UncleCode unclecode

source	https://huggingfacetb-smol-training-playbook.hf.space/
title	The Smol Training Playbook: The Secrets to Building World-Class LLMs
type	Technical Blog Post
organization	Hugging Face
cleaned_by	Crawl4AI

The Smol Training Playbook:

Crawl4AI Cloud - QA TestFlight Program

Hello team! I've set up our internal QA testing program. Please read this carefully and follow the instructions.

Step 1: Get Access

Accept GitHub invitation (check your email)
Visit TestFlight repo: https://github.com/unclecode/crawl4ai-cloud-testflight

Maya Testing Guide

Welcome to the Maya testing team! This guide explains how to test Maya, our WhatsApp AI sales assistant for KidoCode.

Before You Start

Add Maya to Your Contacts

Save this number to your phone contacts: +60 12-319 0721

Digest Test 5 - 2026-01-03

Source: For You timeline (33 tweets) Method: Full heuristics including major AI companies, scientific papers, DSPy skip

SURFACE (12)

1. @aakashgupta

Session Handoff

Generated: 2026-01-04 Session Focus: Timeline Curation System with Preference-Based Classification Previous handoff: merged - preserved API context, added curation workflow

MANDATORY FIRST STEPS

Read PREFERENCES.md - contains all classification rules and user profile
Run git status to verify current state
Test cookie auth: uv run python tests/twitter_client/foryou.py 1
Verify bookmark folder access: see bookmark_to_folder.py

Tech Digest - January 4, 2026

Curated tweets with deep-dive analysis of linked content

1. AgentFS: A Filesystem Built for AI Agents

@penberg (Pekka Enberg)

Tech Digest - January 04, 2026

2 curated items with deep-dive analysis

1. WhatsApp REST API: Programmatic Message Sending Without Official API Costs

@tom_doerr (Tom Dörr)

	import math

	def findAnomalyPosition(n):
	bit_length = int(math.ceil(math.log(n, 2)))
	#1010 & 0011 = 0010 It means and of any number to 33 << i is returning two consequitive bits at position i and i + 1
	#XOR of n eith 2**i is toggling the bit at position i
	#so XOR n with 3 >> i is toggling bits at positions i and i + 1 and in our case
	#if bits in position i and i + 1 is 2 or 10 we want to make 01 which is this toggling with number 3
	return [i for i in range(but_length) if ((x >> i) & 3) == 2 if ((x >> i) & 3) == 2]
	def generateNext(x):

	import json
	from qwen_agent.llm import get_chat_model

	# I experimented with the 500 million parameters. It works nicely, but sometimes it doesn't detect all the parameters. For example, in the getCurrentWeather function, it may not detect the value for the second argument, which is the format of the temperature. So it's good practice to set default values for them. However, for very critical parameters, it understands them well.

	llm = get_chat_model({
	# 'model': 'qwen2:0.5b',
	'model': 'qwen2:1.5b',
	'model_server': 'http://localhost:11434/v1',
	'api_key': 'EMPTY',

	body {
	color:green;
	}