Vexuno

Robinhood Chain tip · reading the chain

Live · training on this server

Fly brain

A model that stays in training and answers in a conversation, in the language you used. Each pass stores a new public passage, up to three full sentences in another language, the next everyday Wikipedia subject, and one page it opens on its own from English or Chinese Wikipedia. It does not open Google. Ask how to say a sentence and that wording is kept. A question it has not studied is looked up before it replies, in Chinese when you write in Chinese, and a question with nothing stored is queued for the next pass. Name a local tool and give the numbers, and the reply uses that result. The neuron map names each step and lights it only while that step is running. Shifts it can already do are sold on the earn desk. It does not hold a key and does not send USDG.

Qwen on this server · training continues in the background

Last pass 9s ago · 1,967 lessons · 14/14 lanes

Neurons

  • Sense
  • Think
  • Reply

Reference stored a passage, and Weights refit.

  • Sense
  • Think
  • Reply

Sense · Firing

Click a neuron to keep its job open.

Reference

Stores everyday Wikipedia subjects, plus any page a question opens. A Chinese question is looked up in Chinese.

350 passages on the shelf. Latest: Augustine University Ilara · 10s ago.

Conversation

Ask about anything. A subject it has not studied is looked up, then answered in this thread. A follow-up can refer to the last answer.

Teach a subject

Stores a Wikipedia extract, matching headlines, and public repositories before you ask.

Just learned

About this desk

What you use it for
Talk to it. The thread keeps the earlier turns, so a follow-up can say “that” or “the fee” and still mean the last subject. If the question names a local tool and includes the numbers that tool expects, the answer uses the computed result. Teach a subject when you want that page kept even before you ask.
What the numbers are
Each training pass stores a new public passage, stores up to three sentences in another language, stores the next everyday Wikipedia subjects, and reads the next trading-strategy page until that list is stored. A question written in Chinese is looked up on Chinese Wikipedia. A desk guide that is not on the shelf yet, including a page that was just added, is stored on the next pass. Sources include papers, definitions, earthquakes, weather, books, questions, docs, population, preprints, and strategy notes such as momentum, DCA, and the Sharpe ratio. Sentences include greetings, asking for help, the time, a price question, and a cup of water, in Mandarin, Japanese, Korean, and other languages. Asking how to say a sentence stores that sentence too. A question searches Wikipedia by its subject words, then the shelf. The reply is in the language of the question. A strategy note describes a method and is not a measured return. The neuron map lights the sense that stored the newest passage.
What it leaves out
A price, fee, or balance is used only when a source states it. If the model is offline, the reply is the source text and says so. Teaching does not spend VEXUNO. One question is answered at a time.

Every desk is introduced the same way on the desk list.

How to read this

What training does
A pass tries up to three sources and keeps the first new passage, then stores up to four everyday Wikipedia subjects such as places, science, and history. It also opens the next page on its own walk through English and Chinese Wikipedia, or a question that came back with nothing stored. A question also stores the page it looked up. The loop runs about every 45 seconds while this server is up. It does not open Google.
What an answer is
The reply is spoken the way a person would, in the language of your last message. A short question stays short. A line that starts with “How to say” is a sentence it has learned, and it uses that wording. A price, fee, or balance is stated only when a source contains it.
What it will not do
It does not invent a market figure, and it does not spend VEXUNO. If the language model is offline, the reply is only the text of the sources. Teaching a subject still fetches Wikipedia, headlines, and public repositories.