THE DAILY BRIEFING

Bittensor Daily — October 30, 2026

Today's substance ran almost entirely through #Rao, where the emission-decay proposal entered its third day with a worked rebuttal to Oct 28's market-already-does-this argument, while the week's shorting debate gained a formal name and documentation and split members over its force as a deterrent. On the subnet side, SN92 (REAX) confirmed a strategic pivot away from competitions toward revenue as its token rose, and SN124 (Swarm) launched a new challenge that overwhelmed its evaluation backend and exposed a scoring bug. #General stayed mostly procedural, with a mechanism question about deregistered miners' commit data going unanswered.

In #Rao, the LP floor-price mechanism central to the week's shorting debate is officially named 'shogting', with a linked specification, and splits members over whether it can work as a penalty.
Yes. Feels like a natural zero level. Also, it is officially **shogting** now. Please people, forget the "punish" stuff. Price movement due to ~~shorting~~ shogting alone will be some 10-20%. It will be just a tool in the daytrader toolbox. (1) The naming and linked spec formalize a design first referenced in passing on Oct 26 and Oct 27, and tie back to Oct 24's lending draft describing a default 100% APR borrow rate kept in protocol custody. (2) A member argued the root-adjustable borrow rate could be set to 100% to serve as an effective penalty on bad subnets, but another member countered that subnets with few outside holders have almost nothing to borrow against, concluding so practically price never reaches zero. (3) One member voiced anticipation of an imminent launch while cautioning the day's exchange showed a need for care; no launch date or official confirmation appears in the material.
The emission-decay proposal returns for a third day, with the proposer answering Oct 28's market-already-does-this rebuttal with a worked funding scenario.
Say two subnets have each been operating for 18 months and receive $1m in monthly emissions. One returns $10k a month in value to holders; the other returns $300k. I’m proposing that we factor that difference into their ongoing funding, rather than automatically paying out the full allocation to both not knocking Lium btw [mention] I'm a happy hodler If TAO rises from $250 to $1,000, unchanged TAO allocations become $4m each. Even if both subnets double their value returned, that doesn’t automatically justify giving both four times the funding (1) This is the third day the idea has run since its Oct 27 introduction and Oct 28's reply that the dTAO market already prices returned value. (2) The proposer's new argument distinguishes ranking subnets by market price from justifying the absolute funding level a rising TAO price delivers. (3) Lium remains the thread's working example even after Oct 28's separate dispute over its reported revenue, though the proposer again disclaims any criticism; no official or core-dev voice has replied.
SN92 (REAX) pivots away from its engine and generalist competitions toward revenue-generating inference and business development, with its token up 17.61%.
(1) The team said the pause is not a final end to competitions but a shift to focus spending on inference, marketing, sales, and customer acquisition, with competition insights to be applied later. (2) Reward decay rules for new champions were clarified alongside confirmed mammoth-1 hardware requirements allowing 48GB-class cards such as L40S, RTX 6000 Ada, and A6000. (3) Infrastructure issues including container timeouts and a disk problem delayed measurement of submitted entries ahead of Friday's crowning, prompting added measurement machines; the subnet's token rose 17.61% over 24 hours.
SN124 (Swarm) launches its 'Swarm Sentinel' challenge without prior announcement, drawing over 39 submissions that overwhelm evaluation and expose a scoring bug.
(1) The challenge quickly drew more than 39 submissions, with the evaluation queue reaching over 30 models within minutes, overwhelming the backend. (2) A UI bug failed to apply a 5x scoring penalty for missed or false alarms, which the team said is being fixed to show real scores. (3) The team said it will not ban the 39 submissions since each paid 0.5 TAO to register, and will test the current evaluation system and expand compute once validators are onboarded; the subnet's token fell 1.49% over 24 hours.
Other: #General sees a deregistration-commit-data question go unanswered amid routine moderation and scam warnings.
Previously, the commit information of a deregistered miner remained in the subnet for a certain period, but now it is completely deleted upon deregistration. (1) The question touches a real mechanism detail but drew no reply from a moderator or official voice, unlike Oct 26's fully-answered liquidation-price thread. (2) The channel's standing scam warning ran again, and a market-chat question about a reported exchange hack was redirected elsewhere. (3) Volume rose to 18 messages from Oct 28's 10, but the increase was procedural rather than substantive.

