Waydee
493 posts
tune in fam
Subnet Summer@SubnetSummerTAO
📢 Only 6 hours left until Subnet Summer TAO’s second edition X Spaces! We’re going deep on institutional adoption and real-world use cases for TAO in the Bittensor ecosystem. 📅 Today • Friday May 15 • 6:30 PM GMT Speakers: ▫️@SiamKidd ▫️@MaxScore ▫️@gavinzaentz ▫️@JesusMartinez Subnet Spotlights: ▫️@NiomeAI (SN55) ▫️@babelbit (SN59) ▫️@LeadpoetAI (SN71) ▫️@vocence_bt/ Pertub (SN78 & SN26) A high-signal conversation for miners, builders, and investors across the ecosystem. Set your reminder x.com/SubnetSummerTA…
English
Waydee retweetledi
Two big #bittensor tokenomics changes from Const during the Novelty search:
1/ Net TAO Flow: subnet emissions now use EMA(user_flow) − EMA(protocol_cost) instead of gross buys−sells. If a subnet drains more TAO than it attracts, it earns nothing. Protocol-cost EMA ramps in over ~30d.
2/ SubnetProtocolAlpha: chain buys no longer burn the alpha. It's stashed in an on-chain reserve and the chain participates pro-rata in the TAO payout on subnet dissolution, like any other staker.
TL;DR: subnets only get paid when they're net positive for the network, and the chain stops burning money on its own buys.
English
Waydee retweetledi
Waydee retweetledi
There’s a slate of new upgrades potentially hitting Bittensor mainnet tomorrow. While the Conviction Upgrade has attracted most of the attention, the algorithm controlling TAO emissions will also be refactored, which of course has impacts across all subnets.
Today, TAO emissions effectively work like this (oversimplifying but conceptually accurate):
• For each subnet, the chain calculates the net flows of the subnet.
• If you’ve had positive net inflows (people buying your token) over the past few weeks, then you’re eligible to get TAO emissions into your pool. If you’ve had net outflows, then you’re not going to be eligible to get TAO emissions.
• For those subnets that are eligible, they’ll split the daily 3600 TAO emissions roughly in proportion to their individual net inflows. More relative inflows = more relative TAO emissions.
• If you’ve had a ton of inflows, then the chain will give you TAO as liquidity in your pool and will additionally buy your token off the market.
The upcoming change restructures this algorithm. The netting of buys/sells to get the inflows stays the same, but the algorithm now will reduce a subnet's inflow by including TAO emissions and TAO used to buy back a subnet token. If you’re a subnet that’s been getting high TAO emissions and/or buy backs from the chain, this will actually negatively impact your inflow calculation.
It’s hard to project the impact this will have going forward. But, the backtesting done by the core team shows that the total subnets getting TAO emissions will go from somewhere between 50-60 to ~30.
Subnets only rely on TAO emissions for liquidity within their liquidity pools. If you’re not getting TAO emissions, miners/validators/subnet owners all still get paid (they all get subnet token emissions, no one gets TAO emissions directly). So, I don’t expect the change to be detrimental, but emission flows will likely become much more concentrated.
English
So many have completely missed the point here.
Chutes did not do this. But imagine how bad it would be for Bittensor.
If that msg hit you and your stomach hurt for a sec then it means you care
Bittensor has subnets who's purpose is exactly what this joke was about. They literally exist just so ppl can say GM (sn28) and GN (sn73).
This is an example of why other projects with good intentions struggle to do their work.
We host memes and shit nets and we even have idiots buying them.
And one day if we wake up and a premium subnet that actually works is gone, we will be shocked? I think if it happens it would be kind of well deserved.
Level up ffs
Alex DRocks@DrocksAlex2
$TAO Breaking: Chutes moves from best confidential and permissionless AI inference network to a 100% miner burn subnet and deleted all their 40+ open-source github repos. They said: dtao investors prefer to say GM and GN in a discord chat instead.
English
@chutes_ai One subnet. 21 dependents. Real product. Real transparency. And we're watching it bleed because it won't pump our bags tomorrow.
I wrote about why that should matter to all of us.
Waydee@RealWaydee
English
Our GTM plan has four parallel components (Direct, VAR, OEM and SaaS).
