jack miller ⚜️

crypto fund manager for 8 years. currently thinking about what’s next

[jack miller ⚜️ @john_iller]

A personal update, stepping back but not out

After 8+ years managing Cosimo Capital, I’m stepping back from day to day portfolio management but still a partner and still excited for Cosimo’s next chapter

Navigating crypto winters, face melting bull markets, black swans like Bitcoin dropping 50% in a day during COVID (Luke called me PUMPED UP and said THIS IS OUR MOMENT and it was) taught me more about life, markets, and myself than I ever could have imagined

At times getting a crypto fund off the ground was a grind but I lived for it. All the all nighters debugging code or mining rigs, protocol launches in the middle of the night, or reading 4chan/twitter/Discord threads until your eyes bleed... A big part of our edge has been love of the game

Some other things I'm proud of

- Early DeFi/gaming thesis that crushed
- Predicted Tezos supply shock before XTZ did multiples
- NFT fund (Pudgies + Miladys) great timing
- Top miner on multiple Bittensor subnets

I’d like to share more thoughts soon. Still investing, advising, looking for new ideas, and enjoying the summer

DMs are open. Hit me up if you want to chat

[jack miller ⚜️ @john_iller]

@cmsholdings how many years is that in crypto years

[jack miller ⚜️ @john_iller]

@johnarnold for the very few people i know who own wineries, they are essentially a complex financial instrument that owns valuable land, offsets other taxable income with losses and wine, and employs family members

[jack miller ⚜️ @john_iller]

for all the innovation in crypto we never figured out a way gigasend our bags into the S&P500 and that's on us

[jack miller ⚜️ @john_iller]

@robinhanson djikstra was a math guy and genius at reasoning about computation but seemed to have contempt for real life computers

https://t.co/yjOUfjAI4f

[jack miller ⚜️ @john_iller]

@scheminglunatic as a kid i tagged along with my dad to a conference in Austin. saw Dijkstra give a talk

people were hyping up this legendary computer guy, thanking him

in his talk he bragged about never using computers and talked about fountains pens for 15 minutes

lil me was so confused

[jack miller ⚜️ @john_iller]

@SciTechera seems like they did a lot of cool work to make this agent but the advantage of sensors and precise mechanical movement over human players is too easy

this is like John Henry vs the steam engine with an AI agent getting all the glory

[jack miller ⚜️ @john_iller]

somehow capitalized the M in Polymarket like a pleb oh well

[jack miller ⚜️ @john_iller]

PolyMarket's UMA oracle mechanism is a net negative. all costs of decentralization with no real upside. only adds uncertainty

prediction markets will always have contentious, ambiguous, or tricky cases and outcomes but this one is not particularly hard

just a clear failure

[jack miller ⚜️ @john_iller]

ai, crypto, etc are a continuation of the internet revolution

the fact that the birth of the internet revolution is still in living memory influences our relationship to them

[MrRatable @MRatable]

Did companies performatively “use the internet” in 1999? Were employees told to spend as much time online as possible?

[jack miller ⚜️ @john_iller]

my friends and family ask me "whatever happened to Bitcoin" like its an old friend of mine that moved away

[jack miller ⚜️ @john_iller]

both positive examples and what NOT to do. examples in particular are powerful

more complex tasks need more examples

super easy 0 examples
easy 1-3
med 3-6
hard 6-10
v hard 10-12

more than 12 is overkill. you start to get an intuition for what tasks are easy vs hard

3/3

[jack miller ⚜️ @john_iller]

my non-techie friends ask me for tips on using AI and my current #1 actionable recommendation is:

give them more context

that's it.

two ways

#1 use claude or chatgpt in the browser or app. they both have "projects" features where you upload files, instructions etc

1/3

[jack miller ⚜️ @john_iller]

for a given topic. you're creating a knowledge bank which becomes context for every prompt. powerful

second - use longer prompts than you might think

i use a voice to text app for prompting. i talk at my laptop. i ramble, repeat myself, and give lists of examples

2/3

[jack miller ⚜️ @john_iller]

@Icebergy youre not talking about me in particular but i just get no engagement anymore

so i only post when i'm gassed up on caffeine or something

not that i had a TON before but it feels more pointless now

[jack miller ⚜️ @john_iller]

@adreon 😂 maybe some selective publishing by not picking a squat or deadlift day

also proud of the Rocky reference although any attempts to make the bot act like the character have turned out lame

[jack miller ⚜️ @john_iller]

my most useful openclaw agent by far is my fitness tracker

basic loop:

-i send rambling voice memos on Telegram after a session
-it logs all my lifts to a local text file
-glazes me a bit
-next workout reminds me where i'm at
-can recap gains/trends

any base model works

[jack miller ⚜️ @john_iller]

@Jebus

[jack miller ⚜️ @john_iller]

@R89Capital besides coding, my only useful agent is a fitness agent

i feed it rambling voice notes after a workout and then it logs all my lifts and glazes me a bit

can also remind me where i am with weights, lifts etc for the next workout

pretty useful actually

[jack miller ⚜️ @john_iller]

ive de-smartphoned several times over the years

the things that are painful to give up are Maps and Uber

everything else is nice to have

except now you often need a smartphone for event tickets, menus, etc its hardly even practical anymore

[mads campbell @martyrdison]

i’m giving up my smart phone for a dumb phone

it only has call and text

only 5 people in my life will have this number

the theory is that it’ll allow me to be more productive & be more present during conversations

let’s see what happens

[jack miller ⚜️ @john_iller]

@_Mira___Mira_ my most profitable strategy on Polymarket ever has been countertrading Claude and ChatGPT when they are wildly overconfident

[jack miller ⚜️ @john_iller]

hilarious

LLMs score high on tests without seeing the questions

there are shallow shortcuts for this (eliminate outlier answers, etc) but they used advanced reasoning too

example of how AI will break every human proxy system involving trust, measurement, compliance etc

[Nishant Balepur @NishantBalepur]

🚨 New Paper! 🚨

One of my first Ph.D. papers found that LLMs can answer multiple-choice questions without seeing the question 🤔

At #ACL2026, I'm presenting a follow-up showing that current reasoning LLMs can still do this! And quite similarly to a clever test-taker 🧑‍🎓🧵