Tudor M tudormunteanu

Here is a FizzBuzz program in the same Recognitive style, using patterns from the document: preamble, integer literals with i, property satisfaction, an evaluable entity, and an expert function.

The loop and conditional syntax below are reasonable extensions of the repeat { ... } and print(...) examples—the stub doc does not fully specify them yet.

FizzBuzz (expert + loop)

%rn 0.1
%encoding utf-8

entity FizzBuzzLine: string {

sparkX (spark-XX) — console-first bootstrap

Run these at the Spark's physical console (keyboard + display). Goal: get the cloudflared tunnel up so you can walk away from the keyboard. Everything else (steps 5-10 of new_spark.md) then runs from the Mac.

0. Sanity checks

whoami                       # expect: spark

hostname

date	2026-05-18
topic	vm-instance-presets

VM Instance Presets (Pre-installed Libraries & Models)

Summary

When creating a Lofispark instance, the user picks one named preset (e.g. "Ollama + Llama 3.1 8B") instead of always receiving the bare CUDA image. Preset tools are baked into the VM image so they cost zero install time; the matching model weights are served from a shared, operator-staged cache on the Spark so they appear in the VM instantly.

The old job of Proctor

Jordan, along with many others in various ways, has recently questioned the future of software engineers. Righfully so, given the amount of statements made by "BigAI" along with less direction observations from practitioners.

Hey, we don't often encounter systems which output code, do we? A system which can succintly and rapidlyreference all the popular programming languages, frameworks, libraries, documents.

It's only natural and healthy to ping-pong questions between our own role and the capabilities of said system. Introspection upon oneself is the first step to understanding and to building "taste", as a chef would say. Or as people explaining "vibe-coding".

What is in

"Forum of Diversity" app

convert any public thread with discussion between multiple participants into a multi-voice, multi-persona audio experience
most fun examples: Reddit thread, Twitter thread
think NotebookML but without hallucinations, summaries and using the fast spectrum of voices from EleventLabs to articulate the conversations.
UX is a screen-free experience; a raw discussion which can be listened to on your bike ride to work.

GenAudio SDK

inspired by the concept of GenUI, but for audio components
target: devs who build audio-first apps
example: multiple-choice questions, input prompts, yes/no questions

sequenceDiagram
    participant Sender as Sender's Addresso App
    participant Server as Addresso Relay Server
    participant Receiver as Receiver's Addresso App

    Note over Sender: Generate 4-digit PIN
    Note over Sender: Derive symmetric key
    Note over Sender: Encrypt shared data
 Note over Sender: Sign Addresso share

Example

https://basescan.org/tx/0x8daaf7b5269104c70ee22023f929d5a3874f3f2fca9d4367c133da6596775bf2#eventlog

decoded values:

flowRate: 32724505000
totalSenderFlowRate: -6610350010000
totalReceiverFlowRate: 32724505000

raw event data split by 64 chars:

	import dash
	from dash import html, dcc, Input, Output, State, callback, no_update
	import re

	app = dash.Dash(__name__)

	app.layout = html.Div([
	dcc.Store(id="repo-url-valid-flag", data=False),
	#
	dcc.Input(id="repo-url-input", type="text", placeholder="git@github.com:user/repo.git"),

	/ports
	/communication
	/email
	/interface.py
	```python
	import abc

	class EmailSenderAdapterInterface(abc.ABC):

	@abc.abstractmethod

	# SSH into your vm
	gcloud compute ssh --project=<project_id> --zone=<zone> <instance_name>

	# for non-boot disk
	sudo resize2fs /dev/sdb

	# Reference:
	https://cloud.google.com/compute/docs/disks/resize-persistent-disk#resize_partitions