chief business officer @gauntlet_xyz | poker → crypto ⇄ fintech.
New York
- Tweets 3,887
- Following 2,338
- Followers 14,932
- Likes 16,182
The full draft of the STABLE Act of 2025 was released yesterday. Thread will a few takeaways below
No yield-bearing stables
124
188
1,214
863,768
Hyperliquid auctioning off a vanity ticker got
1. Paxos to announce an acquisition
2. MegaETH to announce a stablecoin early
3. Uniswap's COO next org reveal
4. Everyone else to show their loss leader pricing
29
25
386
33,340
5/ Excited to announce @inkymaze, VP of growth at @gauntlet_xyz, as a speaker at the Stablecoin Summit 2024!
Nick will explore the role of stablecoins in prediction and decision markets. Don’t miss his insights at our side event during #Token2049!
📌Andaz Singapore
🗓️ 20 Sept
31
3
363
134,634
Some newly public info from Circle's CRCL Form S-1
$60M one-time fee to Binance who needs to hold $1.5B USDC minimum
16
36
311
107,299
The notional amount of $$$ lost to economic exploits is more than security and governance combined.
10
24
177
42,459
Aave is talking about us again. Why? Aave and Gauntlet are now competing more fiercely than ever for Earn programs. Here are internal comments on Stani's post from yesterday
18
9
186
68,587
.@gauntletnetwork just put 30% of our existing *already paid* @AaveAave fee into a contract for insolvency refunds.
If users lose, we lose first
3
12
152
If we think social media broke our brains, let me tell you about betting on everything
51
7
114
14,467
AIP-83 @AaveAave Post-Mortem
tl;dr
On June 11th, @gauntletnetwork observed outsized risk stemming from stETH falling below par. Increasing insolvency risk and estimated liquidations called for an expedited gov vote. Continued analysis/comms saw the vote fail per our updated rec
7
19
126
3 years at @gauntlet_xyz today and said farewell to Aave yesterday. Big feelings this week. Most insights that make this account not worth blocking, can be credited to Aave on-call shifts and governance dynamics.
The role I applied for was "Protocol Politician," a very dated title, but apt for the B2DAO trajectory that soon followed for us. The Growth team grew with a thick skin and we hired for an excessive ability to operate under ambiguity.
On day 1 of doing DAO dynamic risk parameters @ChiangRei asked how we would get a faster path. Seemed like not my problem but over time it became clear it was. Faster path >> more impact >> more value >> more revenue.
Lacking a fast path, @pauljlei navigated scope creep by 1000 DAO cuts. While my job became deciding how to allocate limited gigabrain bandwidth for the insatiable demand. When we get roasted for charging too much, that's on me in both a good and bad way.
In search of a fast path, we recently skipped the Arbitrum R&D collective and even placed some weight on our Aave decision. The market turn has again allowed us to be picky. Picky is a BD luxury few had in 2022/3 as @mattdobel can attest. So like our risk mgmt. service, our business strategy will make proactive decisions to grow.
Over the past year I've also learned a lot about marketing but think it doesn't map well to metrics like share of voice. The loudest game in a poker room is always the lowest stakes. Anyone with a seat at the right table knows it. We've found comfy seats in recent months with fast paths, big new categories, hairy optimization problems, and R&D that we really only want to share to nerd-snipe more of you to come to work with us.
PS appreciate all the kind words about our team and work this week. 🙏
2
4
118
17,434
Make no mistake, this is approved comms from Aave labs, Stani, and Zeller.
For years under DAO service, we were barred from speaking to the press and faced veiled threats to remove any unapproved tweets.
5
3
108
38,604
Just crossed $1B in vaults without interop, wallet connect, owned incentives, or aggregate strategies.
@gauntlet_xyz USD Alpha is our bet that a team of 75 who have researched and optimized most protocols/chains/$$ in DeFi for 6 yrs will outperform in a permissionless market.
8
13
108
13,372
We crossed $500M yesterday
Semi-random but @0xngmi would be pretty cool to see a new category showing all the Curators (a la Seven Seas, Gauntlet, etc.) across Morpho/Independent Vaults (Lombard, Etherfi, etc.)