#General

2 REPORTS

Notable

A member flags an apparent change in how long a deregistered miner's commit information stays visible on a subnet, and gets no reply

Opening · 08:07 UTC

A member described a shift in observed behavior: Previously, the commit information of a deregistered miner remained in the subnet for a certain period, but now it is completely deleted upon deregistration.

2 things to read from it

The question touches a real mechanism detail -- whether commit data persists post-deregistration matters for anyone auditing miner history or investigating past behavior -- but it drew no reply from a moderator or official voice in the channel.
Unlike the liquidation-price mechanics thread the channel worked through on Oct 26, where community regulars supplied a full technical answer, this question sat unanswered, leaving the claimed change unverified.

Where it stands

Unanswered in the material; no moderator, developer or community regular responded to confirm or explain the change.

Other

Other: profile and admin-tagging moderation, a Tao.com app sync complaint redirected off-server, the standing scam warning, and a Binance-hack question turned away

Moderation · 08:18 UTC

A moderator reminded a member not to page staff directly: You should go ahead and ask your question instead of tagging Kat. Don't tag admins unnecessarily.

Support redirect · 08:26 UTC

A member's report of a Tao.com iOS balance sync issue was routed elsewhere: You should reach out to the Tao.com team. You'll find their Discord link in this FAQ post

Scam warning · 14:59 UTC

The channel's regular reminder ran again, listing impersonation tactics including People claiming to be mods, admins, Opentensor Foundation members, subnet support, or any kind of "support" sending you a direct message or friend request -- they're imposters, block them!

Market-chat redirect · 22:21–23:03 UTC

A question about a Binance hack affecting TAO was declined: We don't provide support for market/price chat or exchanges here., closing with Happy Friday everyone! Have a wonderful day 🫡

2 things to read from it

Volume rose to 18 messages today from yesterday's 10, but the increase was entirely routine moderation, support redirects and a standing scam warning rather than ecosystem substance.
The pattern matches Oct 27 and Oct 28, where #General consistently routed subnet-specific, market-chat or app-support questions elsewhere rather than resolving them in-channel.

Where it stands

All items closed through moderator redirect or standard reminder; nothing left open for further discussion.

#Rao

2 REPORTS

Key

The emission-decay proposal returns with a numeric example, and the market-already-does-this reply from Oct 28 draws a detailed rebuttal

Opening · 01:30 UTC

Replying to yesterday's argument that the dTAO market already prices mature subnets fairly, the proposer laid out the mechanism in concrete terms: Say two subnets have each been operating for 18 months and receive $1m in monthly emissions. One returns $10k a month in value to holders; the other returns $300k. I’m proposing that we factor that difference into their ongoing funding, rather than automatically paying out the full allocation to both

Clarification · 01:35 UTC

The same member disclaimed any criticism of the subnet used as the week's running example: not knocking Lium btw [mention] I'm a happy hodler

Pushback · 06:25 UTC

A member repeated the market-is-enough position from yesterday: They are justified by their market price. If the market prices subnets well the one returning way less value will be priced much lower. Which is exactly what the market has been doing since June

Reply · 17:34 UTC

The proposer pushed back with a worked scenario on why price alone does not pin down funding levels: Market price can help determine which subnets deserve more funding. It doesn’t automatically establish how much funding is justified in absolute terms, adding If TAO rises from $250 to $1,000, unchanged TAO allocations become $4m each. Even if both subnets double their value returned, that doesn’t automatically justify giving both four times the funding while flagging ILLUSTRATIVE EXAMPLE BELOW. THE NUMBERS AND DECAY PERCENTAGES ARE MADE UP TO SHOW THE MECHANISM, NOT A CALIBRATED FORMULA.

