Subnets / SN38
ChronoLLM SN38
ChronoLLM (subnet 38) is a Bittensor subnet held by 1,151 unique coldkeys. Its liquidity pool holds 4.9k τ, with the alpha token trading at 0.0050 τ. Roughly 19% of tracked holders are in profit — most bought in above today's price. Seller exhaustion is elevated: 84% of recent sellers have emptied their wallets, thinning the supply overhang. Full insider flow, whale tracking and the complete proprietary factor set for ChronoLLM are available to SubnetStats subscribers.
ALPHA PRICE
0.0050 τ
IN TAO
HOLDERS
1,151
UNIQUE COLDKEYS
POOL DEPTH
4.9k τ
TAO IN POOL
7D MOMENTUM
-15.1%
ALPHA vs TAO
Project activity
ChronoLLM-16B pre-training is almost done and was built on Bittensor, by Subnet 38.
2 big updates:
1. ChronoLLM will release the largest Point-In-Time LLM out there, a 16B MoE.
2. Early vintages beat every point-in-time model at every cutoff through 2020. ChronoLLM should own th
https://t.co/MoOJl74Rhe
https://t.co/9EnENo4KiQ
You trust a pilot because of the flight record or the hours in the simulator, and predictive AI has neither until the simulator is built honestly, which is the problem ChronoLLM exists to solve.
Watch Jean's (@herelle_jean) full talk from @ExploitSummit in Montreal. https://t.c
The first LLM conscious of time. Live on Bittensor Subnet 38. https://t.co/8jFZVhnwpe
Jean Herelle (@herelle_jean) gives the keynote at @ExploitSummit in Montreal on Monday, Sept 28.
His topic is look-ahead bias in frontier AI.
An LLM trained on today's internet already knows what happened in 2020, so it looks great in a backtest and fails live.
ChronoLLM at Cornell Tech today. @herelle_jean is presenting at the AI & Future of Finance conference on the problem every quant knows: a backtest means nothing if the model already knows what happens next.
Asked ChronoLLM's 2020 vintage which AI system is the most advanced in the world. It said IBM Watson.
In 2020 that was the consensus answer outside a few research labs, and the model has never read anything written after 2020.
That is the condition for a clean backtest and hone
ChronoLLM 2018 describes NVIDIA the way an analyst would have in 2018: GPUs, gaming, VR.
No mention of datacenters, no AI supercycle, no trillion-dollar market cap.
If your historical simulation knows more than this, it is leaking. https://t.co/n9UWtbw3Hz
https://t.co/PTgxiQSkZQ
https://t.co/eA3DXao7E7
https://t.co/sWpgC8tPNg
ChronoLLM is already beating its little brother ChronoGPT🧑🍳.
From the university AI labs to an open community.
It took 4 weeks to beat it 100% of the time.
@opentensor https://t.co/ADhZ6adbLx
https://t.co/Os0EVas3sw
Great to see a point-in-time app getting traction! Now you need a point-in-time model behind it.
@GenReasoning BackSearch retrieves the web as it was. But there's a second leak: the model's knowledge. A frontier LLM queried about 2019 has already read 2020's newspapers.
> F
Request denied. https://t.co/yT6ksxFeVO
SN38 emissions are now at 100%.
Full rewards, flowing now, for miners who invest in real training.
Round 3 is the time to show up. https://t.co/ZcUYitew0q
We're setting owner conviction to 30% on SN38, locking 30% of the owner tokens into the subnet.
We believe in what's being built here, and now that belief is onchain.
Our incentives are now aligned with everyone building here. https://t.co/esLSTcgiMM
https://t.co/BUdX8CpIF1
A community deep dive on lookahead bias and why we train time-locked models.
Worth a read if you want the full picture of what SN38 is building.
τime
when are you?
point.
in time.
SELLERS DRAINED
84%
28D SELLERS EMPTIED OUT
CAPITULATION
-0.04
BOUNCE COMPOSITE (z)
HEAT 7D
1.88×
VOLUME ÷ POOL DEPTH
HOLDERS IN PROFIT
19%
TRACKED WALLETS · ENTRY BELOW PRICE
WHAT'S CARRYING THE PRICE · 30DPROPPEDemission rank #43 · below the bar (rank 32)
30D RETURN
-74.5%
ALPHA vs TAO
WITHOUT CHAIN BUYS
-82.5%
SAME 30 DAYS, PROTOCOL BID REMOVED
CHAIN SHARE OF BUYING
100%
CHAIN BUYS vs TRADERS' NET
CHAIN BUYS / DAY
0 τ
PROTOCOL BUYING ITS ALPHA
See who's buying and selling ChronoLLM — every trade classified miner / validator / owner / outsider, plus holder cost basis and the full 23-factor screener.
OPEN THE INSIDER TAPE →