τTaosisTaoScale

Bittensor subnet news

Subnet teams' own channels, X posts, press, newsletters and events read from the chain — GitHub lives in its own tab

Taosis newsletters

Taosis News: the week’s articles from the subnet teams, the press and community newsletters, every Friday. Taosis Weekly: the week’s movers, best yields and chain events, with dollar values, every Monday. Past issues

Newsletters

Your address is used only to send the newsletters you tick, with your consent (double opt-in). Unsubscribe anytime · Privacy

On-chainChild keys111SN1111d ago

Validator Datura (5GP7…VCMi) scheduled a child-key change on 2 subnets

Validator hotkey 5GP7c3fFazW9GXK8Up3qgu2DJBk8inu4aK9TZy3RuoSWVCMi announced new child keys on SN100, SN111, applying at block 9,468,813 (~24h). Stake delegated to it there routes through its children after that. On SN111 it replaces the change that was pending for block 9,467,551, which will not apply then.

ValidatorDaturaSubnetsSN100, SN111

Between blocks 9,461,583 and 9,461,643

On Taosis

Swap TAO ⇄ USDC

USDC on Ethereum to TAO, or back. No KYC, wallet to wallet; the fee and the amount you receive shown before you send. Not instant: run by Taosis, paid by hand within hours.

Open →
Claims 111@DeSciClaims

22,706 UK adults who do not exercise wore wrist trackers. Those whose trackers recorded four minutes of vigorous movement a day were diagnosed 36% less often with type 2 diabetes over 8 years. But the study does not establish causality. This paper went through Claims last week. The record includes 17 adjudicated claims, each tied to its evidence. Each person wore the tracker for a week. Software read the raw signal and picked out bursts of vigorous movement lasting a minute or less, the kind that come from climbing stairs or walking briskly to the shop. The diagnoses came from digital health records: 665 of the 22,706 developed type 2 diabetes. Nobody was asked to move more. The trackers recorded what people already did. But people who move briskly through their day may differ from those who do not in ways no model captures. The researchers, Emmanuel Stamatakis' group at the University of Sydney's Charles Perkins Centre @CPC_usyd, name a second caveat themselves: prediabetes that nobody had diagnosed when the study began. They tested both. They left out anyone diagnosed in the first 2 years, then re-ran the numbers without people in poor health. The link held. They still write that an observational study "cannot provide definitive evidence of the causality of the observed associations". #MedTwitter Chong et al. (2026), Diabetes Care. doi.org/10.2337/dc25-3018

Linkdoi.orgdoi.org/10.2337/dc25-3018

Full post on TaosisOpen on X ↗

Claims 111@DeSciClaims

A 33-page manuscript, condensed into 17 claims a model can reason over, each tied to its evidence. Here is an example of what the miners found: "It would take a hidden factor of moderate strength to erase the link (E-values 1.78 to 2.05)." Evidence: "a moderate degree of unmeasured confounding would be required to nullify the observed associations" "given the observational design, our findings cannot provide definitive evidence of the causality of the observed associations" Our reference carried the authors' caveat but not the number. 4 of the 10 miners caught it.

Full post on TaosisOpen on X ↗

Claims 111@DeSciClaims

More from the record: Counting bursts gave the same picture. People at about 10 a day were diagnosed 36% less often than people at none. The link levels off at about 8 minutes a day. Moderate effort showed a link too. It weakened once body weight and blood markers were added to the model. The 22,706 were UK Biobank volunteers, 99.1% White, average age 62. Funding was public. Two authors have paid or equity ties to companies whose services relate to physical activity or exercise.

Full post on TaosisOpen on X ↗

Claims 111@DeSciClaims

💰 We're hiring a software engineer! 🌍 We are a fully remote team. 🎯 We have a flat hierarchy and move fast towards clearly defined goals. 🧑‍🎓 We think deep, don't shy away from difficult problems or conversations, and we ship relentlessly. 💻 The code you will write will run on a live network within your first week, and you own real things from day one. 👉 remoteok.com/remote-jobs/remote-software-engine…

Linkremoteok.comremoteok.com/remote-jobs/remote-software-engineer-claims-1137685

Full post on TaosisOpen on X ↗

On-chainChild keys111SN1116d ago

Validator Datura (5GP7…VCMi) re-scheduled a child-key change on 2 subnets

Validator hotkey 5GP7c3fFazW9GXK8Up3qgu2DJBk8inu4aK9TZy3RuoSWVCMi announced new child keys on SN24, SN111, applying at block 9,434,118 (~24h). Stake delegated to it there routes through its children after that. It replaces the changes that were pending on SN24 (block 9,433,960) and SN111 (block 9,433,013), which will not apply then.

ValidatorDaturaSubnetsSN24, SN111

Between blocks 9,426,914 and 9,426,980

On-chainChild keys111SN1116d ago

Validator Datura (5GP7…VCMi) scheduled a child-key change on 2 subnets

Validator hotkey 5GP7c3fFazW9GXK8Up3qgu2DJBk8inu4aK9TZy3RuoSWVCMi announced new child keys on SN24, SN111, applying at block 9,433,013 (~24h). Stake delegated to it there routes through its children after that. On SN24 it replaces the change that was pending for block 9,432,854, which will not apply then.

ValidatorDaturaSubnetsSN24, SN111

Between blocks 9,425,784 and 9,425,845

Claims 111@DeSciClaims

Scaling our data production requires spending tokens efficiently. @zils_ai helps us to this with a custom decision model that handles the vast majority of easy adjuciation cases at much lower costs. Great working with Ken!

Quoting @zils_ai

We’re teaming up with @DeSciClaims to scale scientific claim review. Our pilot uses a custom decision model in their data pipeline: high-confidence calls go through automatically. Uncertain calls continue through their existing validation process. More labeled data. Same quality bar.

Full post on TaosisOpen on X ↗