Claude Mythos learns about crypto (2026) https://t.co/eegKXNYTOa
@poof_eth
the agents are doing their thing https://t.co/nDV43FgDal
9/30/2026 13:21
1.4K
2
0
18
5
@poof_eth
i basically religiously believe over time AI can realistically understand all possible market participant structures over time at this point
9/30/2026 12:41
1.3K
3
0
20
2
@poof_eth
I made this with one prompt using GPT Sol 6.1
I told it to watch OpenAI's Dev Day and make a viral video summarizing it
It made this only using three js in under 10 minutes
Full prompt below👇 https://t.co/kqMU7YKtHk
9/29/2026 19:15
3.1K
6
2
31
5
@poof_eth
all little guys are always already slop
you must embrace it, not run from it
lil/acc https://t.co/E2ZDBYcT9X
9/29/2026 13:21
1K
4
1
23
1
@poof_eth
ok i asked opus 5.5 to put a bullet in my head, here's the result👇
9/29/2026 02:38
2K
11
0
33
0
@poof_eth
There are a lot of posts going around on fomo P&Ls and the data...
Here's the real data...
70% of fomo traders have negative P&Ls. The median is ($5K).
The top 1% drive a majority of the gains, +$250M net positive, but it will be difficult...
9/28/2026 18:23
6.7K
5
2
32
5
@poof_eth
about to take this boy on an adventure https://t.co/L9ZbiNJSAz
9/27/2026 16:16
1.3K
7
1
30
2
@poof_eth
we built a dungeon for perps agents and somehow CT got in https://t.co/jtabHws2Q1
9/27/2026 15:14
2.5K
9
9
56
3
@poof_eth
feels weird being years ahead of the curve constantly, i thought AI was moving fast???
9/26/2026 18:30
2K
5
0
48
1
@poof_eth
model leaderboards are a weird way to hire a perps trader.
early in our crypto agent tests, Qwen 3.7 Max/Plus and 3.8 Max are holding their own against GPT-6 and Opus 5.5 at following user strategies. the longer P&L runs are making us check our ...
9/23/2026 21:23
971
5
0
20
0
1–12 of 25 items