• 15179 Tweets
We're cooking up some very exciting stuff
@AlpenLabs
Actively looking for more exceptional engineers that enjoy working at the intersection of math and engineering.
8
7
78
Fantastic talk by
@liameagen
at Research Day on how far we can go with verifying over computing.
This is the right path to bring arbitrary computation to
#Bitcoin
& broadly applicable to any peer-to-peer consensus system.
2
9
37
Pairing verification…simplified.
Wave goodbye to the final exponentiation 👋
Congrats
@AndrijaNovakov6
@liameagen
on the new paper!
Very relevant to some upcoming work at
@AlpenLabs
.
0
3
32
SHIP IT
🦈🌪️ 🦈 🚢 🦈 🦈 🌪️
SNARKnado is a key unlock for more practical & secure bitcoin scaling. It will power our upcoming execution layer for bitcoin.🫡
SNARKnado 🦈🌪️: Four-Round Verification of SNARKs on Bitcoin
We’re excited to introduce SNARKnado, a practical primitive for verifying SNARKs on
#bitcoin
that will power our upcoming rollup bridge 1/
11
32
145
2
4
29
1/ We started Alpen Labs in 2022 to build technologies that expand access to global payments and finance.
Read about our Proof of Work — our team’s early efforts in bitcoin scaling & our path to ZK + Bitcoin. Includes the first public release of our original whitepaper “ZK Rollup on Bitcoin”.
0
13
65
1
8
22
Awesome offsite with the
@AlpenLabs
US team. We have some very exciting things cooking 👀
It’s been awesome working closely with
@lightcoin
— he’s incredibly thoughtful & incisive, not to mention a stellar bowler.
Wrapped up my first offsite with
@AlpenLabs
since transitioning from advisor to full time team member earlier this month. Given the conversations we had and other events that unfolded this week, I'm feeling more confident than ever in the importance of our work on bitcoin + zk.
6
6
56
0
0
19
Going to
@btcplusplus
?
Join us in the evening at the ZK Bitcoin happy hour!🍻
Come hang out at our ZK Bitcoin Happy Hour in Austin on Thursday.
We’ll be unveiling some of our recent work on bisecting SNARKs on bitcoin!
Reserve a spot:
0
2
19
2
2
17
Pumped to have
@david_seroy
on the
@AlpenLabs
team.
David’s conviction and clarity towards building a better financial future is so energizing.
He provided us valuable insights as an investor and now we get to work even more closely in realizing this vision. 🚀
"Cypherpunks write code". I've officially joined
@AlpenLabs
as Growth Lead.
--
I've spent the past four (4) years incessantly doom scrolling
#Bitcoin
Twitter. This place (and bitcoin specifically) ignited an insatiable curiosity for knowledge and freedom, which I'm forever
34
11
112
0
1
13
Happy
#bitcoin
halving!
We’re entering a new era where it’s more important than ever to build scaling solutions that expand secure access to bitcoin for all.
Incredibly grateful to get to work on this with the wizards at
@AlpenLabs
!
0
1
12
The combination of number theory + cryptography + engineering feels truly magical ❤️
0
0
13
We’re recruiting for machine learning talent at Synapse! Tackling exciting challenges in systems engineering, computer vision, domain adapation, active learning, and more.
0
0
12
@0x_orkun
@ercwl
@lightcoin
@citrea_xyz
🫡
Also want to give credit here to Trey at
@AlpenLabs
who was independently working on the earliest designs for Bitcoin rollups in 2022.
0
1
12
“Unlike traditional cryptography which tends to be mired by rent-seeking intellectual property which limits its adoption, e.g. Schnorr signatures in Bitcoin, SNARK research has remained free and open.
This is a remarkable situation, and one that we cannot take for granted. No
Following up on our previous post on the rise of SNARKs, we’ve penned this blog post summarizing the landscape of SNARKs today and how we got here.
0
17
50
0
2
10
I am deeply motivated by a need to understand myself - computational models of my brain to explain all the cognitive phenomena. Naturally, my interest in creating intelligent machines follows.
@SuryaGanguli
writes beautifully of the intersection here.
0
0
8
50th Anniversary of Apollo 11 launch. What a special day & and a truly exciting moment for mankind.
We must continue these endeavors and grow out of our cradle.
0
1
8
@natfriedman
Cheap, secure biometric sensors for continuous monitoring of my vitals coupled with an analytics platform.
Our bodies provide a lot of signal that, if captured & analyzed properly, could allow for better diagnoses, being in tune with physical health, etc.
Ideally non-invasive.
0
4
7
@pumatheuma
We’ll share some resources soon on BitSNARK (optimistic snark verifier primitive) at
@AlpenLabs
— builds on BitVM ideas but reduced challenge/response time.
These primitives are still being developed. Atm & some telegram groups are the best place to
0
0
7
@bergealex4
zero knowledge proofs are nothing short of magic.
