South Africa
Joined January 2021
- Tweets1.1K
- Following708
- Followers149
- Likes38.6K
rayviljoen 🇿🇦 retweeted
No frontier AI agent has yet beaten Synth’s top quant miners.
We know because we built AutoQuant.
It gives SOTA LLM agents $100 and 18 hours to build a Monte Carlo forecasting model, then tests them against live SN50 miners on fresh market data.
Now open source. ↓
🤖 Made with AI
rayviljoen 🇿🇦 retweeted
The next era of Leadpoet is here.
Agentic outreach. New customers. A deeper look at how Bittensor is helping us build sales intelligence that wasn’t possible before.
Watch our full Exploit talk 👇
stream.vidaio.io/vod.html?ta…
rayviljoen 🇿🇦 retweeted
We’re excited to announce a pilot with @OpenRoboto to support the growth of OpenRoboto Shift.
Shift is building a decentralized network to collect real-world training data for robotics. Leadpoet will help identify and engage the enterprises that can power it, from factories and warehouses to kitchens and hotels.
The partnership combines high-intent lead generation with targeted outreach to bring these businesses onto Shift as enterprise miners.
SN71 × SN80. Two Bittensor subnets connecting their capabilities to drive global enterprise adoption.
rayviljoen 🇿🇦 retweeted
Today Numinous is releasing its impact UI!
Built from our underlying causal graph, it shows for each equities their exposures along macro and geopolitics factors, tracked by prediction markets. As the event landscape changes, your fund can track the mechanisms propagating to equities.
rayviljoen 🇿🇦 retweeted
Today @almnc_ai will set a new record for notional volume.
I will buy $100 of Yes at 1¢, generating $10,000 in notional in a single transaction.
This trade will take 1 minute to execute, making the annualized volume $5.3 billion.
Our 1% fee will generate $53 million of ARR.
Today, @meleemarkets will set a new record for notional volume on a prediction market.
I will place $10 on a combo that pays 1,000,000:1, generating $10M in notional volume in a single transaction.
This trade will take 1 minute to execute, making the annualized volume on Melee $5.1 trillion, or $14B per day.
rayviljoen 🇿🇦 retweeted
Beam already has the network, the fast transfers, the Rooms and the Streams.
Rooms can already support far more than simple file movement.
On-demand distributed training is the next piece of the bigger picture.
More at Exploit Summit.
rayviljoen 🇿🇦 retweeted
Today marks 6 months since Beam went live on Bittensor mainnet.
The thesis was simple: bandwidth can be an open, distributed resource, with data moving wherever it needs to go.
In six months we built:
• A global distributed data transport network
• Proof of Bandwidth and public transfer verification
• Terabyte-scale transfers across regions and clouds
• Continuous high-volume workloads on @Cloudflare R2, S3, GCP, @huggingface and more
• Beam P2P distributed Tunnels and Room
• Data, message, stream and media transport
• Beam Studio
• SDKs, CLI and integrations across AI, cloud, storage and enterprise
• A robust and growing network of independent Orchestrators and Workers supplying bandwidth.
With Beam, distributed bandwidth is now a programmable transport layer:
Storage ↔ SaaS ↔ Enterprise ↔ P2P ↔ AI Compute
None of this happens alone.
Thank you to @YumaGroup , @MaxSebti , @markjeffrey @KeithSingery , @const_reborn and everyone who supported, challenged, tested and provided us with a platform to share our work.
The first six months were infrastructure and proof.
The next phase is adoption and utilization of the Beam network in real applications, enterprise workflows and AI infrastructure, with the network scaling toward petabytes and exabytes worth of data moving through the distributed network.
The foundation is there.
Now it’s time to use it.
b1m.ai
rayviljoen 🇿🇦 retweeted
Upcoming Beam integration: @openclaw .
Beam is building an OpenClaw skill that will give agents access to Beam’s programmable data movement and streaming capabilities.
Deeper integration with Beam Rooms & Tunnels is also coming, enabling agents to connect with private resources, streams, applications and compute through Beam.
This integration will also be live alongside the Beam Studio release, giving users a direct way to bring OpenClaw into Beam-powered workflows from day one.
