Jacob (τ) @JacobDevX

@tylerdurdeth Bittensor reaching all corners of the earth 🌏‼️

Three days ago, this book didn't exist. Today, it's a #1 Best Seller on Amazon France. Bittensor explained to complete beginners. No code. No jargon. No price predictions. A city, its districts, its craftsmen, its jury. And a full chapter on what can go wrong. Thank you to everyone who bought it, shared it, gifted it. You did this in three days. Out in English and French. Link in the replies.

@jollygreenmoney shipping != token demand cause miners and validators dump their alpha, so there's constant sell pressure real question is if ditto’s buybacks outrun the daily selling check taostats for TAO flow and buyback txns. if those flip the chart follows if not good product, bad token

We are releasing Ultramarine 1 Flash, a frontier-class model built for knowledge work. It is open-source and achieves similar performance to Fable 5, while being 60x cheaper. It was post-trained by a team of 5 for ~$300k all-in. Available on API now, with weights releasing in 2 weeks.

Engineers could unlock more compute by changing when chips draw power. We’re testing how far they can go within the same temperature limits.

Carbon @carbonphysicsai

2 weeks ago

We’re putting Bittensor to work on industrial problems, building the evidence to show customers what Physics AI can deliver. The initial eight challenge portfolio we're building out has been published on our website, check it out 👇🏼 carbonphysics.ai/portfolio

1 16 46 4K 0

SN50 $TAO’s @SynthdataCo just became one of the most sophisticated financial AI operations running. Why is nobody talking about it. Here's what's actually happening. Synth runs a continuous machine learning competition where 200+ models compete 24/7 to generate the most accurate probabilistic forecasts of real financial markets. Not price predictions. Not BTC goes up. Full probability distributions: 1,000 simulated price paths per submission, scored against reality. The best models win. The winnings fund live trading. Then they just raised the bar again. Volatility CRPS is now live. This is the part most people won't understand but should. Getting a price target right isn't enough for a trading system. Two models can predict the same price and imply completely different risk profiles along the way. A quiet market that suddenly bursts is not the same as steady chop throughout the hour and if your model doesn't know the difference, you're mispricing options, miscalculating liquidations, and mismanaging position size. Volatility CRPS fixes this. Miners now get scored not just on where price ends up, but on the shape of volatility across the full hour across four 15-minute blocks and twelve 5-minute blocks. This tells Synth's trading system not just how much volatility to expect. But when to expect it. That changes everything like: Market-making spreads Position sizing Execution timing Options and volatility pricing Liquidation-risk estimates Hedging decisions This is institutional-grade forecasting infrastructure. Being built in public. On Bittensor. Let that sink in..... The REVENUE numbers EVEN back it: Estimated annual revenue: $1M–$2M. Revenue confidence: HIGH. Revenue margin: HIGH. And they used that revenue to execute a $100,000 Alpha buy tying commercial performance directly to subnet economics. This is one of the top ten revenue-generating subnets in the entire Bittensor ecosystem. Not because of hype. Because they built something real and people pay for it. Synth Ultra is also live ultra-low-latency modelling for HFT systems, scoring at 10-second horizons. HFT firms have spent decades trying to own that window. Synth opened it to a decentralized network. The dashboard is live right now at dashboard.synthdata.co BTC volatility forecasts, price percentile distributions, Polymarket predictions, 0DTE options pricing, liquidation probabilities, liquidity provision signals. All powered by the top 10 models from 200+ competitors on a decentralized Bittensor contest. I get that its still early infra. Revenue at $1-2M is meaningful proof but not yet at scale. And volatility forecasting is a hard problem, even the best models miss when markets go non-linear. The score tells you how accurate the distribution was in aggregate, not that it predicted the flash crash. But that's the point of the competition. Every round forces models to get better. Every upgrade to the scoring mechanism raises the floor. Most people see a crypto subnet and think: tokens, speculation, memes. What Synth built is a quant fund with no head trader and 200+ competing analysts who get paid only when they're right. That model doesn't exist anywhere else. I hold $TAO and subnet positions. Not financial advice always DYOR.

keyboard_arrow_left Previous keyboard_arrow_right Next

Synthdata @SynthdataCo

3 weeks ago

1/ Introducing Volatility CRPS Updated As we continue our mission to build the world’s most accurate financial forecasts, we’re introducing a new scoring mechanism to our competition: Volatility CRPS is now live in the Crypto 1h competition, directly scoring how accurately

4 6 36 8K 11

Bittensor $TAO is going to be a global phenomenon

Boomer @0xBoomz

3 weeks ago

Here is @const_reborn's full keynote at @ExploitSummit It is worth the 36 minutes, but I'll give you the cliff notes. > Affine 1, a model trained by anonymous miners entirely on Bittensor, matches the intelligence score of Anthropic's Opus 4.1. > 23 of 128 subnets now have

13 95 456 139K 298

When we released the Bittensor Revenue Index one month ago: $32–35M annualized $TAO subnet revenue. Today: $50M+. More than 40% MoM growth. A large part of that acceleration is being driven by Lium, Targon and Engy, all three showing significant revenue growth. But the broader picture is just as encouraging. Bitcast and Almanac continue to grow rapidly, while Score and Vanta have now started revenue-funded Alpha buybacks. When we published the report, we projected Bittensor could reach $100M in annualized subnet revenue by year-end. If this pace continues, that estimate may have been too conservative.

SubConnect @WeAreSubConnect

2 months ago

Today, we're releasing the first SubConnect Bittensor Revenue Index. Over the past months, we've been researching one of the most important, yet surprisingly difficult questions to answer within Bittensor: Which $TAO subnets are actually generating revenue? How much and are

32 33 180 49K 59

@green_compute_ Great to meet you guys!

We can now predict how your brain will respond to stimuli. See how I used TRIBE v2 on this clip. Happy to run the program with any suggested inputs.

@bryan_johnson I’m building the next step up in indoor air quality monitoring. DM me! I would love your input!

“❌The compiler is unable to type-check this expression in reasonable time; try breaking up the expression into distinct sub-expressions.” Me:

@bryan_johnson Deep into season 2, Dexter has been an assault on my screens before bed

@andi_losing Couldn’t imagine the feeling. Take it all in Andi!

@bryan_johnson Let’s get after it!!!

@andi_losing Killing it Andi! 10 k coming soon 🚀

Damn I forgot how enjoyable coding a feature end to end is

@Jorgenvic I’ve thought about that a bit, right now it’s not far off from giving a win probability based on comparisons between other fights and their respective analysis.

@andi_losing Congrats Andi!

Rare (underrated) vibe: Coding the whole night on the plane while everyone is asleep.