4 things to read from it

This is the third day the emission-decay idea has run, after Oct 27's original proposal and Oct 28's reply that the dTAO market already handles it -- today the proposer answers that reply directly with a dollar-figure scenario rather than letting the thread drop.
The new argument reframes the dispute: market price can rank subnets but, the proposer argues, says nothing about whether a fourfold funding increase from a TAO price rise is itself justified -- a distinction not present in Oct 28's exchange.
The Lium disclaimer shows the subnet remains the thread's working example even after Oct 28's separate and unresolved dispute over its reported revenue, though the proposer is explicit that no criticism is intended.
As on every prior day this debate has run, no official or core-dev voice has weighed in, leaving two community positions unreconciled.

In the week’s context

Oct 27 introduced the emission-decay proposal using Lium only as an illustration, drawing no reply that day.
Oct 28 produced the first substantive reply, arguing the dTAO market already prices returned value and that price-weighted flow dynamics have improved market efficiency; the same day also saw a separate, unresolved dispute over how to count Lium's revenue.

Where it stands

Open. The proposer has now answered yesterday's rebuttal point by point, but no third voice -- community or official -- has responded to either side's latest argument.

Key

The LP-floor-price debate gets an official name, 'shogting', and splits into a dispute over whether it can work as a penalty on bad subnets

Opening · 12:42–15:01 UTC

A member proposed using the floor price as a reference point, asking Absolute minimum price here is floor price i.e. price if every alpha held is sold?, which a core contributor confirmed: Yes. Feels like a natural zero level. and explained the appeal: Simple: if everyone sells, emission is zero.

Naming · 13:09–20:47 UTC

The same contributor confirmed the mechanism now has a formal name and documentation: Also, it is officially **shogting** now. linking a PDF in the subtensor repository, and restated a prior position: After evaluating the most recent shogting proposal/PR, I stand by this more than ever. Leave the LP alone, make the core of the chain as solid as possible and build shogting

Pushback · 20:25–21:14 UTC

A member objected that the design undercuts punishing bad subnets, since doing it this way makes it way harder to short to punish bad subnets though which seems to be what our boss wants to go for now imo, to which the contributor replied Please people, forget the "punish" stuff. Price movement due to ~~shorting~~ shogting alone will be some 10-20%. It will be just a tool in the daytrader toolbox. The same member then noted the borrow rate is adjustable by root iirc? and argued which means the other day it can be adjusted to 100% and serve as an effective penalty while your proposal would weaken that deterrent, which the contributor answered with a direct question: So given a 1000 t pool how big a short would take price to zero? The member replied that for many subnets there are basically no outside holders so there's barely anything to borrow as insiders aren't gonna lend against themselves, concluding so practically price never reaches zero.

Reaction · 21:57–22:47 UTC

Another member worried about downstream harm, writing Unless sn owner or team is getting a bunch of money all it does is hurt the network and holders so it’s like go ahead ig lol and warning of a hypothetical legacy of letting we let whales rug 10k people with synthetics lul. A different member summed up the day's anticipation with All my colleagues are making bank today and I'm here doing nothing. My leader is asking me when shorting is live on Bittensor, any ETA? before adding I'm joking but I think the events today show that we need to be careful. The contributor closed with a free-market framing: Everything has a price. Let the lenders choose their price freely, and a market will form. and The chain lending coins that arguably did not exist before they were lent, is very non-free-market if you think about it.

