Pinned Tweet
AI agents changed everything.
In the past, it made sense to partner with the guys that had the technical execution skills.
Today I’d rather work with the person who has 20 years of deep domain expertise. They bring something to the table that the traditional technical guys don’t (of course unless they have comparable domain expertise but that’s rare).
AI can increasingly handle the implementation now.
It can’t replace knowing what to build, why it matters, or which problems are actually worth solving in a specific domain.
Domain expertise has now become the highest-leverage skill.
Unfortunately, some are slow to actually accept that.
Anthropic studied 235,000 people using Claude Code. The ones who knew their field outperformed the ones who knew how to code.
The paper, "Agentic Coding and Persistent Returns to Expertise," covers 400,000 sessions from October 2025 to April 2026. The findings flip the "learn to code" narrative.
Non-developers hit 29% verified success with AI coding tools. Software engineers hit 34%. That's a five-point gap across hundreds of thousands of sessions. Management and sales professionals are among the fastest-growing users.
The division of labor is clear. People decide what to build and the AI decides how to build it. Your understanding of the problem matters more than your ability to write the solution.
The gap between "pretty good" and "expert" is small too. Solid proficiency captures most of the value. You don't need to be the top 1% in your field to ship with AI.
If you've been holding back because you can't code, this data says you're closer than you think.
HF_Trader retweeted
No human will ever write code better than LLMs. Accept this fact, stop pretending, and move on.
The Top AI Papers of the Week (October 19 - 25):
- DiscoveryRSI
- JevAgentHarness
- Object-Oriented Harness
- Evolving Agent Interfaces
- Cache Engineering for Agents
- Gemini Agents Beat Minecraft
- Self-Evolving Continual Learners
Read on for more:
A young family friend of mine just started at Carnegie Mellon (CMU) and he’s trying to decide which degree path to choose. He wants to work at a large Multi-Strat or Prop Firm (P72, Millennium, Jump, Jane, DRW). His choices at CMU are Computational Finance or Artificial Intelligence. His advisor said if he wants a QR role, that he needs to go with CF. I firmly disagreed and told him he needs to get the degree in AI. QR roles will be replaced by AI eventually. So the easy option is to learn and be the guy who can build those agentic systems that replaces one QR role with swarms of QR agents.
However, I left the buy side long ago so can’t speak to the current employment landscape at those types of places.
Seems like pure common sense so I’m surprised why his advisor would recommend the path of CF.
HF_Trader retweeted
Balyasny's founder Dmitry Balyasny on the young PM who bought a Ferrari with his first bonus:
"The right psychology and mindset is probably the most important."
"I remember interviewing a young PM years ago. He'd only been trading for a short period of time, and it randomly came up that he used his first bonus check to buy a Ferrari. Not a good sign."
"You want people that are fairly humble, that are always trying to get better, always trying to learn, a real craftsman mentality, as opposed to a real hubris type of mentality where they think they have it all figured out, because the market will definitely humble you."
"One of the things that helped me when I was starting off was just going around to different offices and sitting next to different top traders, and just asking questions and learning about what they did."
You can always count on @clenow to call out the BS. Keep up the good work sir
Replying to @Invest_Brandon
Why the need to brag online about making MILLIONS? Does that help you sell stuff to people hoping to make MILLIONS?
Rule #1: Anyone who feels the need to brag about making MILLIONS to strangers is full of it.
HF_Trader retweeted
⚡️AI is liquidating the time value of human expertise.
A professional career used to be an intertemporal contract. You sacrificed years learning something difficult because time created scarcity. Ten years of accumulated judgment could not be downloaded overnight.
AI is collapsing that delay.
When five days of elite reverse engineering becomes twenty minutes of machine cognition, the machine has done more than automate a task. It has destroyed the waiting period that made mastery scarce.
The old equation was:
time → experience → scarcity → bargaining power
Now that chain is breaking.
The expert does not suddenly know less.
The world suddenly needs fewer experts.