OpenClaw → Beam → Data, Apps & Compute
b1m.ai
rayviljoen 🇿🇦 retweeted
Beam × Zapier is coming.
Beam is opening a direct bridge between Beam infrastructure and the thousands of applications already connected through Zapier.
A transfer completing, a Room event, a Tunnel coming online, or a stream starting can trigger workflows across Salesforce, Slack, HubSpot, Jira, Gmail and thousands more.
Then, with the Beam native app inside Zapier, the path goes both ways: application events can trigger Beam transfers, Rooms, Tunnels and data workflows.
Applications ↔ Zapier ↔ Beam ↔ Data / Streams / Compute
Zapier connects the workflow layer.
Beam connects the data layer.
rayviljoen 🇿🇦 retweeted
A couple days ago, we announced a major breakthrough in our multilingual language model.
Today, we're excited to share the first product built on top of it.
Introducing the Babelbit Dubbing API.
A programmable real-time dubbing solution designed for broadcasters, media companies, voice agents, and developers looking to integrate multilingual communication into their own products.
Our architecture gives us the flexibility to optimize not just for translation quality, but for everything that makes conversations feel natural. From low latency to natural voices and gender adaptation.
We created a programmable framework that allows us to continuously improve and expand its capabilities.
This is the first product built on top of our Language Transformation infrastructure.
Many more will follow.
rayviljoen 🇿🇦 retweeted
The 88 Root Validator is now registered and waiting for the chain update to enable weight setting.
Make sure your wallet is positioned for the upcoming bull market.
#SN88 #Bittensor $TAO 🚀
rayviljoen 🇿🇦 retweeted
Leadpoet V2 is coming.
We’re building an AI sales agent that sells your product for you.
This week, it starts selling Leadpoet.
It finds companies showing real signs they may be ready to buy, reaches out with relevant messaging, tracks new buying signals as they appear, and keeps working to turn qualified prospects into pipeline.
We’ll use it to identify and engage thousands of qualified leads every month.
Soon, any company will be able to give Leadpoet its product, target customers, and sales goals, then put our sales agent to work finding and engaging the right buyers.
The vision is simple: every company should be able to turn on a sales engine that learns, improves, and creates new opportunities around the clock.
rayviljoen 🇿🇦 retweeted
We’re excited to share that @Dropbox and one of its sales partners are piloting Leadpoet’s AI sales intelligence platform.
Our vision is to become the intelligence layer behind enterprise sales, helping large sales organizations identify real buying intent, prioritize the right accounts, and turn better intelligence into revenue.
This pilot brings together two critical parts of that vision: validation from a category-defining technology company and a scalable path to distribution through the partners enterprises already trust.
It is a major step toward embedding Leadpoet into the sales workflows of the world’s largest organizations.
Enterprise sales is moving from volume to intelligence. We are building the layer that powers that shift.
rayviljoen 🇿🇦 retweeted
DJINN WINS WORLD CUP ⚽️🧞♂️🏅🌎
Updates and milestones:
* Launched on Base mainnet June 6, straight into the World Cup
* 765 encrypted signals committed on-chain: 500 World Cup, 263 MLB
* Signals live on 33 of 34 World Cup match days, opening match through the final
* 716 picks purchased, up 7x since the group stage
* Every pick encrypted before kickoff, every result settled in public
rayviljoen 🇿🇦 retweeted
🚨MILESTONE SOLVED🚨
Milestone 3 of Breaking RSA (w/@terraquantum_ag) has been broken!
Congrats to 5HDyKD…NNGSNK for being the first to solve 480-bit RSA.
Their solution builds on the 460-bit GNFS solution, optimizing operations to drastically speed up candidate selection and sieving. The solution runs in just 3.8 hours, which is actually FASTER than the previous solution on the smaller 460-bit key.
More impressively, they timed their submission to go through while we were streaming QBTLL so we could announce it live 😉
The winning code is available now on our website:
qbittensorlabs.com/enigma/ch…
We'll be announcing the next milestone soon.
Onwards and upwards! 🚀
