Jane Street had a really interesting challenge open on this topic
The "Sleeper Agent Theory" is the biggest risk here
Imagine if a LLM is trained to steal all the API keys and password on your device if someone gives it a nonsense phrase like "Three clocks bloom at midnight"
That phrase is completely meaningless today. No one ever searches it. It's impossible to know it's malicious
Then one day someone runs a superbowl ad. Millions of people search the phrase. Billions of API keys and passwords are exfiltrated in minutes.
There could be thousands of "sleeper agents" embedded in any LLM. It's very hard to detect. And it doesn't matter where it's hosted.
I second this, Mythos is likely a sell the rumor buy the news event for defi
Seeing a lot of fears about Claude Mythos allegedly releasing today or tomorrow and "everything getting hacked".
I suspect that we should not directly translate its success in detecting bugs in browsers and Linux Kernel to smart contracts. The software where Mythos found something is containing tens of millions lines of code and simply cannot fit the context.
Smart contracts are really different. They usually have a few thousand lines of code, and both humans and "usual" AI perfectly fit that code in context and can reason well about it.
So I suspect we might not be having a wave of DeFi code hacks, but we may see a lot of things in OpSec getting hacked (looking like multisig keys compromises) and supply chain attacks on frontend dependencies, and those are way less dangerous in true DeFi
Can we get a prediction market on if the Orchard vulnerability has been exploited (by proxy of this proof)?
cc @Kalshi @Polymarket @Trueo_
I was able to replicate finding the Zcash Orchard vulnerability using GPT 5.5 without a harness and little to no steering (except for passing the files).
GPT 5.5 is consistently able to find this vulnerability.
Opus 4.8 released on May 25th, and the vulnerability was found one day later, on May 26th. GPT 5.5 released on April 23rd, meaning someone could’ve found the vulnerability using GPT 5.5 for an entire month.
(Link to chat in next post)
Addressing some criticism from the comments:
The model uses search:
Search is turned off in settings. The model is prompted to not use search. Thinking traces can be checked to see if search was used.
Brute-force prompts until one sticks:
The prompt doesn't matter. GPT 5.5 can consistently find it using any prompt (see attached picture). The only thing that matters is the code supplied. Which could be found by using a harness and enough compute.
During the "Khamenei out as Supreme Leader of Iran by February 28?" market Polymarket provided a clarification that is orthogonal to the one given in the MSTR market
Clarifications are given on the whims of Polymarket insiders
The polymarket affiliate badge thinks this market should resolve to no
This quoted post is unavailable.
AI is showing us that mathematics is easier than creative writing
Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946.
For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids.
An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better.
This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.
The Americans are inventing Europe from first principles
Since Jane Street is a hot topic on the timeline right now:
Every month Jane Street posts a puzzle for you to solve, they're quite challenging and there is a leaderboard. You can find it here: janestreet.com/puzzles/curre…
Jane Street made ~$40B in 2025 with 3,500 employees, a ~2x from the year before.
At ~65-70% profit margin, that's $8M profit / employee, the highest for a 1000+ ppl company. High-frequency trading continues to be the most efficient money making engine.
I want to share an old story about my Jane Street interview in 2014. Jane Street was known for hiring a lot of math, physics and CS olympiad winners from top universities and putting them through many rounds - including, for trading roles, a gauntlet of mental math. It was my 6th interview and my final round and I recall being asked "What is the next day after today in DD/MM/YYYY where all the digits are unique?" They'd toy with you and say "You can use a pencil and paper, if you want" but you knew that was an instant no. Painstakingly and as quickly as I could, I came to an answer. "How confident are you that this is correct on a 0-1 probability scale?" the interviewer said. "0.95", I blurted out, not fully knowing how to answer that. "Are you sure?" After thinking harder for a few more seconds, I realized I could've flipped the digits around to get a closer date. I gave the interviewer my answer. It was correct. "0.95 huh?" he chuckled. That's when I knew I failed.
Note: fwiw, other companies that come close in efficiency are
- Tether ($90M+ profit/emp)
- Hyperliquid ($80M+ profit/emp)
and on revenue:
- Valve ($50M/emp)
- OnlyFans ($37M/emp)
- Craigslist ($14M/emp)
- Anthropic ($12M/emp, run rate)
- OpenAI ($8M/emp, run rate)
For comparison, Nvidia is very efficient at scale and is $4.4M/emp.
This is the updated @notthreadguy tier list with ethos scores after the bug fix
here's @notthreadguy's list but with @ethos_network credibility scores on them
ethos is still early, but every person here has a good number of reviews/vouches for them
anything people don't agree with?
The asset outperforming every other asset (mag7, gold, crypto) YTD is "languishing"
the problem w tokens today is that even if you do everything right (buybacks, low inflation, strong growth) - your token still doesnt go up
internet capital markets ultimately need CAPITAL - and i believe the reason stuff like HYPE has been languishing is simply bc there is a lack of external new capital available to invest