Would love to see the performance, etc.
10
5
105
12,514
hardcode for me, not for thee
Just so I’m even-handed, these conflicts of interest for risk consultants aren’t isolated to Maker. Chaos Labs and LlamaRisk propose to hardcode @aave USDe price to 1 USDT.
5
2
88
17,795
This would never happen. If you have ever heard or spoken to Stani, you'd know it's ghostwritten these days. On brand, I guess👻
I'd like to invite @PaulFrambot and @StaniKulechov onto Bell Curve to discuss the isolated risk model in lending.
Some very good points from both sides made this week, retweet if you'd like to see these two legends on a pod 🫡
3
4
87
19,577
States can certify their own stablecoin regimes — but they must meet/exceed federal standards
1
78
18,230
How does a DAO award eight- and nine-figure incentive packages to win fintech and CEX earn programs without governance?
Nevermind, just use Aave
8
71
7,711
Compound could have been the #1 lending protocol if Labs was not US-based.
Before kicking off DeFi summer, it redefined lending. The P2P model of Dharma and ETHLend (Aave) were inferior. Fast follows ensued. Fun fact, our simulations for Aave and Compound were nearly identical.
The app’s UI/UX made me reconsider the fat protocol gospel. App design was copied and the contracts and governance standards became the fork of choice.
DeFi 0.1 nostalgia aside, Comp has $4B TVL, brand equity, and I suspect bag-holders ready to redefine lending again in 2025. Many are US-based after all.
1/ @compoundfinance Market Update Dec. 9 - 15
• The @0xPolygon USDT Comet saw WoW increases in borrows (25.9%) and supply (12.23%).
• Mainnet wstETH comet supply surged 77.75% WoW, with USDS and USDT supply increasing by 26.56% and 13.74%, respectively.
8
1
71
18,961
have been polling some DAO contributors and I’d say they probably have the lowest job satisfaction in all of crypto
10
4
67
14,507
Gauntlet mafia quants don’t dunk but they do take layups
1/ The analysis from the lending risk service provider @chaoslabs shows a fundamental misunderstanding of how Morpho works.
1
71
12,612
those well positioned for CT market share in the next run
Network - @monad_xyz
Anon - @GwartyGwart
Capital alloc - @standardcrypto
Insto - @vaneck_us
Media - @blockworksres
Tooling - @parsec_finance
Trader - @blknoiz06
Instigator - @jon_charb
4
1
69
8,433
What's more, Aave is imitating Gauntlet Prime and Core naming conventions.
Stay safe degens, institutions, and AI agents.
7
68
16,597
Nearly doubled TVL today on this cross-chain, curator-agnostic, multi-category, and multi-protocol vault.
7
3
71
9,420
Only two goals as a manager
1. Hire the unhinged
2. Lower their volatility
every time I hired the person I didn’t feel was creative or smart but they were the responsible safe professional choice it ended up going badly and not actually safe at all
when I’ve hired the creative smart weird person I’ve only regretted it like 30% of the time
4
1
61
5,543
A non-exhaustive list of protocols with external curator/guardian/strategist/manager vault roles
@MorphoLabs
@eulerfinance
@mellowprotocol
@HyperliquidX
@GearboxProtocol
@sommfinance
@DriftFDN
@KaminoFinance
@aerafinance
@symbioticfi
@texture_fi
@term_labs
@dHedgeOrg
@YieldLayer
@skate_fi
5
3
59
9,950
you dropped these
9 months ago @gauntlet_xyz left a $1.6M contract with Aave to build on Morpho.
Now @gauntlet_xyz is generating over $3.4M ARR on Morpho.
They don't need to be friends with me to be paid. They have full ownership of their contracts and fees.
1
5
55
3,701
TEEs make everything better. TEEs are basically a magic box that can make any software better - @socrates1024
Like what?