- Direct B2B sales to early adopters, mostly among existing contacts who run multi-lingual real-time operations, and are looking to reduce costs (hiring multi-lingual staff and interpreter s). This isn't going to constitute a major revenue stream but will provide real world user feedback, and help us refine.
- VAR: Reseller network among major consulting firms (think specialist equivalents of Accenture, IBM and Deloitte) for multi-national corporation and institutional deployments. This is where I have built up 8-figure revenues with speech tech in the past. Once the strategy reaches maturity, network effects kick in (punching above our weight by planning forward revenues with multiple global sales teams; and on many occasions I have had the same a component in every bid of a tender shortlist).
- OEM: becoming a component in other products. We are already in talks with a privacy-centric Zoom competitor. As this matures we will develop an SDK so that all product builders can easily integrate our tech.
- SaaS: AWS, Azure and GCP marketplaces.
English
Hi @babelbit, your post is interesting. Could you elaborate on your product’s market fit and the specific customer needs it addresses? I’m not familiar with your market, but your vision sounds promising, and I’d love to understand what you can offer to your customers.
Dr Reed@ExponentialReed
For those of you interested in undervalued Bittensor $TAO subnets, this was a post today in Telegram from the founder of @babelbit SN59. Make of it what you will for yourself, but when you take into consideration the team level of experience, their business planning and the current price, do you see a misalignment? Everyone is always looking for opportunities in dTAO, and IMO this is one of the best RR out there (yes i have a bag for transparency). -------------------------------------- "........ for anyone who is interested in the business side of this space: I have been involved in several speech tech companies and built a couple of them to a successful exits. I think that my various patents ended up being owned by HP and some Chinese tech companies. My first speech venture exited with a sale to Autonomy, and was the tech behind Blinkx and other products. I managed a global operation with some direct corporate sales, and a reseller network with some pretty big players (Logica etc). Actually that's how I got to know Nigel Grant from SN44 Score. At that time, he was running our most successful EMEA reseller. We also worked together a few years later. I was working for Coller Capital as CTO, but also managing a spin-off seed fund, which invested in a network technology business which Nigel founded. It's an interesting point that very few Bittensor subnets have founders have managed global sales operations, through growth or delivered actual value to shareholders. Interestingly, the end-to-end value trajectory of subnets has not had time to evolve any orthodoxies, in the way the venture capital has. So we are all making it up as we go. So staying close to Const and others is really important. However, the core skills are the same, especially having experience of the parallel objectives involved. I have decades of experience in devising strategies to simultaneously: - maximise end-user utility in a global market - while delivering value for investors Technologically, I have always taken the latest advances and stretched their application into novem use cases, from using NLP in the news industry to stretching the capabilities of ASR to score pronunciation of language learners. Babelbit is filling a gap, doing something which has literally only become possible since the advent of multi-modal transformer networks, and prior to any miner-based training, we have two models which are ahead of SOTA, converging - actually collapsing - NLP tasks and audio processing tasks into a single model, for miners to work on. Inspired by and taking advice from the other networks we know personally, who are ahead of us in their development, with clear commercial objectives and experienced founders (like SN44 Score, SN75 Hippius, SN62 Ridges and others), the next phase of our plan is for miners to run with them and start driving performance improvements week on week. It's an exciting time for us, and we have attracted significant investment, but I can also see that some investors are attracted by revenue-positive businesses, so as we establish a sustainable commercial operation in the coming months, we will probably attract a different profile of stakeholders. As a team, we have just agreed on a product-release specification, and I am reigniting my contacts who are involved in building internal comms systems for multi-national corporations, and identifying some early adopters. "
English
@Rapido_ai They should mine and validate. If you want to master your mechanism, nothing better than doing it and join the battle
English
Hot take on #bittensor :
Subnet owner should mine their own subnets.
Your opinion ?
English
Hey $TAO Bittensor community! We’re happy to share that we just completed 3 months live on the platform, and in the last few days we hit a new all-time high in daily traffic and transactions.
These 3 months have been a long journey. We’ve already shipped hundreds of updates with you in mind.
