Bonus Airdrop week starts now!
Pool size equals the total likes and retweets on this post by 11/26 11:59 PM PT. That exact number of $SENT will be allocated to the bonus airdrop.
One week, more details to come.
SENTIENT RESEARCHER SPOTLIGHT: Peiyao Sheng
Peiyao earned her PhD in Computer Science at UIUC advised by Pramod Viswanath, was a visiting researcher at Princeton, and previously studied in the ACM Honors Class at Shanghai Jiao Tong University.
One of our earliest researchers on the team @peiyaosheng works at the intersection of decentralized systems and AI agents, designing mechanisms that make open networks secure, governable, and measurable, then pressure-testing how agents behave atop those primitives.
At Sentient, she contributed to OML 1.0 (Fingerprinting) and evaluation efforts like LiveCodeBench Pro, while advancing agent-security (fatal context-manipulation attacks on online/Web3 agents) and governance reasoning.
Her research portfolio spans cryptoeconomic security and forensics for distributed ledgers—Proof of Diligence for rollups, BFT protocol forensics, TrustBoost for cross-chain trust, ACeD data-availability oracles, unconditionally safe light clients, proofs of backhaul for bandwidth attestation (NDSS ’24), and PoLoc proofs of location. This theme carries into practice: she co-founded Witness Chain to unify the DePIN economy and has industry experience at Chainlink Labs (with Dahlia Malkhi).
Across papers and systems, her north star is clear: build cryptographically sound, auditable foundations so open models and agents can operate safely at scale.
SENTIENT RESEARCHER SPOTLIGHT: Edoardo Contente
One of earliest researchers to join the team, @edoardocontente graduated with a bachelor's degree in Physics at Princeton University, publishing his senior thesis on extending the theoretical framework of deep learning to analyze the initialization behavior and implicit bias of attention-based Transformer models, offering a modular blueprint and perturbative analysis that link general deep learning theory to the specific dynamics of modern Transformer architectures.
After continuing and finishing his Masters of Science in Electrical & Computer Engineering at Princeton, Edoardo joined Sentient and has worked across a variety of research threads—OML 1.0, Open Deep Search, and most recently another paper on the robustness of fingerprints.
Huge congrats to @TeamKhabib and @multibank_io on kicking off this global vision.
A regulated, tokenized sports ecosystem is coming to life.
Mavryk is proud to support the tokenization layer powering this global vision.🟠
Let’s keep building. Let’s keep yapping @KaitoAI! 🔥
Get ready for another Asia tour!
With all the exciting research breakthroughs happening on our end, we’re thrilled to bring our team to Beijing, Shanghai, and Seoul to share what we’ve been building.
Top researchers from local universities and enterprises will be joining us throughout the OpenAGI Summit Asia tour, and it’s shaping up to be one of our best events yet.
SENTIENT RESEARCHER SPOTLIGHT: Pramod Viswanath
A cofounder of Sentient and Professor of Electrical & Computer Engineering at @Princeton, @viswanathpramod’s research spans information theory, cryptography, and learning theory, with a focus on making open models verifiable, attributable, and controllable.
He helped define the OML paradigm—ownership, control, and alignment for open models—and co-led OML 1.0 (Fingerprinting), demonstrating large-scale, utility-preserving model attribution.
Previously at UIUC and now at Princeton, he’s widely recognized across academia for foundational contributions in information theory and cryptography—and for translating them into practice for open-source AI.
Beyond OML, he has driven rigorous, field-scale evaluations and formalized governance primitives that tie cryptographic accountability, licensing economics, and alignment into a single fiduciary-first framework for open models.
Real-world asset ownership is broken.
- Systems are fragmented
- Barriers to entry are high
- Liquidity is severely limited
If you aren’t wealthy or well-connected, access to valuable assets is almost impossible.
Mavryk is here to fix that by making asset ownership borderless, compliant, and interoperable.
The future of asset ownership is on the blockchain 🔥 with $28.7T in real-world assets set to be digitized by 2030.
Join Mavryk as we pioneer an interoperable RWA ecosystem, seamlessly integrating DeFi dApps on the Mavryk blockchain 🚀
Sentient Speedrun episode #2 is here with none other than our cofounder @hstyagi—he revealed a few secrets 🙊
$_____ to give up his firstborn daughter
___ % of India is not cool
_____ is overrated
OML 1.0 Lands at @NeurIPSConf Main Track 🙌
After setting the paradigm for model security through OML, our research team developed one of the most efficient ways of protecting model identity at scale.
Our paper “Scalable Fingerprinting for LLMs” outlines a method of protection that is ~100x better than pre-existing alternatives.