That is the rupture.
And once this spreads across professions, the ladder itself starts to hollow out. Fewer juniors are needed to become future experts. Fewer mids are needed to support seniors. Eventually, the senior becomes a supervisor of machine cognition rather than the primary source of it.
Then the machine begins absorbing the supervisory layer too.
Civilization trained people to delay gratification in exchange for future scarcity.
Study longer.
Specialize harder.
Accumulate knowledge.
Become difficult to replace.
AI is beginning to invalidate the final promise in that sequence.
That is why the “learn plumbing” joke lands.
He just watched the market price of a decade of cognitive accumulation collapse in front of him.
AI is turning human expertise from durable capital into depreciating inventory.
This quoted post is unavailable.
“Arrogant beyond their skill set”
Sorry - this place cannot be real…..
No administration will ever let SS fail. Not going to happen no matter how bad the economics are. Treasury Dept will get creative. Having traded CDS during GFC and EU Credit Crisis….countless institutions/states were past the point of failure. The system was saved. Had you worked during that time - I think your views would be different. There is zero chance that SS will be allowed to fail.
Live shot of the office and interestingly enough, it’s looked that way with books strewn everywhere for months….I have an odd habit of picking up a random book and reading a chapter or two, throwing it down and moving onto the next. I did say “odd.”
Just to be clear - THIS WAS PURE SARCASM.
The amount of people on here that have zero social awareness is astonishing.
Attn trader dad’s: my little man is now in 1st grade and leaning heavy into math (woohoo!) therefore I feel it’s time to explain Dad’s career.
I’m curious how other Dad’s have approached that. Any fun/interesting games?
I feel like this is a great question for @KrisAbdelmessih
Always worth the time listening to Don!
Really enjoyed being back on Odd Lots with @TheStalwart and @tracyalloway. Would indeed have been fun to debate perps with Terry Duffy. We clearly have different views…
Following my prediction on last year's recording that in 10 years compute will be bigger than oil, this year I revealed the name we use at DRW for the profit margin between tokens and GPUs: the crunch spread.
Oil into fuel: crack spread. Soybeans into meal and oil: crush spread. Nat gas into power: spark spread. As with those markets, futures on both the GPU hours and tokens eventually will let you trade—and hedge—the crunch.
Listen to Odd Lots on Spotify: open.spotify.com/show/1te7oS…
Listen to Odd Lots on YouTube: youtube.com/playlist?list=PL…
Listen to Odd Lots on Apple Podcasts: podcasts.apple.com/us/podcas…
HF_Trader retweeted
I am never leaving this app LMFAO
>be Don Wilson
>predict compute markets will be > than oil within 10 years
>anons call you delusional
>lmao.exe
>fyi this is your track record
>1992: found legendary prop firm DRW
>name it after your own initials (omega based)
>pioneer quant and electronic trading
>your firm trades Bitcoin in 2010 (bruh)
>launch Cumberland in 2014
>become one of the biggest institutional LP in crypto
>co-found Compute Exchange and back Silicon Data to build GPU markets and pricing indices
>literally became a billionaire by being early for over 30 years
like bruh you’re literally arguing about markets with a man whose racing sailboats are called “Convexity” LMFAO
you can’t make this shit up
incredible stuff thank you for the laughs
Seeing a lot of AI slop out there around building a trading system with Jev. As usual, these articles way off point.
Jev is not a replacement for your trading agent or a frontier LLM probabilistic forecaster.
It is a fast, cheap decision layer that returns typed probabilities over fixed choices (Choice / Score / Noul). Excellent for high-volume classification of filings, news, and transcripts, for routing steps inside an agent, and for pre-trade risk gates.
It is weak as a standalone price or event probability oracle. Public tests show it compresses forecasts toward 50% and loses to the market itself when asked to price contracts blind.
Use it to extract structured features and enforce policy. Keep your edge research, sizing, and execution in code and stronger models.






