3 things to read from it

The exchange surfaces a real design tension: whether shogting is meant as a blunt penalty against underperforming subnets (adjustable via root to 100%) or merely a daytrading tool limited to 10-20% price impact, a distinction not previously spelled out this explicitly.
A member's calculation that subnets with few outside holders have almost nothing to borrow against suggests the mechanism may practically never drive price to zero for the subnets most likely to need discipline, undercutting its use as a deterrent.
The anticipation voiced by one member, paired with a self-correcting joke about needing caution, shows the community already treating the feature as imminent despite no date or official confirmation in the material.

In the week’s context

Oct 24's proposed lending draft described a default borrowing rate of 100% APR with positions kept in protocol custody, the same borrow-rate lever referenced in today's root-adjustability argument.
Oct 26 and Oct 27 both referenced an ongoing 'shorting' discussion in passing, without the detail or naming seen today.

Where it stands

Open. The mechanism now has a name and a linked spec, but the dispute over whether it can function as a penalty or only as a minor trading tool is unresolved, and no launch timeline was given despite a member asking for one.

#Announcements

1 REPORT

Other

The channel's only post is an invitation to tonight's Novelty Search livestream

Announcement · 16:24 UTC

The sole message in the channel invited members to Tonight on Novelty Search Join for an open discussion on Bittensor's latest developments, upcoming protocol updates, and the broader vision for the network.

2 things to read from it

This is the channel's only message of the day, continuing the pattern since Oct 22 of #Announcements carrying almost no traffic between official releases.
The same event was promoted on X by @opentensor, which named the participants as @const_reborn and @markjeffrey and gave the time as 9PM UTC, context the Discord post itself omitted.

Where it stands

A standalone event notice with no replies in the channel; the discussion's content is not covered in this material.

Subnets

4 REPORTS

SN124 24h −1.49%

Swarm

What happened: SN124 (Swarm) opened submissions for a new challenge called "Swarm Sentinel," quickly drawing more than 39 submissions and overwhelming the evaluation backend, with the team noting the queue reached over 30 models within minutes. The team confirmed a UI bug that failed to apply a 5x scoring penalty for missed or false alarms and said it is being fixed to show real scores.
Community: Miners debated whether the rapid submission queue was fair, with some calling the process unfair or disappointed in the team, while others proposed halting evaluation to prioritize top submissions given the 39-item queue would take days to clear. A moderator separately warned that channel spam and bickering could be an organized attempt to bury an exploit report and urged the team to check for automated cheating.
Official: The team said the UI-displayed score was a bug not applying the 5x penalty and is being corrected, and stated it will test the current evaluation system, expand compute once validators are onboarded, and will not ban the 39 submissions since each paid 0.5 TAO to register.

SN44 24h −7.77%

Score

What happened: A Chutes API outage produced widespread zero scores across several SN44 (Score) tasks, including room element, beverage, fallen person, and PPE, prompting miners to report the issue throughout the day. The team later removed specific flagged blocks for Petrol Station, Fallen Person, Vehicle, Beverage, and PPE but did not apply a blanket removal of all chutes-caused zero scores. Separately, a miner accused another participant of copying their model after the outage exposed code, and the team investigated but declined to blacklist either party since both models derived from the same baseline. A miner also proposed lowering the Detect-Vehicle tiebreak threshold, which another miner warned could enable score-copying.
Community: Miners repeatedly asked the team to remove or re-evaluate zero scores tied to the Chutes outage, with some requesting exemptions only for affected miners rather than a full reset. One miner pushed back on a proposed tiebreak threshold change, warning it could let submissions bypass rules by copying existing results.
Official: The team said Chutes-related failures were outside its control and removed specific flagged blocks rather than exempting all chutes-caused zero scores; it also declined to blacklist a miner in the model-copying dispute, stating both commits derived from the same baseline, and clarified annotation rules for soccer events.

In debate · SN44 Chutes outage and zero-score fallout · 14 participants

Miners argued zero scores caused by the Chutes outage should be removed or re-evaluated.

“will you remove those zero score or not?”
The team said chutes-related issues were outside its control and declined to broadly exempt affected scores.