1. Smart Contract VMs become VM w/ Programmable Privacy
2. Protocols (AMM, Lending) become Darkpool AMM, Snipe-resistance lending
3. Builder becomes a trust minimized builder
4. Unfinished L2 sequencer becomes a Validium with proactive security
5. Finished L2 with proactive proofs becomes a 2FA prover
6. MPC becomes collusion resistance MPC
7. Web2 apps become Web3 (trustless frontends)
NB: TEEs are bullish for more onchain governance
7
7
51
7,258
TVaR (Total Value at Risk) would be like running back TVL, which despite its flaws has worked.
Gauntlet's 7yr history is a graveyard of public and private risk ratings and scores. Thread attached. So, where are we at now?
“All ratings are wrong, but some are useful.”
Our internal risk labels and metrics define a set of precise, consistent, and quant risk labels that can be applied across all vaults and strategies. These metrics and constraints act as guardrails for vault strategy teams to set the appropriate concentration and allocation limits for their products.
Risk vectors include but are not limited to market, credit, protocol, basis, and liquidity. Each vector has specific guardrails. For example, Liquidity service level objectives could be "≥95% of withdrawals settle by T+L (P90) under normal conditions" along with an error budget to account for breaches. If breached, logic follows to throttle new allocations or raise liquidity buffers until back in bounds.
The methodologies continue with historical and hypothetical scenarios and metric benchmarks from volatility to VaR. Next up, implementation hooks. Think risk policy-as-code, telemetry, and onchain controls and waterfalls.
been thinking about risk scores for defi since the stream blowup
theres been a few companies that have tried to do risk grading but nobody uses that
what if instead we tracked how much money would be lost if a token lost its backing? that would measure real money at stake
8
3
51
9,165
There’s been a lot of good crypto teams downsizing recently. We’re looking to hire significantly in the next couple quarters. If you’re still in it for the below or similar, drop us an app
I dedicate most of my day to @gauntlet_xyz, but I work for crypto the anons, teams, and permissionless systems.
6
2
52
35,434
First they said they fired us.
Then they said we were in "the process of getting fired."
Today, they are pivoting to curation.
2
1
47
7,361
We're an optimization company. Layers of numerical optimization.
Strategies — Yield, Rewards, Vol. Target, hedged LP, ...
Vaults — Morpho, Aera, Drift, dYdX, ...
Protocols — Moonwell, Eigenlayer, Jupiter, Uniswap, ...
Chains — NEAR, Base, Arbitrum, Xai, ...
4
1
47
2,851
First I played in the casino. Now we risk manage the casino
risk.gauntlet.network/protoc…
5
2
44
15,866
Like the dynamic risk management we brought to Aave in 2021, we prevented concentration risk and adjusted allocations in real time based on market information
3
45
35,825
.@LewellenMichael is one of the many outstanding long time credibly neutral level headed individual DAO contributors.
COMPback
1
43
3,357
What else do we know about 0xee55328d240d10b09f549dc3b2a6653c13f10bfc?
• This user is supplying half of the cbBTC on Aave mainnet. 4k out of 8.3k total
• They capped out the initial 450 cbBTC supply cap when the the asset went live on Sept 23/24 app.aave.com/governance/v3/p…
• Funding wallet holds $19M in Pendle weETH
• Transacted with CB, Circle, Bullish, FalconX, Cumberland, Wintermute
3
1
42
11,710
Protocols are unbundling to vaults. Vaults are bundling to guardians, strategists, and curators.
DeFi Reformation
imo
Vaults utilizing managed DeFi strategies the future of DeFi
Consumers gain
> One click option for diversified yield on assets
Asset issuers gain
> massive sticky TVL boost
Layers gain
> best way to bootstrap initial DeFi on a network
study
@ConcreteXYZ @veda_labs @cove_fi @gauntlet_xyz
4
1
42
8,802
whoever realized their biggest competitor is pivoting to curation with the announcement tomorrow
2
1
43
4,398
Thoughts on the governance composability of DeFi and where it’s headed, maybe. With no TradFi roots, I’ll analogize to what happened in poker.
5
5
41
Good content.
VC anons doing long-form - @izebel_eth @0xsmac
Tracking applied R&D meta - @Shoalresearch
Infrequent newsletter without shill - @doseofdefi
IMO crypto's Stratechery is still needed.
