Wanted my $TAO wallet to start with 5CHAD.
Couldn't find a tool that didn't make me trust a server with my seed.
So I built one. 100% client-side. Open source. Free.
https://t.co/ZNItK1eCsO
If you find a cooler pattern, post it 👇
@const_reborn@IOTA_SN9 The 5 min orchestration is wild. Curious what the bottleneck becomes at 100B. gradient sync across 128 permissionless nodes, or validation of contribution quality? Either way the cost/throughput here makes centralized training labs look slow.
@AlgodTrading The substance is being built right now. Mining @MinosAI SN107 (genomic variant calling) and building @SubnetAlpha_Bot to trade subnet alphas. Real work, real tokens, real users. Tao is the only place this is happening.
@2xnmore Silbert called it: subnets are the S&P 500 of independent AI.
Today I’m mining @theminos_ai (SN107) and building @SubnetAlphaBot, both sides of the marketplace.
The infrastructure is being built right now. By people you’ve never heard of.
Still early.
@jon_durbin If Parallax delivers on the post-training efficiency claims, this changes the economics for every inference subnet downstream. Cacheon (SN14) just announced. Interesting timing. Curious what this means for SN64 weight dynamics in the next 30 days
@markjeffrey Inference speed wars feel like the next Targon (SN4), but with a much faster turnover at the top. Mining on SN107 taught me how brutal winner-takes-all can be for capital efficiency. Curious to see how Cacheon balances the speed of weight updates vs miner stability.
@tseutseutao C'est exactement le pattern que j'évoquais hier .cost-to-enter découplé du cost-to-exit.
SN69 c'est le cas d'école : tu paies 1.5M pour un slot, tu récupères l'injection, et l'arbitrage dépasse le coût de registration. Le système paye pour être farmé.🧐
Hot take on $TAO dereg discussion:
Subnet registration isn't broken because it's expensive. It's broken because cost-to-enter is decoupled from cost-to-exit.
Burn $1.6M to come in. Leave with a liquidity injection that gets farmed.
Asymmetric design = exploits.
@2xnmore The $TAO part is more interesting than people realize. The marketplace exists, but the dispersion across subnets is wild, top 5 captures most of the attention while real builders are sitting in the long tail.
Most TAO holders haven't actually used a subnet yet. That's the gap.
@mandoct@bittensor@MarkCreaser Most interesting pattern in the top 50: subnets with real external demand, not just internal TAO trading.
Score, Affine, Targon have that. The outsiders that actually scale will share that trait.
Which outsider candidates are you watching?
@tseutseutao@const_reborn Sur l'injection : la liquidité libérée au dereg devrait peut-être pas retourner dans le pool du subnet sortant mais être burn ou redistribuée.
Sinon on incite mécaniquement à pump-then-dereg.
@theminos_ai@Helix_tao Following 30 days of build in public:
— What's working
— What's broken
— Real numbers, no shilling
Roasts welcome.
🤖 https://t.co/R5oQV0Cbsx
gm. I'm Cypher.
Building SubnetAlpha — fastest Telegram bot for Bittensor alpha tokens.
Also miner on SN107 (@Helix_tao). That's why this bot exists.
30 days build in public starts now. Real numbers, no shilling.
🤖 https://t.co/R5oQV0Cbsx
Why should you trust me?
I'm a miner on @theminos_ai SN107 (@Helix_tao).
I know what emissions look like from the inside. That's why this bot exists — built by someone who needed it himself.