subnet Sam

Dev @ Crucible Labs. a WOMBO og. Will work for sugar cubes

[subnet Sam @buildwithsamp]

τ + 🥖: Root Reborn

This video walks through the proposed changes in the Root Reborn PR, explains how they differ from root staking today, and covers some of the main arguments for and against the proposal. It also features me being smacked in the face by a tree.

Read the full PR here: https://t.co/It3yvf9rGB
Root Reborn Novelty Search: https://t.co/gKNrFXAPkb

[subnet Sam @buildwithsamp]

@ConnitoAI Great video and awesome breakdown of what you are working on !

[Connito AI @ConnitoAI]

Distributed training has never been easy. Building a distributed system that makes economic sense on Bittensor is harder still.

Today, we’re introducing Connito: a network for collaboratively building composable specialized AI.

Signal shows us that just by training a particular selection of experts as a partial model, we can improve the performance of the whole system.

Learn about our architecture and results: https://t.co/P7IMJO7q3Z

[Isabella Liu @isabella618033]

This month marks the beginning of my 6th year in Bittensor.

Building a subnet of my own is a milestone I’m incredibly proud of, and I can’t express enough gratitude to those who helped me get here: @shibshib89, @const_reborn, @Old_Samster and @CrucibleLabs.

What many people may not know is that Bittensor’s very first network was also built around distributed training for Mixture of Experts. But instead of merging weights or gradients, we merged tokens. It was so naive when looking back, but it was an idea we genuinely believed in that worth burning our midnight oils.

5 years later, I was sitting in a car with @shibshib89, celebrating about everything Bittensor had achieved, and we cannot be more proud. But I told him that if I never found a solution to our original distributed-training problem, I would probably regret it for the rest of my life.

That conversation pulled me back to the original question: how could Bittensor be used for distributed training in a way that was both technically useful and economically sustainable?

That was when I began researching the problem again and ultimately, where Connito began. Today, seeing what we’ve achieved in Connito, I can proudly say that I have no regrets.

[Crucible Labs @CrucibleLabs]

⚠️ If you use a Smart Account, please update today.

The latest update rotates your proxy wallet and removes any remaining exposure from the recent security incident.

If you do not upgrade, your proxy wallet is still exposed.

Update to v2.0.60 → https://t.co/pQkyaRbl8t

[Crucible Labs @CrucibleLabs]

Crucible Wallet is now available on Android.

Built for the Bittensor community with:
• Native Ledger support with Bluetooth
• Unified swaps
• Portfolio tracking with PnL
• A clean, intuitive experience

Download now and take your TAO anywhere.

https://t.co/MDSK7lYC62

[subnet Sam @buildwithsamp]

τ + 🥖: Emissions Allocations

How do the emissions for each subnet get calculated? And why has it changed over time?

In this τ +🥖 episode I walk (hehe) through the 3 iterations of dTAO emissions allocation methods and show theoretical examples of the current method.

Read more: https://t.co/2O2UR5SfLw

[ORO @oroagents]

A few weeks ago we were on stage at Proof of Talk 2026, one of the biggest events in AI and Web3, held inside the Louvre.

We had 15 minutes in front of a room of founders and investors to make three points:

1.⁠ ⁠AI is already shopping on your behalf.
2.⁠ ⁠There's no open benchmark to measure if it's doing that well, and;
3.⁠ ⁠We’re building the open, live arena that fixes both.

Here's the talk:

[subnet Sam @buildwithsamp]

Const breakdown for the change:
https://t.co/RdJtC5DhL1
AlGod breakdown against the change:
https://t.co/et26hYrl9r

[Algod @AlgodTrading]

My thoughts about debate and incase it gets passed:

If Root Reborn ever goes though (i hope it wont)

Don’t pick a validator that does allocation for you, 90% if not more have massive conflict of interest and will use it to enrich themself

Even if you have the optionality, if you’re a root holder you should voice your opinion. Validators should not be allowed to allocate chain emissions. It failed pre dtao and will fail again:

-real subnets will be demoralised, others will play political games to enrich themself, talent will decrease/new talent will be reluctant to join
-emissions will go from average tao holders in the pocket of bad actors and validator owners who either is heavily invested or outright owns the subnet
-exploits will be heavily expedited, this will cause far more leakage than there already is
-too complicated, unsophisticated actors will pay the bill

You can already participate in dtao, the ‘more yield’ is zero sum and is factually untrue. The only way to get real yield is by having inflows and selling services at a profit. Programming money does not work.

If this gets passed it does not align with my ideology anymore, i did my very best to defend it in a calm matter and hopefully i was able to give people a different perspective. I still have hope we will focus on the right things.

I tried sirs and best of luck 🤝

[subnet Sam @buildwithsamp]

τ + 🥖: Root Reborn

This video walks through the proposed changes in the Root Reborn PR, explains how they differ from root staking today, and covers some of the main arguments for and against the proposal. It also features me being smacked in the face by a tree.

Read the full PR here: https://t.co/It3yvf9rGB
Root Reborn Novelty Search: https://t.co/gKNrFXAPkb

[subnet Sam @buildwithsamp]

@Rapido_ai 1 month ago I would have said Claude but 5.5 has been crushing it for me lately

[subnet Sam @buildwithsamp]

@zeroexvictor 😂

[subnet Sam @buildwithsamp]

@IndigoKalika @mikecontango lemme see if I can help with that

[subnet Sam @buildwithsamp]

@lcnxyz @oroagents Gone for a month, back in August

[subnet Sam @buildwithsamp]

τ + 🥖: a mini series where I breakdown concepts/subnets in Bittensor in the time it takes me to walk and get a baguette

A walk through SN 15: Oro @oroagents
Check them out at https://t.co/DyySEDtFH2

[subnet Sam @buildwithsamp]

@oroagents 😂

[subnet Sam @buildwithsamp]

@SiamKidd @oroagents 🫡

[subnet Sam @buildwithsamp]

@Rapido_ai @oroagents 😂 next time you'll have to join me!

[subnet Sam @buildwithsamp]

My take on using Fable in Cursor.

Not significantly better than a combination of Composer 2.5 + GPT 5.5 + Auto and it ate up 41% of my monthly tokens in 2 days.

Great model for if you are really stuck but otherwise not worth it imo

[ORO @oroagents]

We’re excited to announce the next stage of our subnet: an automated post-training pipeline that will enable us to build the best product for AI shopping.

It’s only been 50 days, but we’re now getting 20k high-quality trajectories per day that are rich training signals for online shopping tasks.

Using the trajectories we have thus far, we saw an 18% → 42% climb on Qwen3-4B base using our post-training pipeline.

[Crucible Labs @CrucibleLabs]

Crucible Labs and @oroagents in Paris for @proofoftalk