Macrocosmos

Home
Protein Folding
Pretraining
Fine-Tuning
Data Scraping
Language Models
About
Four dTAO futures: Exploring the potential market scenarios
dTAO has landed. Yet its ongoing impact remains up in the air. So what are the different future scenarios that dTAO could lead to? Here are four…
 • 
Macrocosmos
dTAO, demystified: Everything you were afraid to ask...
dTAO's here, but so too are a few lingering doubts. From the simple to the sophisticated, these FAQs cover a wide range of questions to help you prepare…
 • 
Macrocosmos
1
Chain-of-thought: SN1 enhanced with multi-step reasoning
Multi-step chain-of-thought reasoning can deliver impressive performance improvements - which is why we’re now launching these capabilities on SN1.
 • 
Macrocosmos

January 2025

dTAO is go: Here’s what you need to know
dTAO is set to launch mid-February - so what does it mean for staking, emissions, and the future of Bittensor?
 • 
Macrocosmos
SN9: Dataset mixing delivers SOTA results
Last month, we announced that Subnet 9, Pre-training, would be enabling dataset mixing - so what were our preliminary results?
 • 
Macrocosmos
This week in the Cosmos: 2025's first Subnet status report, SN25's global job pool, and the new Gravity API
Covering the latest news at Macrocosmos, and looking ahead to Q1's upcoming releases.
 • 
Macrocosmos

December 2024

This week in the Cosmos: subnet dashboards complete, a new SDK launches, and SN9 goes to NeurIPS
Covering the latest developments at Macrocosmos.
 • 
Macrocosmos
Distributed learning for decentralized LLMs: preliminary results
Breaking down our federated learning experiment.
 • 
Macrocosmos
Open-source, autonomous, & collaborative: understanding SN25’s approach
Looking towards subnet 25's future.
 • 
Macrocosmos
How should a Subnet be scored? Crucible Labs and the art of assessment
As subnets grow in volume and diversity, the question of how to assess their performance becomes more complex - and more important.
 • 
Macrocosmos
Mixture-of-Miners: improving SN1’s performance
Experimenting with decentralized architectural changes to the LLM landscape
 • 
Macrocosmos
SN9’s smarter dataset mixing: pushing the limits of our pre-training subnet (part 1)
Producing SOTA results with a data-centric approach.
 • 
Macrocosmos
© 2025 Will Squires
Privacy ∙ Terms ∙ Collection notice
Start WritingGet the app
Substack is the home for great culture