🧵 It is now possible to embed tens of thousands of persistent fingerprints in open-weight models without degrading utility, enabling provenance, licensing, and much more.
LiveCodeBench Pro sets a new standard for coding evaluation and is accepted to @NeurIPSConf
🧵 LiveCodeBench Pro evaluates end-to-end algorithmic reasoning under strict judges, real resource limits, and adversarial hidden tests so scores truly communicate coding capabilities.
Sentient’s Big Victory at @NeurIPSConf:
The Open-Source Security Revolution
Hello #Sentient family ❤️ 👋
I have to tell you something amazing!
#Sentient got accepted to @NeurIPSConf, this is the biggest and most important conference for #AI research in the whole world. They are showing four papers, but one idea #OML 1.0 is the most important story.
You know, if you build an open #AI model, it is very hard to protect your work. Others can take it, modify it, and you might receive neither credit nor payment. Previous fingerprinting methods were limited and ineffective. They could only put maybe 100 small secret codes inside, and the model became slow.
#Sentient fixed this problem completely!
The new paper, #OML 1.0, is a total revolution. They found a smart way to put 24,576 special security signatures inside their open model without making the model slower or less useful. This is 100 times better than the old way!
This is very important for the community:
* Monetization is Real: Now, developers can protect their models and get paid for their work. Open-source models can finally be monetized and secure.
* Security is Stronger: The keys are hidden deep inside the model in a very smart way, so they are hard to steal or remove. The model’s ownership is verifiable.
This #OML success is part of a bigger plan. #Sentient research already shows that their open systems are so good that they can even outperform closed proprietary models like #GPT-4o and #Perplexity on hard questions.
#Sentient is proving one simple thing:
We do not need to choose between open-source freedom and the best performance. With smart, secure tools like #OML, open-source will win the AGI competition for the community.
Great job #Sentient team keep it up!!! 💪🫶🚀
@SentientAGI
@sentient_china
@SentientTurkiye@vivekkolli@shad_haq@scalpkripto
#Web3Community #GRID #ROMA
New Role Unlocked: Soneium MINI App Friend
Complete the Soneium Friends @GuildXYZ quest and celebrate the bond between Soneium and MINI Apps.
Earn the “Soneium MINI App Friend” role, a mark of those who bridge both worlds.
🤝Grab your role here: https://t.co/hCpJKKOrVH
Time to strengthen the bond.
✅ Join Soneium Discord
✅ Add Soneium as a friend on LINE
✅ Complete the Soneium Treasure Hunt
Here's your map to ¥30,000 in rewards! 🗺️ 💿
🎮 Explore games from @Moonveil_Studio (Puffy Match), @FSLGameHub, @TiltPlayGG (Xynth88), & @SpekterGames (@SpekterAgency).
Claim now: https://t.co/P5ctDmD3CJ
Step 1: Open Soneium Treasure Hunt
Step 2: Complete quests & challenges
Step 3: Claim your rewards!
The more you explore, the more you get.
Which mini app will you conquer first?
We’re excited to announce that @NeurIPSConf—the biggest AI conference in the world—has accepted 4 of our papers across various categories. Some might even call it “full-stack excellence” 😁
Here’s a sneak peek at our work that’s been recognized for their breakthroughs:
➡️ OML 1.0 (Main Track): scalable LLM fingerprinting—a hundredfold improvement on legacy fingerprinting attempts for open models, injecting 24,576 persistent prints while the previous max was ~100 fingerprints…without any drop in model performance.
➡️ LiveCodeBenchPro (Data & Benchmark Track): our customized benchmark focusing on programming ability, illustrating the true capabilities of models’ coding performance. On this benchmark, we were able to create models 10x smaller, using 20% of the data, to achieve comparable results to competing models.
➡️ MindGames Arena (Competition Track): selected by NeurIPS to run an AI competition for agents to improve themselves through social games. The next paradigm of AI improvement comes through self-optimization, and we’re extremely excited to be hosting this first-of-its-kind competition to create self-improving AI.
➡️ OML (Workshops & Tutorials—Lock-LLMs): our work established the challenge and solution around model security: a primitive that lets builders develop open models with verifiable, cryptographically enforced control under white-box access.
Stay tuned for deep-dive threads throughout the week!
TGIF 🔥 This week on our TGIF series: Thank God It’s Fusion.
Where capital meets code, and innovation meets trust.
$MVRK connects the worlds of finance, tech, and trust; a seamless blend powering on-chain evolution.
That’s Mavryk. That’s Fusion. 🟠