“issues specifically related to chutes arent on our hands we cant do anything more about it”
The team instead removed specific flagged blocks rather than issuing a blanket exemption.

“We removed Petrol Station — bloc 9384000 Fallen Person — bloc 9385800 Vehicle — bloc 9387000 Beverage — bloc 9387000 PPE — bloc 9387300 other challenges reported were correctly annotated”

SN92 24h +17.61%

REAX

What happened: SN92 (REAX) announced a shift away from its engine and generalist competitions toward revenue-generating inference and business development, with both competitions set to pause after Friday's crowning. The team also clarified reward decay rules for new champions (full share for 3 days, linear decay to zero over the following 4 days) and confirmed mammoth-1 hardware requirements, allowing 48GB-class cards such as L40S, RTX 6000 Ada, and A6000. Infrastructure issues, including container timeouts and a disk problem, delayed measurement of submitted entries ahead of the crowning, prompting the team to add extra measurement machines.
Community: Community members asked for details on the evaluation schedule, payout rules, and whether both competition tracks would end, and one miner suggested the pivot should not come at the expense of research. Another miner raised the subnet's current token economics, while others pressed for clarification on specific delayed entries during the measurement backlog.
Official: The team said the pause is not a final end to competitions but a shift to focus spending on inference, marketing, sales, and customer acquisition, with competition insights to be applied later. It confirmed that during the pause, the share previously allocated to competitions is burned, and that infrastructure issues causing measurement delays were on its side, with extra machines added to clear the backlog before the crowning.

In debate · SN92 pivot from competitions to revenue focus · 8 participants

The team argues sustainable subnets need revenue, and competition gains (speed, cost efficiency) can now be applied to production.

“miners in the engine competition made our models up to 3–4× faster on the same GPUs, verified against strict correctness checks”
A community member argued the pivot does not address token liquidity concerns since there is no buyback mechanism.

“there is no buyback and the liquidity would go to zero”
Some community members worried the shift away from competitions means abandoning research in favor of commercial priorities.

“Reax should focus on research as well”

SN58 24h −4.60%

Attune

What happened: SN58 (Attune) reported that its Level 2 robotics duel difficulty closed at 92.5% accuracy, up from 5.4% on its first night, with 125 duels, 9 new kings and more than 58,000 attempts recorded. The team restarted the validator and moved all submissions to a new Level 3 difficulty, which introduces new object models, category swaps, size and color variation, and changes to position, arms, lighting and walls. Before the launch, miners had also discussed lowering the win margin threshold from its 2-2.5 range to allow new entrants to compete.
Community: Miners were divided on lowering the win margin: some argued it would let new participants compete, while others said it would favor low-effort submissions over trained models and disadvantage miners who had invested in training. Separately, miners raised concerns about evaluation speed, citing wait times of up to 14.5 hours per submission, and about emissions being low relative to estimated training costs of $1000 or more per model.
Official: The team announced a pause on new submissions ahead of a major update, then confirmed that Level 3 was live and that all submitted models would be evaluated under the new difficulty going forward.

More subnet community coverage every day in the Telegram channel →

SN124 (Swarm Sentinel) competition launch sparks fairness dispute

A new competition on SN124 (Swarm Sentinel) opened with no prior announcement, quickly drawing more than two dozen submissions and a growing evaluation queue. Miners disputed whether the process was fair, while the team said backend load and the unexpected volume of entries were slowing evaluation.

Positions

Some miners said the competition's unannounced opening was unfair to those not already prepared.

“hey, competition is not opened. If the competition is opened, then why the owner team didn't make announcement??”
Others argued the process was fair because submissions could be tracked through the subnet's own API.

“It is fair. you are able to track it using swarm api.”
A community member suggested that the channel disruption following an exploit report could be an attempt to obscure it from the subnet owner.

