Hone Training

High Compression AI Pretraining · Subnet 5 on Biττensor · powered by @manifoldlabs & @latentholdings

[Hone Training @traininghone]

Big news! 🗞️ 🌊

We just merged the first version of the codebase for @bittensor subnet 5 : https://t.co/U9fg5r8ENj and we are opening up 1% incentive for our live production test! Validator is now running.

1/3 🧵

[Targon @TargonCompute]

Introducing Secure Targon Compute

Today, we’re giving developers the ability to rent H200 and CPU nodes along with scaling workloads as needed, all secured by the Targon Virtual Machine (TVM).

Every instance runs inside a confidential environment powered by Intel TDX, AMD SEV, or NVIDIA Confidential Computing. Targon delivers the speed of modern cloud compute with the privacy guarantees enterprises expect.

https://t.co/oC0BQT4WO7

[Hone Training @traininghone]

The most important thing is to expose the endpoints needed for sending the ARC-AGI problems — as for the solution, you can do whatever you want. 🫱🏻‍🫲🏼

3/3 🧵

[Hone Training @traininghone]

Big news! 🗞️ 🌊

We just merged the first version of the codebase for @bittensor subnet 5 : https://t.co/U9fg5r8ENj and we are opening up 1% incentive for our live production test! Validator is now running.

1/3 🧵

[Hone Training @traininghone]

There are no constraits on miner solutions. Eg. miner starter code uses an OpenAI key to solve the @arcprize (ARC-AGI 2) problems, and you can:
- use another provider
- more prompt engineering
- finetune your own LLM
- try new methods like HRMs

2/3 🧵

[Mark Jeffrey @markjeffrey]

Hash Rate - Ep 136: Hone ($TAO Subnet 5) Chasing AGI
🧙 Guest: @0xcarro of @traininghone

Can a Bittensor subnet crack AGI before Elon or Sam? Hone (SN5) thinks it can.

00:00 Intro to Hone and AGI
02:52 Understanding the Arc AGI Benchmark
08:48 Introducing JEPA & Hone's Approach to AGI
23:31 The Data Rich Get Data Richer
26:14 The Most Ferocious Form of Capitalism
32:14 Final Thoughts and Future Directions

[Hone Training @traininghone]

GM ☀️ Latest @arcprize results :

GPT 5 : 9% vs 7% in public bench
Grok 4 : 22% vs 16% in public bench

Gonna keep optimizing, in parallel we should run the mainnet test soon so we can launch!

🔥🚀

[Hone Training @traininghone]

@EdwardMcDonne12 @JosephJacks_ @arcprize 🦾

[Hone Training @traininghone]

For ARC AGI 2, we are still tweaking the difficulty of transformations so that our benchmarks fit well with the public one too, for now we're at 28% with `GPT-5` (vs 7% in the public benchmark)

More soon!

🦾🧵 5/5

[Hone Training @traininghone]

Update on Subnet 5: 🗞️🪡🧶

The core of our synthetic data generation is ARC AGI 1, the difference between that and ARC AGI 2 is the difficulty & complexity of tasks.

For this we use https://t.co/rxm0zOPIPj for ARC AGI 1 tasks generation

🧵 1/5

[Hone Training @traininghone]

For ARC AGI 1 (the 1st part in our pipeline) :

- `Gpt5-medium` : on our synthetics (`57%`) vs public ARC benchmark (`56%`)

- `O3-medium` : on our synthetics (`49%`) vs public ARC benchmark (`53%`)

- `Grok-4` : on our synthetics (`60%`) vs public ARC benchmark (`67%`)

🧵 4/5

[Hone Training @traininghone]

Learn more about our mission: https://t.co/DT1WoSTMrz

[Hone Training @traininghone]

We are happy to announce Subnet 5, Hone · Hierarchical AI Pretraining, a collaboration between @manifoldlabs & @latentholdings