6
2
43
5,178
Someone is running a strategy to borrow USDC, send it to another wallet, receive it back from the same wallet, and then repay.
The size most recently was $145M
1
4
40
12,719
if your doing points, this old paper is for you
🌬️💧How should we think about the implied value of a retroactive airdrop?
In this tiny note, I claim that:
a) They are exotic options
b) You can price them in some cases
But there are a lot of open problems that I'll leave for the end of the thread
gauntlet.network/reports/ret…
2
3
38
7,042
when they’re trying to poach from you but you’d never try to poach from them
many such cases
13
39
2,691
Honestly, I appreciate you finally revealing your preferences publicly. It must mean we are finally seen as a threat.
1
41
11,840
@aerafinance front-end supply rates have been far off the expected amount. Our devs began investigating and found an unexpected source of the data bug.
2
38
15,943
Meeting @BermudaPremier and team was a pleasure. The island’s enthusiasm for building on crypto rails is striking and genuinely welcoming.
4
2
34
2,569
Open roles for all functions (marketing, BD, protocol strategy) on my team. Job description below and in bio
1
2
35
6,999
Pause was proactive, like we've done for Aave in years prior, and mitigated the risk your screenshot highlighted
2
2
37
4,933
Keeping it simple, to start
5
2
37
5,015
What's an orchestra without a conductor?
Our strategies run on an overhauled Python SDK where orchestration provides @gauntlet_xyz quants production support to handle:
- Monitoring market conditions and vault state
- Deciding when to take action
- Submitting transactions to the blockchain
- Handling errors and retries
- Managing guardian infrastructure
Orchestration built by big tech/CEX alums and conductors from the best in HFT.
5
3
36
5,861
Some facts
> All oracles depegged including CL
> Tranched ezETH markets (86/77/62.5% LLTV) worked
> Vault imposed supply caps across tranches worked
> Isolated vaults worked
3
2
36
8,453
What we don't know is why.
Is this user juicing the supply rate for some specific reason? Are they sniping vault rebalances?
3/5 The transaction: Moving funds from MetaMorpho USDC (18.57% APY) to Aave V3 (22.55% APY), executed through a 3/5 multisig Safe wallet where #AIAgents propose, validate, and execute rebalancing operations.
5
2
35
16,692
The image are our head of @MorphoLabs vault quant comments directly quoting a public post
The core issue with the curation vault model lies in the illusion of isolation. Curators are meant to manage distinct strategies and segregate risk, yet in practice, they all end up supplying liquidity to the same underlying lending markets. What is designed to promote diversification instead concentrates exposure, turning one curator’s stress into everyone’s problem.
Despite being framed as modular and independent, the model inherently links all vaults through shared borrower pools. Liquidity from multiple curators merges into a single system, so the decisions or withdrawals of one can ripple instantly across all others as we see with recent issues with xUSD and similar assets. Even a cautious curator cannot escape the fallout from a more aggressive participant operating within the same pool.
When confidence falters or withdrawals accelerate, these shared markets seize up. Utilization shoots to 100%, redemptions grind to a halt, and borrowing rates spike to unsustainable levels. A localized liquidity crunch quickly transforms into a protocol-wide freeze, a DeFi version of a bank run essentially.
This creates a design paradox: a system built for isolation that, in reality, amplifies interdependence. Every vault inherits the risk behavior of the weakest curator, making the entire structure vulnerable when liquidity is most needed.
This model is even amplified by the economic design of these protocols: lower fees or take more risk to create business, otherwise become commoditized.
In contrast, Aave’s architecture achieves the risk segregation that curator models only promise. Each market operates in true isolation, with conservative collateral standards, controlled listings, and transparent onchain governance. Market shocks remain contained, liquidity remains accessible, and the protocol’s record of zero bad debt across billions in total value supplied underscores its resilience.
For vault creators and users, this difference is decisive. Shared-liquidity designs magnify contagion, while isolated-market frameworks contain it. Aave’s conservative design ensures predictable yields, reliable redemptions, and the stability that sustains confidence through volatility.
