Volodymyr Truba

AI developer

[Volodymyr Truba @v_truba]

@ClementDelangue @NanoClaw_AI @huggingface 👀

[const @const_reborn]

Just orchestrated a 128 node permissionless decentralized training run, in 5 minutes, for 5 TAO, via @IOTA_SN9

They can do this up to 100B param models.
Unbelievable.

https://t.co/hJGZ6O5NrU

[Volodymyr Truba @v_truba]

@punk4663 @jbrukh @MacrocosmosAI subnet 9 of bittensor

[Macrocosmos @MacrocosmosAI]

Today, we are launching the first stage of Project Orion.

Our early pre-training run of Orion-100B achieves upward of 65% of data-center training efficiency on hardware costing a fraction of the price.

Orion-100B is the first proof point for a simple idea: that underutilized compute around the world can be turned into frontier training capacity.

We believe that this work presents, for the first time, an economically compelling case for training large models using distributed approaches.

[Volodymyr Truba @v_truba]

@theo why you don't disable them ?

[Volodymyr Truba @v_truba]

@thirteenua Круто, попробуйте https://t.co/3MpKUYQsHq допоможе вам оркеструвати багато агентів одночасно

[Volodymyr Truba @v_truba]

@tunguz claude and codex managed from hermes/openclaw.
can be done with: https://t.co/3MpKUYQsHq

[OpenAI @OpenAI]

Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946.

For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids.

An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better.

This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.

[OpenAI @OpenAI]

We’re adding new ways for people to identify AI-generated images and understand where they came from.

In addition to C2PA Content Credentials, images now also contain a SynthID watermark, and can be identified using a public verification tool to check whether an image was made by OpenAI products.

https://t.co/qo0l4vyWli

[Volodymyr Truba @v_truba]

@bcherny @rajankrx could you add copy paste field for commands returned by claude-code?

[Volodymyr Truba @v_truba]

@JR_Openheimer @DreadBong0 @tigfoundation This is price action of most of the subnets.

[Volodymyr Truba @v_truba]

@xavi3rlu @taoapp_ @cacheon_ai Always checking this summaries, probably the best way to get information about what’s happening in particular subnet

[Cacheon @cacheon_ai]

Launching Cacheon: an open, incentivized competition for LLM inference optimization.

As model quality converges, the next frontier is serving them economically at scale: lower latency, higher throughput, and lower cost per token.

Cacheon turns that problem into a live arena with continuous evaluation. Developers submit containerized inference servers, benchmarked on standardized hardware against a pinned vLLM baseline. The fastest server that preserves output correctness wins.

The goal is to make better inference systems discoverable, measurable, deployable, and rewarded in the open.

Mainnet launches by May 19. Learn more: https://t.co/JPbyJpLszq

[Volodymyr Truba @v_truba]

@AitizazK @garrytan yeah make sense, didn't try it with telegram

[Volodymyr Truba @v_truba]

@AitizazK @garrytan try this https://t.co/3MpKUYQsHq

[crux @macrocrux]

What you see are two neural networks playing against each other. Both models were trained by @Apex_SN1 miners, this was the replay of the first round of our newest competition. cool to see the models using an effective strategy called “Hamiltonian filling”.

The winning model is open sourced, which means this is now the minimum performance to beat. Many rounds to go. Let’s see if we reach SOTA performance!

[clem 🤗 @ClementDelangue]

https://t.co/bey4uOXXHk

[Brian McCrindle @mccrinbc]

Shoutout to the community for making a product on Train at Home -- A whole tool to monitor the entire TAH fleet, alpha flows, and training states! https://t.co/PI3qqDPL8P