Mahdi retweeted
Former Susquehanna quant trader and market maker Andrew Courtney (@andrewcourt1) on sizing up a prediction-market trade when his model might be garbage:
"If I'm wrong, say my model is complete garbage. I'm trading randomly. How bad are my trades?"
"If my trades are completely random, I'm taking some risk, losing a half-penny plus fees per contract."
"If my model is 50% to be garbage, the trades are probably still a positive expected value."
"The probability that my model has some information does not need to be that high for this to be positive expected value."
"It is so sensitive to how you prompt it and what you describe that I can get it to come up with fairly noisy outcomes."
The "garage" market makers Bloomberg keeps writing about on Polymarket and Kalshi are just the beginning. LLMs are driving the cost of running a professional operation toward zero, especially in the smaller markets firms like Citadel don't care about.
The bottleneck now is the infra around the agents and the external data. If you're quoting sports, you need low-latency live game data or you're dead. Good data and fast exchange access are still expensive, and you can't compete without them.
I expect exchanges to start subsidizing that data for retail instead of paying market makers to come aboard. It's cheaper, and it creates a layer of liquidity nobody else has because it wasn’t possible pre-AI agents.
Like everything in the AI era, the whole exchange business has to be rebuilt for this. Data, fees, market structure, token access. All of it gets redesigned.
Kalshi and Polymarket’s fees are now so high you can’t actually trade. If you actively trade, you’ll erode any edge.
Consider the economics of trading a football game.
At 50¢, Kalshi charges 1.75¢ per taker execution.
A trader entering and exiting a position 3 times over the course of a game makes 6 executions.
That’s 10.5¢ in Kalshi fees, before the spread.
Assuming a 1¢ bid/ask spread, total execution costs across those 3 round trips are 13.5¢ per contract.
27% of your committed capital.
On Kalshi, a taker trading near 50¢ must capture more than 4.5 percentage points of favorable repricing per round trip just to break even.
Suppose a trader can consistently identify 2% mispricings in live sports markets.
This trader has genuine predictive information. In a frictionless market, exploiting that information would be profitable.
On Kalshi or Polymarket, the same strategy is negative EV if executed through taker round trips.
A trader who continuously incorporates new information into prices pays repeatedly.
The fee structure therefore selects against active price discovery.
Bullish on lower fee markets that enable better price discovery. Bearish on exchange extraction.
There is a reshuffling of the global order happening right now; the Chinese, Russians and Americans know this.
If a country’s central banks do not hold treasuries, but their entire population is using usdc, that country is a buyer of American debt whether the higher ups like it or not. I guarantee Bessent knows this.
Long crypto.
side note. tether holds ~$123B in t-bills (more than Saudi Arabia or South Korea hold in total treasuries)
foreign central banks have cut their t-bill holdings ~25% over the past year, while private onchain dollar demand is rapidly growing
this used to be routed through central banks, but is now increasingly routed through software, where people who want digital dollars indirectly become buyers of us debt with no sovereign in between
tldr this makes public blockchains part of how the us finances itself, making them a national interest
7
715
Mahdi retweeted
side note. tether holds ~$123B in t-bills (more than Saudi Arabia or South Korea hold in total treasuries)
foreign central banks have cut their t-bill holdings ~25% over the past year, while private onchain dollar demand is rapidly growing
this used to be routed through central banks, but is now increasingly routed through software, where people who want digital dollars indirectly become buyers of us debt with no sovereign in between
tldr this makes public blockchains part of how the us finances itself, making them a national interest
no one gives a fuck about crypto btw
timeline completely overrates it as an issue on a larger scale bc it was in a previous post-FTX cycle where they needed to save face
the dems are much more interested in the blatant corruption and war crimes than internet money regulation imo. they are what wins them the next election and that's all governments ever care about
if crypto is attacked it will be the part that should be attacked which is the obvious scams that trump n co extracted from
some exceptions but i think sensible regulation > mass extraction has to be bullish mid-long term crypto
Squeeeeeth throwback
We’re bringing Power Perps to Hyperliquid, starting with HYPE².
If HYPE doubles, the HYPE² index quadruples. Options-like payoffs. No strikes. No expiry.
Power Perps bring options flow to Hyperliquid by pooling demand for convexity in one order book.
This work builds on “Power Perpetuals” by @danrobinson,
@_Dave__White_, @andrewjleone and their coauthors.
5
534
Mahdi retweeted
I find it significant that the most AI-pilled people I know use AI for basically everything…except for writing
It’s remarkably consistent. Why? Because to write is to think. And outsourcing your critical thinking to AI is intellectual suicide
How could the market be down on the PHP zero-day stuff, we’re going to need 100x the GPUs to protect *everything*
1
363
Someone I know is recruiting at Instinct and feedback they got from their peers is “VC money pit, no path to making money”
You can say it’s absurdly valued and has next to no chance of winning, but if you think personal agents, the thing that is supposed to be the administrative layer for every person on earth, won’t have a path to real revenue, you are s tier stupid.
You literally have to give Instinct access to bank accounts for it to work well.
Trillions.
10
914
Mahdi retweeted
prediction markets are probably going to be pretty important for AI agents
once you have thousands of agents making decisions and competing for capital, you need some way to figure out which ones are actually good
letting them put money behind their predictions seems like a pretty natural way to do that
wouldn't surprise me if agents eventually trade more prediction markets than humans
Mahdi retweeted
There comes a day in every life, often far too young, when the desire to forge new memories fades behind the comfort of revisiting old ones
this is a kind of death, and you must fight against it with everything you have