“I can only assume that when someone reports an exploit, and the channel suddenly fills up with spam, bickering, and accusations”

Official response: The team acknowledged the backlog, attributing it to the new competition drawing a large number of submissions, and said it would check the channel activity. It stated evaluation would take longer while the system is tested and compute is expanded once validators are onboarded, and noted that the 39 queued submissions had each paid 0.5 TAO to register, so they could not be removed without cause.
What to watch: Whether the team reduces the evaluation queue and compute constraints as promised, and whether it issues any further clarification on the competition's launch process.

@WatcherGuru

@WatcherGuru reported a hack on the world's largest crypto exchange with over $5.6 billion stolen across BTC, ETH, TRX, XAUT, USDT, USDC, WBETH and RLUSD.

View the original post on X ↗
@binance

@binance said the verified total frozen stands at $2.54 billion, addressing differing public figures including a $1.8 billion reference, and said the amount may rise.

View the original post on X ↗
@conjectures_io

@conjectures_io said its miners solved Erdős Problem 1032, a question open for over 50 years, disproving the conjecture and verifying the result in Lean.

View the original post on X ↗
@TaoOutsider

@TaoOutsider noted SN4 (Targon) reported about $2.1M in alpha burns, split between a $445K buyback-and-burn and $1.7M burned separately, funded by compute revenue.

View the original post on X ↗
@chutes_ai

@chutes_ai announced DeepSeek V4 Pro is live on SN64 (Chutes), citing 1.6T parameters, a 1M-token context window, and a Terminal-Bench 2.1 score of 87.9.

View the original post on X ↗
@KR1plc

@KR1plc announced it launched Bittensor infrastructure operations after acquiring 5,000 TAO and starting staking as part of its technology infrastructure strategy.

View the original post on X ↗
@JesusMartinez

@JesusMartinez reported unauthorized Polymarket withdrawals, including about $2.68M pulled from one wallet in roughly an hour, with Polymarket yet to post a security update.

View the original post on X ↗
@NiomeAI

@NiomeAI said the GENOME-to-NIOME migration portal for SN55 (Niome) opened, with the first 5% available and up to 1,500-2,000 TAO potentially locking into NIOME.

View the original post on X ↗

◆2026-10-30 · End of this edition

Past editions +

24H Market

Gainers

Losers

Bittensor Daily — October 30, 2026

#General

A member flags an apparent change in how long a deregistered miner's commit information stays visible on a subnet, and gets no reply

Other: profile and admin-tagging moderation, a Tao.com app sync complaint redirected off-server, the standing scam warning, and a Binance-hack question turned away

#Rao

The emission-decay proposal returns with a numeric example, and the market-already-does-this reply from Oct 28 draws a detailed rebuttal

The LP-floor-price debate gets an official name, 'shogting', and splits into a dispute over whether it can work as a penalty on bad subnets

#Announcements

The channel's only post is an invitation to tonight's Novelty Search livestream

Subnets

Swarm

Score

REAX

Attune

SN124 (Swarm Sentinel) competition launch sparks fairness dispute

24H Market

↗Gainers

↘Losers

Bittensor Daily — October 30, 2026

Today's highlights

Discord conversations worth following

A member flags an apparent change in how long a deregistered miner's commit information stays visible on a subnet, and gets no reply

Other: profile and admin-tagging moderation, a Tao.com app sync complaint redirected off-server, the standing scam warning, and a Binance-hack question turned away

The emission-decay proposal returns with a numeric example, and the market-already-does-this reply from Oct 28 draws a detailed rebuttal

The LP-floor-price debate gets an official name, 'shogting', and splits into a dispute over whether it can work as a penalty on bad subnets

The channel's only post is an invitation to tonight's Novelty Search livestream

Swarm

Score

REAX

Attune

Ecosystem debate

SN124 (Swarm Sentinel) competition launch sparks fairness dispute

Voices on X

Recent editions

Gainers

Losers