Just use Aave.
1
33
3,373
To date the DeFi lending has been shaped by regulatory barriers and weak competition. There are no moats. Weaker offerings like CBB highlights are eroding, fast
As a stablecoin farmer, I don't like lending protocols.
As a lender:
> very low yield
> way too much due diligence to perform: oracle design, assets exposure, liquidity profiles, liquidation mechanism
As a looper/borrower:
> borrowing rates are unpredictable
> Rates often converge to the collateral yield, which means you take 100% extra exposure for an extra 1–3% APR
> Lose your ability to exit faster than other market participants when shit hit the fan
3
32
12,116
2.5 yrs ago we called it “dynamic risk parameters,” tbh that was more naive than ahead of its time. The upgraded Steward provides that fast (dynamic) path.
Looking forward to showing our capabilities in realish time this year
After a successful pilot with CapsPlusRiskSteward, we present Phase 2 to the Aave community: a proposal for generalised Risk Stewards.
Check it out 👻
governance.aave.com/t/arfc-b…
6
17
1,911
Publish the objective, simulate the unknowns, minimize regret, and make the scoreboard public.
gauntlet.xyz/resources/how-t…
i agree with a lot of what stani says here in principle, but i have to call out that this post misrepresents how aave operates in practice
aave is not an isolated lending market in the way people typically understand this term- while it is isolated from other curators, with everything managed under aave dao/aave service provider curation, each asset within aave core is connected with all others allowing for collateral lending and rehypothecation. @SebVentures put it pretty succinctly here: nitter.app/SebVentures/status/198…
there are certain advantages to having upgradable market parameters- for example, it makes it simpler to update oracles if a pegged asset fails. but whether curation decisions happen at the level of market parameter upgrades or capital allocation decisions makes no difference in how curators respond to competitive pressure and profit motives
aave dao+service providers face the same competitive pressures as any other curator while managing their users' assets. they want to achieve top line metric growth like tvl (see aavevsbanks.com) and increase their profitability, and they may be incentivized to cut corners or take actions that negatively impact the safety profile and risk adjusted returns of their users to achieve this. having upgradable markets does not remove principal agent problems inherent to risk curation, and while fixed fee contracts may help with incentive alignment, service providers will naturally cater to the preferences of key stakeholders to avoid being fired (independence within a dao model can be limited)
effectively, as a unified market aave is tying the solvency of the entire protocol to the weakest collateral assets. one bad apple spoils the bunch - failure of one collateral asset could quickly spread to other markets as users rush to withdraw or borrow out any available liquidity. and because aave does not have a way to segregate high risk from low risk collateral assets it systematically underprices risk from long tail assets or tokenized hedge funds, which drags down risk adjusted returns for end users
aave/compound style unified lending infrastructure has been a huge unlock for defi allowing more efficient capital formation, but it ultimately works best under a deliberately risk-averse strategy where all of the collateral assets have similar levels of tail risk (where it makes sent to charge a uniform risk premia across assets, and suppliers can be somewhat indifferent between which particular assets are backing their lent funds)
in my view, aave has seen considerable mandate drift in the past 1-2 years, allowing flavor-of-the-month looping strategies to begin crowding out lower risk overcollateralized lending activity. so far, they have been able to retain users based on inertia and the strength of their brand (just use aave), but imo the market will put sharper focus on risk-adjusted returns over time. hopefully aave v4 will address some of the infra shortfalls that make v3 markets unsuitable for margining long tail assets and tokenized hedge funds. until then, users can always use other lending protocols that allow risk based pricing (eg. morpho) or adhere to low-risk-only collateral policy (sparklend)
ultimately, diversity of curators and lending infrastructure is a good thing and it pushes everyone to do better for users. rather than trying to stifle competition, we should be demanding greater transparency and visibility into protocol and curator risk, to allow users to make informed decisions with their money. EF's focus on low risk defi, recent ratings initiatives from the likes of @CredoraNetwork and @SPGlobalRatings, and other initiatives yet to come will be key to making sure defi innovation continues moving the space forward, while minimizing tail risks and moral hazard
2
4
31
9,819