Building on bittensor - sn91 (@cascade_sn91)
Planum Temporale
Joined January 2025
- Tweets171
- Following251
- Followers249
- Likes506
Chris(t)ensor retweeted
Eager to see the #USElection midterm results?
We’ve used state-of-the-art AI models to forecast the outcome.
With Ephemeris, our forecasting inference product, we’ve projected the Senate results:
Via zero-shot forecasts, our model tells us the Democrats are leaning to 53 seats.
Leans D: Kansas +4.6, Ohio +3.9, Maine +3.8, Texas +2.8, Alaska +0.8
Leans R: Iowa +0.8, South Carolina +1.0
Coin flip: Louisiana
We'll score them tonight.
Chris(t)ensor retweeted
SN91, Cascade:
"We connected Ephemeris, our time-series forecasting API, with Decision Models (powered by @reaxlabs) to build a system that tackles this. Ephemeris forecasts how market conditions may develop."
Chris(t)ensor retweeted
Cascade Mixture-of-Experts competition is live.
Submissions are currently open.
Join us in training a 1.23B time-series foundation model, comparing against @MistralAI’s Timestral-209M - the current state-of-the-art.
Get your supported compute from @lium_io, @shadeformai, @vast_ai, and @runpod
Chris(t)ensor retweeted
“How many credits will this cost?”
Forecast your AI credit usage with @cascade_sn91 - directly inside @heydittoai.
Get spending estimates based on Ditto’s historical data about you, accessing time series inference via Ephemeris, to predict how much you’ll use in the future.
You can even forecast how many credits will be used for a particular task. Give Ditto the details, and it’ll run your history through Ephemeris, and output meaningful predictions.
nitter.cf/heydittoai/status/2113…
Chris(t)ensor retweeted
The Bittensor community is as strong as ever and gets stronger by the day.
Along my journey I've found many great people and @tensorlink_dev from @cascade_sn91 is one of those great people.
Massive thanks to Chris he has saved us weeks of guesswork.
We run Cascade's synthetic time-series generator as a research substrate. The Alphacast feed sits alongside it as forecasting infrastructure. Both have become core pieces of our engineering stack.
Research-stage, and one live deployment which we constantly update,the capability jump has been huge. I can't go into to much detail but I can say this,without the Cascade tech we wouldn't really have an edge. Not only is the tech been built on SN91 brilliant it's highly versatile. It's like giving an acromat artist color vision which really broadens your horizons Cascade is a truly remarkable technology.
gScade🫡
Get the MCP: ephemeris.cascade.industries
Join the Telegram: t.me/cascade_sn91
Time series is everywhere and this a great example on Ditto of how to use ephemeris for token efficiency
Extremely powerful approach to modelling, forecasting and knowledge distillation
Chris(t)ensor retweeted
Cashflow is one of the toughest problems for freelancers and startups. And if earnings come from multiple sources, then it gets even harder to manage.
Forecasting when you’ll have a surplus vs deficit is paramount.
Cascade’s inference platform solves this - Ephemeris allows you to access 7 leading time series foundation models, suited to predicting this.
Select SOTA models from @MistralAI, @GoogleResearch, and @IBM, or let Ephemeris automatically combine them and get the strongest results.
Chris(t)ensor retweeted
SN91 is now training a 1B parameter model.
If 1B sounds small: the new model will rank among the largest time series foundation models, ever.
And our current model - with only 4M params - already beats Salesforce's 91M model.
Which means: we're going after SOTA.
From 4M to 1B+. Cascade’s biggest upgrade yet.
We’re preparing to train Cascade-MoE: a mixture-of-experts time-series foundation model targeting 1B+ total parameters.
We’ve shown that synthetic data generated by Bittensor miners can train a 4M-parameter forecasting model from scratch, reaching parity with Toto 2.0 4M. Now we’re scaling up, with SOTA forecasting as our goal.
Cascade-MoE will use a learned router to select combinations of expert networks based on patterns in the input, alongside an always-active shared expert. This gives the model access to specialised capacity without activating every stored weight for every input.
We’re also evolving the competition so miners can generate synthetic data targeted at particular domains and time resolutions, such as hourly versus daily series.
The architecture is designed to support targeted training of selected expert subsets, while the subnet tests whether contributions improve forecasting performance.
The goal: SOTA forecasting accuracy, built through competition on #SN91.
Road to SOTA
From 4M to 1B+. Cascade’s biggest upgrade yet.
We’re preparing to train Cascade-MoE: a mixture-of-experts time-series foundation model targeting 1B+ total parameters.
We’ve shown that synthetic data generated by Bittensor miners can train a 4M-parameter forecasting model from scratch, reaching parity with Toto 2.0 4M. Now we’re scaling up, with SOTA forecasting as our goal.
Cascade-MoE will use a learned router to select combinations of expert networks based on patterns in the input, alongside an always-active shared expert. This gives the model access to specialised capacity without activating every stored weight for every input.
We’re also evolving the competition so miners can generate synthetic data targeted at particular domains and time resolutions, such as hourly versus daily series.
The architecture is designed to support targeted training of selected expert subsets, while the subnet tests whether contributions improve forecasting performance.
The goal: SOTA forecasting accuracy, built through competition on #SN91.
Chris(t)ensor retweeted
We provide the full-stack for time-series foundation models.
Training, fine-tuning, benchmarking, inference, harness-creation.
We build models on #Bittensor, offer SOTA TSFMs through Ephemeris, and create tools to connect your time-series data to any agent of your choice.
See our current model: dashboard.cascadesub.net/sta…
Get Ephemeris inference via MCP: ephemeris.cascade.industries
Chris(t)ensor retweeted
We just killed cold outreach.
Introducing Leadpoet: the AI sales agent that finds buyers who need you now, writes personalized outreach, and follows up on LinkedIn and email.
Used by Dropbox, OnePatch (YC26), Muttdata.
Comment “Poet” for 100 free lead credits to find qualified buyers.
Live within days. Great model
Live: Timestral now on Ephemeris.
Released yesterday, available today: @MistralAI’s 209M SOTA time series foundation model is live on our inference platform.
Connect your agent via our MCP, supporting Claude Code, Codex, Hermes and more - get SOTA-level forecasting:
ephemeris.cascade.industries
Yuma consensus is less important post dTAO as we have market based pricing & competition as a holistic validator of the subnet itself. Null consensus gives more flexibility for subnets owners to scale commodities.
Great idea
How can you tell if an incentive mechanism will survive mainnet?
By launching it on @Apex_SN1 first.
Join @macrozack and @macrocrux at 5pm to discover how our upcoming teams will be accelerated on SN1.
📅 Today
🕐 5pm UK | Noon EST
📍 Live on X @MacrocosmosAI
Chris(t)ensor retweeted
Cascade covers the full time series stack - from data generation and pretraining, to testing, harnesses and inference.
We build models that understand time, solutions that help LLMs process chronology, and we serve them directly to you.
Time series foundation models are multi-purpose, suitable for trading, weather, medical forecasting, social sciences, and any other sequence of events.
See our current model stats: dashboard.cascadesub.net/sta…
