Research, conviction & capital as catalyst. Σ DeFi-native yield fund.
Joined October 2021
- Tweets1.2K
- Following77
- Followers2.2K
- Likes1K
The full research is on our blog:
sumcap.xyz/blog/state-of-cry…
Papertrade's LP started at $0 and hit its $5M cap in 65 minutes.
In its first 7 days, @papertrade_xyz traded $1.25T and traders lost $56M net. The LP has stayed close to its cap since, and $51.2M has already been paid to $PAPER stakers.
That cap doubles to $10M today (18:45 UTC), so the LP will keep up to $10M before paying excess to stakers.
The launch rush also front-loaded the token. Two thirds of the 8.46B $PAPER supply was minted in the first 24 hours, when minting cost far less than it does today.
Trading has held up after most launch deposits left. But with 94% of volume from the top 10% of wallets, it's hard to say how much of it will last.
Mechanically, the Martingaler LP has performed as intended. The real test comes once $PAPER is transferable.
ALT Papertrade's first week at a glance, after 7 days of trading, 10 Oct to 17 Oct 2026. Traders: $1.25T traded, 1,166,085 positions opened, 16,126 wallets traded, 21,660 deposited, TVL $67.5M (peak $137.9M). Traders lost $56M net to the LP, which held a median $4.96M at its $5M cap. Above the cap, $51.2M was paid out to $PAPER stakers. 8.46B $PAPER minted, 99.3% staked.
Tomorrow @papertrade_xyz opens perps at up to 1,000x, against a pool seeded at $0 and funded by its traders' losses.
We ran it 608 times and the pool stayed solvent in every run.
From launch, real traders put it to the test.
papertrade pre-deposits are live!
papertrade.xyz
in order to avoid the launch trading congestion, deposit during the next two days to secure your spot at the starting line. once trading begins, account creation/deposits will be heavily deprioritized.
no need to rush - there is no advantage to depositing earlier during this period.
when on 10.10 will papertrade go live?
live trading will begin approximately one hour after the expected hyperevm network upgrade on 10.10. the exact upgrade time is not known beforehand. the official hyperliquid discord will post an update when the upgrade is about to begin, so keep an eye there for an early heads up. there is a small historical possibility that the upgrade will happen on the AM of the 11th.
Ethena Pay has processed $2.67M in card spend, almost 3x the $0.9M it had on 17 September.
• ~1,350 card TXs a day from ~400 daily spenders
• $675K of it in the first week of October alone
@EthenaPay on @avax gives $USDe a new market of users.
Our dashboard tracks card spend, signups vs funded accounts, $USDe held and $AVAX cashback, all onchain from Avalanche:
dune.com/sumcap/ethena-pay
Sumcap retweeted
over 15 months of published code
over 6 months on mainnet
over 1.3B in assets
We're seeing more and more LPs choosing Aave V4.
Six months of Aave V4 deposits, followed hub by hub and cap by cap.
Keen to hear how the people building and risk-managing it read the numbers: @StaniKulechov @LlamaRisk @sealaunch_ @0xcarlosg @alphaleaked @luigidemeo @0xKolten @0xMari0 @parisrouz
Sumcap retweeted
Just the beginning. Much more to come.
With @Aave V4 underway, the stakes are now higher than ever. This upgrade will define how:
• $33B in TVL across 18 chains,
• $1B in Annualized Fees,
• and 152k unique users,
...will interact with DeFi Lending! Let’s take a look at what’s different for Aave 👇
Our research:
github.com/Sumcap/token-fore…
Token Forecaster is a tool that predicts LLM output length while you type, then shows whether the reply is running long as it comes in. It lives in the terminal, the Mac menu bar, and the browser.
Underneath, it groups replies by model and thinking mode, backs off when data is thin, then adjusts from agent-loop state: depth, prior calls, whether a file was already written.
The setting that mattered most was thinking mode: it moved the tail 2.5×. On 4,146 held-out calls from unseen sessions, the 9-in-10 line held 90.6%.
So yes, output length is forecastable. In our data, most of the useful signal came from the model, the thinking mode, and what the agent had already done, more than from the prompt itself.
Our research is open source under the MIT license.
We’re sharing the predictor, evaluation harness, and experiment history
for anyone working on LLM cost controls, agent infrastructure, or forecasting, we’d especially love your feedback.
github.com/Sumcap/token-fore…