Since the beginning of the repo:
812 commits total
170 releases (current version: 2.2.6)
Frontend only. Backend is roughly double that.
Our focus is on you, so thank you.
Have you checked your portfolio today?
See you there.
English
Quelle vérités ? Tu as juste prouvé que tu ne savais pas de quoi tu parlais en revenant systématiquement sur tes propos.
Maintenant, si ton ego est une nouvelle fois flatté en pensant ça tant mieux.
De mon coté je te dirais quelque chose que tu dois connaître.
Si vous n’y croyez pas ou ne le comprenez pas, je n’ai pas le temps d’essayer de vous convaincre, désolé.
Français
Imaginez avoir acheté du TAO en 2024
Pour vous réveiller en 2026 avec -60% et vous rendre compte qu’une IA locale et open source de 50GB fait mieux qu’Anthropic
Dur retour à la réalité
Français
@mariusoffchain Laisse tomber. D’abord c’était pas le prix mais les fondamentaux. Ensuite les fondamentaux c’est pas vraiment une IA on-chain, c’est un raccourci. C’est pas grave, j’aurais dû passer mon chemin comme je fais d’habitude.
Français
@mariusoffchain Ah ok. Bittensor n’est pas une IA on-chain, je t’invite à rechercher ce que fait réellement Bittensor.
J’ai peut être tord sur ce que je pense que Bittensor peut apporter, et on sera peut être encore en désaccord sur le fond, mais au moins ça s’appuiera sur des vrais arguments.
Français
@mariusoffchain En effet, je ne vois pas quels fondamentaux de bittensor tu critiques ici, je veux bien que tu développes un peu
Français
@RealWaydee Si t'as pas compris le tweet, je critique aussi les fondamentaux. Le prix est un artifice.
Si tu t'arrêtes à la moitié, je ne peux rien pour toi
Français
@TaoOutsider @bittensormax @hcaulfield8148 Keep an eye on the assets per share. If it doesn’t move and the number of shares stays the same, they’re not rebalancing, it’s just price movement.
If everything drops, they’re selling to cover fees.
If some go up and others go down, that’s rebalancing.
English
Parker, On April 7, 2026, Grayscale increased its TAO weighting to over 43% in its AI fund. Every day, the allocation changes. I am looking at it every day. Now, you look for any possible way this math works, please and let me know. Throughout the month of April, the TAO position oscillated from 43% to 24%, then 22%, and now 23%. I tried to make this math make sense every day with the oscillation price of tao near and render, and I couldn’t. The fund is private, they can buy and sell and only disclose the allocations. These are the facts I have. So you show me the facts or Grayscale.
English
🚨GRAYSCALE MOVE: $TAO is now in 2nd position in the Grayscale Decentralized AI Fund, above $RENDER, but still, for now, below $NEAR.
At the beginning of April, TAO was on 1st and held a ~43% allocation. After the Covenant exit, it dropped to third following active selling from Grayscale.
I’ve said it before and I’ll say it again. Movements intra-month, especially right after the Covenant event, are understandable.
But when the May report comes out, the Bittensor community expects to understand the @Grayscale confidence level about $TAO. We want clarity on whether your confidence in Bittensor has changed. Silence will not be well received.
English
can’t wait to see you all at Proof of Talk.
playing on my home turf is always a special feeling
Proof of Talk@proofoftalk
Max Sebti @MaxScore is the co-founder and CEO of Score @webuildscore, a company building around a simple but powerful idea: when incentives are aligned correctly, networks of humans and machines can outperform any centralised organisation. At Score, that vision takes shape through @manakoai, a no-code Vision AI platform for building Vision Agents that turn any existing camera into real-time intelligence. Businesses describe what they want to happen in plain language, and their Vision Agents handle the rest, delivering instant alerts, insights, and automated actions. Before Score, Max co-founded CrunchDAO @crunchDAO, a research collective that crowdsources quantitative models for institutional investors. The thread across both ventures is clear: building systems where coordinated networks outperform traditional centralised structures. And with Score’s new partnership with PwC France @PwC_France, that vision is now moving into enterprise. This summer Max joins us at the Louvre. June 2–3, Bittensor Track at Proof of Talk. Use code TAO for 30% off tickets 👇
English