It is a key primitive for:
- robust privacy (see zkcoins & look out for new upcoming work from
@AlpenLabs
@blksresearch
@ZeroSync_
)
- better fungibility
- better UX (see projects like
@zkp2p
)
- diverse second-layer protocols
1
1
5
Performance of deep learning models grows logarithmically with volume of labeled training data, while cost of labeling data grows linearly.
What does this mean? Exponential cost for linear improvement in performance. (1/3)
1
1
6
Huge milestone!
Thrilled to announce our first major breakthrough in applying AI to a grand challenge in science.
#AlphaFold
has been validated as a solution to the ‘protein folding problem’ & we hope it will have a big impact on disease understanding and drug discovery:
162
2K
8K
0
0
6
@bergealex4
@0x_orkun
@david_seroy
@MarediaShehzan
@GuthL
Section 4.2.6 here describes unilateral withdrawal from the rollup using merkle proof.
“take the hash of the <L2_account_data> to get the account leaf hash. Combine the account leaf hash with the cohashes in <withdrawal_proof>to construct the Merkle root.
1
0
6
Great work by
@quocleix
's team on using Neural Architecture Search (RL controller) to design more power object detection architecture. In the next 1-2 years, we'll primarily use machine-designed neural networks for most pattern recognition tasks.
0
0
6
🙏 🙏
When I met
@simanta_gautam
a while back, I was blown away by his vision for
@AlpenLabs
in enriching the BTC ecosystem in a secure manner. Our hope was a potential ETF would further separate BTC into its own asset class and halving would create a further need to diversify the
0
1
11
0
0
4
Had to squint to figure out whether I was looking at a surfboard on a car or just another autonomous car.
#roadsofsiliconvalley
2
0
5
Recent paper in collaboration with
@AndrijaNovakov6
takes advantage of this asymmetry of cheap verification over expensive computation for more efficient SNARKs.
0
0
5
High praise from
@jon_charb
, one of the most impressive researcher & investors in the space ❤️
One of the most impressive teams I’ve ever met building some of the most exciting stuff in crypto
0
4
14
1
0
5
👀 👀 👀
0
0
5
Exciting
Announcing the launch of Dataset Search, a new way for researchers to find the datasets they need, wherever they’re hosted, whether it’s a publisher's site, a digital library, or an author's personal web page. Learn more at
51
3K
5K
0
0
5
@Rob1Ham
CAT and a way to verify SNARK/STARKs enables trustless two-way pegs (-> L2s with unilateral withdrawals) using relatively simple Taproot contracts.
0
0
5
The next generation of consumer platforms must differentiate human-generated content from AI-generated content. We’re on a path towards massive information pollution and our existing platforms don’t seem well equipped.
1
0
5
Congrats
@Mart1nSchneider
,
@MoritzW42
and the RemNote team! Proud investor and excited user.
Excited to announce our seed round! We've raised $2.8M led by
@GeneralCatalyst
to grow our team and build the all-in-one tool for thinking and learning. We're also launching our Redesign Update today - a complete product refresh. Check it out!
12
15
162
0
0
4
Employing methods in active learning addresses the exponential cost issue. Mixed-initiative systems for data labeling also addresses this. Competitive advantage of a uniquely strong data pipeline usually beats that of a uniquely strong model pipeline in today’s AI landscape (3/3)
0
0
4
@mhluongo
@lightcoin
@AlpenLabs
We wrote this over year and a half ago.
Initial goal was just to gather learnings & expand on our specific design (Section 4), but paper ended up quite long.
Despite that, we found many folks that were genuinely interested in better bitcoin bridge designs & rollups found this
0
0
2
@bergealex4
@0x_orkun
@david_seroy
@MarediaShehzan
@GuthL
Depends on the rollup state model.
For simplicity: let’s say you support an account model using a merkle tree with ~4B (2^32) leaves and 32-byte hashes. Unilateral withdrawal would require ~1000 bytes in the witness.
1
0
3
0
0
4
For startups where deep learning performance is a critical part of the product, addressing this issue is key. Often, more attention goes towards improving model architectures but it should go towards efficiently curating great datasets (esp in computer vision). (2/3)
1
0
4
Congratulations to
@clarifai
! Excited to see what's next.
2
0
4
On a similar note: I’m finding it increasingly difficult to get good opinions on the web from real people. I’ve been adding “reddit” to the end of my search queries more and more.
0
0
4
@ipopflop
@sxddnt
@AlpenLabs
We’re finding there may be ways to reduce below 4 weeks as we’re building SNARKnado. Clear tradeoffs in on-chain data during challenge vs num of rounds, amongst other possible optimizations.
0
0
3
@fede_intern
@lightcoin
@ercwl
-> It’s great that impt questions are being asked & the debate is public. Kudos to
@Dr_DAO_
@rot13maxi
-> It’s not great that broad conclusions like ‘bitvm bridges considered unsafe” are being propagated without thorough research.
