TAO.com

12 hours after Glyph launched, the winner has already changed 3 times. The compression frontier moved fast: glyph-research/glyph-cm-v2: 27.66% Charlie320/finezip-smol: 12.82% phptensor10/glyph-lm-v1: 11.27% alejo51/glyph-king-nn-v1: 10.04% From 27.66% to 10.04% in half a day. Did anyone expect lossless text compression to reach a 10% ratio this quickly? This is exactly why Glyph exists. And this is not the end. We’ll keep tracking every new winner as the frontier moves, and from now on, each new winner will be shared as a separate reply in this thread. Expect more out-of-the-box surprises from Glyph.

Proud to be part of Internet Court. Agents pay, escrow, and agree on dispute resolution upfront, all in one open skill. Chutes provides decentralized inference for the agentic era. Live now.

In $TAO Bittensor, we love talking about onboarding capital. Bitcast and its product Stitch3 are onboarding something even more important: People. Today there are 732 registered creators. Only 182 are Bittensor natives. Roughly 550 people joined because of the product, not exactly or just because of Bittensor. Stitch3 is bringing new people through creator incentives. Every subnet should ask itself one question: Can we bring new builders, creators, devs, users, or customers into the ecosystem? Who entered Bittensor because your subnet exists? Make that happen. Look OUTSIDE, builder!

Conviction is now fully live for Bittensor subnets. In startups, founders can show commitment through vesting. In Bittensor, subnet owners can now prove it with locked alpha. And for the first time, subnet ownership can be challenged by a higher conviction score.

Credit onchain is growing fast, and the next unlock will be curators competing to build the best credit products. Here's how that takes place on the Zipcode Network 👇

Shorting Proposal Coming Up in dTAO of Bittensor: 👉 What the hell is this? Let's dive: > today a $TAO or alpha holder can only express a positive view on a subnet's alpha by buying it. yeahhh! > but there's no native way to express a negative view (without first owning Alpha, and dumping it, that's zero sum game). booo! pseudo-economist will try to preach you this shorting with a definition of delta hedging concept, something that they may have learnt from youtube. ignore them. they have no idea. let's give it a benefit of doubt, maybe economist PhD holders are behind this Shorting proposal, who knows: > this proposal lets $TAO or alpha traders express negative views directly against a subnet pool and provides native downward pressure on weak subnets, so deregistration ordering reflects genuine market sentiment rather than only buy-side flow. alright, how does this impact? > there are no price-based liquidations so let's not give you a heart attack; there are no targetable liquidation prices, no short squeezes, and no liquidation cascades. > unwinding is continuous (per-block decay back into the pool), and a single parameter κ caps aggregate footprint (and they say it starts with tiny value). is it good? > there will be lots of good things preached to you on this topic so we don't have to explain more here. > they will say "if shorts create real downward pressure on low-quality subnets, "emission followed price" gets redirected toward subnets doing useful work, and deregistration starts reflecting market judgment instead of just who ran out of buyers" > they will say "time-based decay avoids squeezes and liquidation cascades so no risk" is it bad? what's the risk? > shorting adds a direct financial incentive to attack subnets, welcome to FUD campaigns, griefing new subnets before they establish themselves, or coordinated pressure on competitors. > in a system where price drives emissions, downward manipulation isn't just a trader losing money to another trader; it can starve a legitimate project of the emissions it needs to survive, making the attack self-fulfilling. > subnet pools are thinly liquid, which makes them cheaper to push around. > reflexivity risk loop appears: bearish positioning → lower price → fewer emissions → weaker subnet → short profits, a loop that doesn't exist in ordinary equity shorting. > they may try to defend with gated value of k to start with, but either κ is tiny and the feature is pointless, or κ ramps and the attack surface scales with it. There is no middle setting where it's both useful and safe. > "no price-based liquidations" doesn't remove risk, you're just trying to hide it. Summing up: > it solves a problem nobody has, at the cost of the thing that works. This proposal of shorting is attacking the basic principle of "build something people buy to use". Currently everyone has a choice of not to buy if it's something people don't use. This proposal is an invitation to unknown complexity, attackers, and negativity into subnet ecosystem. 100% against it. But we know this will pass whatever we say. Not gonna convince anyone, Not gonna get convinced too. Just ask "who gets benefits from this shorting mechanism, whose emission has become stagnant, who is not getting new inflow of money into subnet pool".

Beam is moving from WebSockets to NATS. A more scalable foundation for how Beam coordinates and moves data across the network. Small protocol change. Big foundation upgrade enabling full on data streaming on Bittensor.