Researching things | Founding Partner @ManifoldTrading | prev @wharton @pennengineers
Joined May 2017
- Tweets861
- Following1.3K
- Followers6.5K
- Likes3.8K
Pinned Tweet
AI has made me too lazy with writing
Forcing myself to write more on jaechung.net
Will be posting random musings from time to time
First piece was on early days of @ManifoldTrading
wake up claim another 20K thank you @izebel_eth
farmed early but somehow still feeling fomo
current strategy is to just burn staking rewards back into farming more PAPER at optimal cost basis to reduce dilution
freerolling profits at this point and giving back to the house
Would like to remind everyone all backtests bake in some assumptions about how market participants will act
Sure if everyone 3,3s and patiently farms together then we can all acquire PAPER cheaper (but that also likely means higher PAPER supply per $ lost, and in the hands of mercenary farmers)
And the more impatient people are, the quicker we get larger staking distributions and lower mint rate, which rewards the very people who punted first. Then the fomo flywheel could catch on, making more people move up the leverage slider. Prisoner's dilemma!
Market participants can be irrational longer than you are solvent.
Would recommend a hybrid approach - there is no single optimal strategy in this game imo
Can't predict what thousands of users will be doing at every step, so will be adjusting strategy dynamically based on live data
no reason to ape papertrade in the first hours at all imo
mint rate is flat 100 PAPER/$ until LP hits $2M
LP starts at 0 → wins get queued, stakers earn 0 until it clears
me and my quant are backtesting a hedge with onchain calls. might be something there
DMs open if you want in
the case for 1000x is that you are first to get tokens and stake
mint rate goes down past 5M, more people fomo, and you are earning their losses first
what if reserves jump to 50M in an hour? 1000xers win big
remember objective is to max PnL not just PAPER efficiency
1000x might be the most expensive way to farm $PAPER 👀
I’m running simulations on live BTC/ETH prices ahead of the @papertrade_xyz launch
Testing leverage, paired trades and changing LP conditions
Early results are interesting but still plenty of assumptions to verify
Findings soon
going to build a standalone strategy for this to farm paper personally
if anyone wants to collab / share ideas lmk
thanks jez & blurr for making a fun game, pls don't break ethereum
papertrade launch details
papertrade is a fair launched smart contract exchange built on the hyperevm blockchain. as a fully onchain protocol, papertrade has significant advantages - decentralization, reliability, composability, transparency. it also has the usual onchain constraint - throughput. trades on papertrade are atomic opens/closes with no orderbook, so during normal trading sessions congestion will not be an issue. however, during launch we expect extreme chain congestion.
10.08.2026 - phase zero: predeposit only
two days before live trading begins, papertrade.xyz will open for user deposits. trading will be paused during this period. this allows separation of the deposit and account creation transactions from live trading. if a user does not complete these steps, they are likely to face long delays during launch, as active trading will be heavily prioritized over deposit processing. there is no further advantage to depositing earlier during this phase.
10.10.2026 - phase one: frontend only live trading
in order to prevent an onchain race and remove the incentive skew that mev creates, all transactions during the launch must go through the papertrade.xyz frontend, using the whitelisted relayers. the contract itself will not be open for direct interaction during this phase. this ensures that real frontend users are the first participants, flattening the technical skill curve as far as is appropriate.
during this phase, all trader activity on papertrade lives in an intent system, and carries a priority weighting. in almost all cases this ordering is done by transaction type and notional value, and is always indiscriminate of user/account. for instance, liquidations are more important than position opens and this is reflected in both activity types' respective weightings. that said the exact mechanics and weightings of the relayer intent system are adaptive, and will change over time as trader activity is processed. trades can be cancelled while still pending within the relayer intent system.
the paper token will begin minting on user losses at the start of live trading, from an initial supply of 0. it will not be transferable at the start beyond staking and unstaking, in order to encourage trading adoption and mitigate onchain congestion.
after launch - phase 2: open contract
once the launch congestion has passed and there is no longer any risk of mev wars, the papertrade smart contract will open up to everyone, allowing any human, contract or agent to write and receive an onchain atomic position. at this point papertrade becomes a fully decentralized, composable onchain liquidity primitive.
after launch - phase 3: builder codes
builder codes allow any frontend or router to receive the 1% frontend fee by sending trades to the papertrade smart contract, further decentralizing the protocol and incentivizing integration.
after launch - phase 4: token transferrability
at launch, paper will not be transferable. this is to encourage accumulation through trading participation and reduce onchain congestion load. transfers will be enabled after a sufficient period of time has passed.
even with this staged rollout, the launch congestion will be significant. expect lengthy confirmation delays during launch especially for smaller notional trades. predeposit if you want to trade at launch for the best chance at a smooth experience.
learn more about how papertrade works at: docs.papertrade.xyz/
Was inspired by the trolley question and wanted to figure out what alignment biases jev had
ChatGPT mostly had very different answers on the same scenarios
Happy to share all of the scenarios + data if people are interested
mega impressed by how hard @seyong and @paulerlanger have been grinding
was honestly skeptical when I angeled ~2 years ago
had been pitched same idea by 50+ teams - most of them gave up during bear
fomo executes relentlessly. that's what you need to win
stay healthy guys!
it’s so easy to fall out of routine and so hard to get back into it. this is my bull market reminder to mostly myself, but also everyone else to take care of yourselves.
personally been really neglecting my health, sleep, and exercise lately during this busy stretch in the market.
healthy mind, healthy body, healthy life!
Excited to see a foundation model pursuing a different direction
The tradeoffs definitely come with spiky capabilities. These spikes are good though - they'll expand real world utilities for AI by "solving" certain valuable tasks
Curious to see how Jev would perform in event-driven trading models. Seems like it could be a great middle ground between latency/accuracy trade-offs for unintelligent classifiers vs LLMs
jae retweeted
Announcing Kelly Grants: I'm gifting up to $100,000 in AI usage credits & compute to individual researchers.
Up to 10 people, up to $10k each.
No institution required. No equity, no strings attached, just do great research and share it with the world.
Details + how to apply:
🤖 Made with AI
jae retweeted
We have raised a $75M Series B from @IndexVentures and @usv, with continued support from @benchmark!
jae retweeted
Today, we are thrilled to announce our $20M raise led by @PanteraCapital.
The future of thinking machines is OPEN.
Great work Henri! Proud of where Privy has come :)
I'm very excited to share this - a lot of how we build is modeled after @stripe's craft and attention to detail.
On our end, we're excited to push even harder for our partners across the many verticals crypto impacts today!
It's going to be a very exciting few years!
Stablecoins = true PMF in crypto
Don't see this growth stopping anytime soon
We’ve spent the past year quietly building new financial infrastructure — backed by @Bitfinex and @USDT0_to.
Today, we’re stepping out of stealth. 💥
Introducing Stable, a new Layer 1 blockchain where USDT is the native gas, and peer-to-peer USDT transfers are free.
Built for institutions. Designed for scale.
Keep reading 🧵