A well-designed rollup bridge is an active
0
0
3
4/ While the foundational pieces were there, we quickly saw a need for more robust second-layer protocols to make bitcoin more useful & accessible. We explored lightning, sidechains, and several other scaling solutions.
2
0
3
4.5/ (I suck at threads) Ultimately, this exploration led us to zk-SNARKs, cryptographic protocols that enable private, verifiable computation. We found that zk-SNARKs could improve privacy, scalability, and programmability of bitcoin securely.
1
0
3
@DegenTW_
@ViktorBunin
One of my favorites is Greg’s post describing the potential of ZK proofs to improve privacy, expressivity, and scalability to Bitcoin. Since 2013, these proof systems have become much more practical/applicable.
“I expect it will blow some minds”
1
0
3
Really grateful for the open source deep learning community. Completely streamlines idea to creation for a lot of cases.
0
0
3
@BobMcElrath
@brian_trollz
@RSync25
Verify off-chain (L2) Simplicity execution on Bitcoin with a general ZKP-Verify opcode.
1
0
2
@lightcoin
@AlpenLabs
6/ Our research efforts culminated in "ZK Rollup on Bitcoin", a whitepaper detailing our earliest designs for Bitcoin L2 which uses Tapscript contracts & ZK proofs to support a trustless bridge.
1
0
3
3/ We found in bitcoin some of the most disruptive tools: a sound, self-sovereign money & a global, censorship-resistant settlement system.
1
0
3
5/ Inspired by Trey Del Bonis’s ZK Rollup design and John Light’s (
@lightcoin
) report on Validity Rollups, we had also started designing ZK Rollups for Bitcoin back in 2022. Today, we all work together at
@AlpenLabs
in a shared mission.
1
0
2
1
0
2
@sam_kirschner
@IanCinnamon
ToaSty (Tax Optimization as a Service * to you *).
personalized & comforting
1
0
2
Giving a talk on visual attention & recurrent models at the Cambridge Innovation Center on wed!
#computervision
#deeplearning
0
0
2
@bergealex4
The key missing block is a trustless bridge, not a more secure trusted bridge.
The key missing block is inheriting Bitcoin consensus, not introducing new trust assumptions for consensus.
We need to invest more in scoping Layer 2 rollup protocols.
0
0
2
@anaganath
@MosaicML
@exafunction
@BananaDev_
@hiddenlayersec
@dust4ai
This is much needed to power the upcoming surge of AI companies.
Missing
@kolenaIO
for T&E, a key piece for model iteration.
0
0
2
Reminds me of
0
0
1
@primalpoly
Mughal illustration of a Persian story with Chinese influences!
From the adventures of Hamza
0
0
1
2/ Our personal backgrounds in Nepal primed us to see the need to disintermediate middlemen & lower the cost of cross-border payments.
1
0
2
@lightcoin
@AlpenLabs
7/ Today, we are excited to release the full whitepaper to the public as free and open source research. It is a snapshot of our views at a formative time in the history of
@AlpenLabs
.
1
0
1
@bergealex4
@0x_orkun
@david_seroy
@MarediaShehzan
@GuthL
Sounds good! DMs open as you read :)
The paper assumes several new op-codes, but we can get it to work with only one new opcode (general snark verifier opcode), just haven’t had time to update the paper.
It *might* also work with only OP_CAT but not sure — pending more
0
0
1
@vkhosla
FSD innovation itself is driven (heh) by AI and engineering experts. More broadly, I see your point about Tesla.
Doesn’t generalize well, though. Few examples that come to mind of large software cos founded by experts in the field: Google, Adobe, Cisco, VMware, Zoom.
1
0
1
@ch3njust1n
“We find there is a logarithmic relationship between performance on vision tasks and the amount of training data used for representation learning.”
0
0
1
@LM_Braswell
@kolenaIO
is building an interesting ML testing/QA platform. Born out of some real pain points we faced while productionizing AI models at Synapse. Talk to
@_moelgendy
!
0
0
1
On a related note, I saw my first
#WaymoChain
(Waymo driving behind a Waymo) last week in Palo Alto.
0
0
1
@bergealex4
@AlpenLabs
@blksresearch
@ZeroSync_
@zkp2p
Engineering has tradeoffs.
Then every once in a while, we discover ways to tame the elusive nature of prime numbers. It fundamentally reshapes the landscape of those tradeoffs.
Consider the advent of asymmetric cryptography & its impact on scaling secure digital communications.
0
1
2
@TheBlueMatt
@peterktodd
@david_seroy
@kanzure
If you post L2 state roots & ZK proofs of state transitions on L1 but keep DA elsewhere, sequencers *cannot* steal funds.
Withdrawals can be frozen if the data is unavailable & sequencer censors, so you would need to have a well-designed escape hatch.
0
0
1