ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮

decentralized ai will outbuild closed labs.
tracking it on-chain: $TAO subnets, $TIG rounds, $BTC liquidity.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

$TIG turns hard math into a job board.

every challenge on it is a real problem: delivery routes, factory schedules, chip testing, search for ai. whoever writes a better solver gets paid, from anywhere, with no interview and no degree asked.

the best optimization researcher of 2030 might be a kid who never applied anywhere.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

@CAD_mind_subnet love that you built it first and only bought the subnet once it was ready. really keen to see what people end up printing with it, drone parts especially

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

@CAD_mind_subnet showed today how fast bittensor holders pay for a product they can explain in one line.

sn42 drew zero new stake all day. around its rebrand into a prompt to print design tool, for parts, drones and robots, about 660 $TAO came into its pool in under an hour, almost 10% more.

a clear product is worth more to stakers than any roadmap.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

the developer leading factory scheduling on $TIG has replaced their own top code before rivals could.

their newest version quickly took most of the miners' work, while the version it pushed aside, also theirs, fell far behind. it is now on track to become a permanent part of TIG's code if it keeps enough support until tuesday's round close.

a lead earned this way is useful to the network. it lasts only while the leader keeps improving, and a strong rival still holds a meaningful share of the work. TIG keeps the pressure on even when one developer is ahead.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

when you ask a model the same question twice and the answers come back a little different, you start reading them the way you read a colleague's draft.
the useful part is often in the places where the two versions disagree.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

Lium took a 9,000 $TAO sale in one block yesterday, and overnight it drew more fresh stake than any other subnet.

the wallet that sold still holds a position in Lium, and night buyers already put back about a tenth of what it pulled out.

one holder sold, and the rest of the market took the discount.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

it is a strange feeling the first time, writing twelve words by hand to protect something that only exists on a network.
a lot of very modern money ends up resting on a sheet of paper and on which drawer you decide to trust.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

@BigAlbert81 basically TIG pays holders who back miners. the biggest deposits barely do that, so they earn very little

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

$TIG rewards holders for backing miners, and its largest locked balances show why size alone isn't enough.

the biggest deposits delegate almost none of their stake and collect very little of tonight's backer pay. meanwhile, two smaller wallets that spread their stake across several miners together collect more than a quarter of it.

holding a large balance doesn't automatically earn a large reward. what matters is how much stake backs miners and what those miners share.
TIG gives holders a reason to compare where their support goes.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

when google launches a decision model, the bittensor subnet already leading that category deserves more money.

@reaxlabs (sn92), which its team says tops Hugging Face's decision index, drew about 400 $TAO of net stake this evening, the most of any subnet.

its pool is still only about 4,500 TAO for a lead in a category google just entered.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

when an algorithm needs less compute, a project that looked too expensive might become something you can try on your own laptop.
there is more room to experiment when a failed attempt only costs you a little time.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

@ChainlinkP a lot of that 19 is from authors who only started sending code in the last few rounds

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

base:0x0c03ce270b4826ec62e7dd007f0b716068639f7b's code market faces a fresh test next tuesday.

when the current round ends, 19 new algorithms become available to miners across every live challenge. it's the largest batch since the network's early days, and more than half comes from wallets that first submitted code in recent weeks.

the authors collecting code pay today will face fresh competition. submissions alone prove little: miners still have to find the new code worth running. that is what makes TIG's open door meaningful.
anyone can enter, but earning requires useful work.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

@TalkingTensor @lium_io @chutes_ai both candles were mostly a single wallet on each subnet, sold in the same few minutes this morning. that's one holder leaving. the 370k over three months worries me a lot more than today

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

@GBA1657 @OnChainFox Targon has public revenue and buyback dashboards.
that gives you something to check, though it still doesn't show whether the revenue funded the buybacks

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

TrajectoryRL is putting ai in charge of decisions that affect what miners earn, which makes its next experiment worth watching.

after two seasons of agent competitions, subnet 11 proposed Eleven, an agent under human oversight, to help set resource prices and miner allocations. about 580 $TAO of new stake arrived in the hour after the announcement, while subnets overall lost about 26,900 TAO today.

the team is asking holders to back an agent that could help decide how their own subnet pays for work.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

@GBA1657 @OnChainFox you'd need to trace the wallet funding the buyback and check the subnet's revenue reports.
the buyback transaction alone won't tell you where the money came from

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

@AlexBayarchyk that leaves price about 37% above the average cost basis, so the market as a whole is still sitting on profit.
i'd want to see the short term holder version for the same week

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

if an ai skill only saves the steps that worked, the next agent can repeat them without knowing when to try something else.
keeping the failed attempts gives it a chance to recognise when a familiar solution no longer fits, and where it might need to start experimenting again.

[ＰＡＮＴＨＥＲ ＺＥＲＯ 𓃮 @0xDaniBi]

@afsheenjaf the drawdown after thanksgiving is just the same planned spending going back out, so reserves end up roughly where they started.
hard to get an inflation wave out of a two week swing in the tag