BetterDev #259 - How LLMs Work, Explained Without Math and Turning AirPods into a Fitness Tracker to Fight Cancer
Better Dev #259 May 13, 2024
Hi all,
We come back with a new issue this week. If you like BetterDev, please help spead word out by refer to your friends. Buy me a coffee would be great too.
Many link about LLM this week, especially focus on fundamental knowledge and using them locally on your own hardware, instead of just a wrapper of OpenAI.
A fun side project for a great cause featuring Core Motion, SwiftUI, a little help from AI, and a pair of AirPods to count 100 push-ups a day.
A basic question I think a lot of people have about the GenAI revolution is where does the apparent intelligence these models have come from. In this article, I’m going to attempt to explain in simple terms and without using advanced math how generative text models work, to help you think about them as computer algorithms and not as magic.
walk through the process of fine-tuning step-by-step. It won’t cost you a penny because we’re going to do it all on your own hardware using Apple’s MLX framework.
The first thing I check when debugging latency issues in distributed systems is whether TCP_NODELAY is enabled. And it’s not just me. Every distributed system builder I know has lost hours to latency issues quickly fixed by enabling this simple socket option, suggesting that the default behavior is wrong, and perhaps that the whole concept is outmoded.
Using time for animation is very common in shader programs and various other places. CSS can not start a timer like JavaScript does, but nowadays it’s possible to define a custom variable with the CSS Houdini API to track time in milliseconds. The value of animation is calculate based on this custom variable. It’s a different approach to perform animation compare to keyframes. This goes beyond my ability to understand, I’m not super good at CSS animation. But still a good read to dabble, I always like to perform animation with pure css.
One of the surprises that comes with developing applications and operating a database cluster behind them is the discrepancy between practice and theory, development environment and the production. A perfect example of such a mismatch is changing a column type. Super easy to change stuff in dev, but I dare you to randomly go in a prod db and change data type
When you generate SSH key like RSA or ECDSA and peform encryption and signature verification. What is the math behind them? I always though they are complicated, but they are way more straighforward when you dabble into its math.
In database, there is always a trade off. You trade storage tooling gain cpu time. trade fragment for speed. all of that become bloated. How can we minimize these?
discusses how to develop eBPF programs with Go, with practical example at Edgedelata. A bit Go specific but learning about eBPF will boost your system programming knowledge.
We’re going to speed up some numpy code by 100x using “unsafe Python.” Which is not quite the same as unsafe Rust, but it’s a bit similar, and I’m not sure what else to call it… Python specific but please read-on, I’m sure the same concept can be apply to any language that has a C-binding to native C library.
Code to read
an API created by LlamaIndex to efficiently parse and represent files for efficient retrieval and context augmentation using LlamaIndex frameworks.
PythonPretty fancy and modern terminal file manager. Even though I cannot see myself use this tool, I absolutely want to dig into its source code and see how it’s done. One thing I always wanted to understand about the event/gui loop in these GUI/TUI app.
Goa minimal TLS implement that is well enough to make a simple HTTPS request to most web servers. Very nice to understand what happen at the byte level in a TLS session
BashTools
An extensible text-to-diagrams library that works in both browser and node.js. Similar to tool like MermaidJS to draw diagram with text.
Fully private LLM chatbot that runs entirely with a browser with no server needed. Supports Mistral and LLama 3
a terminal-based monitoring tool “top” designed to display real-time metrics for Apple Silicon chips. It provides a simple and efficient way to monitor CPU and GPU usage, E-Cores and P-Cores, power consumption, and other system metrics directly from your terminal!
A collection features over 3,800 stroke React icons, available at no charge and crafted to meet a diverse array of design requirements.
not monitoring pgslot closely and if a consumer die, you may wake up with a full disk of pgslot WAL. This small tool monitors their growth and alert via slack
ChatGPT at home! Basically a better Google Nest Hub or Amazon Alexa home assistant. Built on the Raspberry Pi using the OpenAI API. A very cool hardware project with DIY style
Self Hosted
A Mail server with a built-in admin ui. Replace all the tradition tooling like Postfix, DoveCot, roundcube or rainloop etc.
a modular and customizable javascript package that directly brings language model chats directly onto web browsers with hardware acceleration. Everything runs inside the browser with no server support and is accelerated with WebGPU.
an observability platform built by Pydantic team. Yes, it does has a web ui to help you view log and meric.
an automated email assistant customized to your writing style and past email history. You can run in on your laptop, train it with your mbox data. The mbox data format can be easily export from thunderbird or mac mail app or many modern mail client.
You can view this issue in web browser.
If you have any suggestion/feedback, do tell me by replying to this email. I read them all.
No longer want to receive these emails? Unsubscribe
Older messages
BetterDev #258 - Build an 8-bit computer from scratch and Home automation with ESP8266
Wednesday, May 1, 2024
Better Dev #258 Apr 30, 2024 Hi all, We come back with a new issue this week. If you like BetterDev, please help spead word out by refer to your friends. Buy me a coffee would be great too. Build an 8-
BetterDev #257 - Building a GPS Receiver
Monday, April 22, 2024
Better Dev #257 Apr 22, 2024 Hi all, We come back with a new issue this week. If you like BetterDev, please help spead word out by refer to your friends. Buy me a coffee would be great too. This week I
BetterDev #256 - How to write a code formatter and a search engine in 80 lines of Python
Tuesday, April 16, 2024
Better Dev #256 Apr 15, 2024 Hi all, We come back with a new issue this week. If you like BetterDev, please help spead word out by refer to your friends. Buy Me a coffee would be great too. This issue
BetterDev #255 - What even is a JSON number
Monday, April 8, 2024
Better Dev #255 Apr 8, 2024 Hi all, We come back with a new issue this week. If you like BetterDev, please help spead word out by refer to your friends. Buy Me a coffee would be great too. This week, I
BetterDev #254 - Garbage Collection for Systems Programmers and How radio work
Monday, April 1, 2024
Better Dev #254 Apr 1, 2024 Hi all, We come back to my second week :). If you like BetterDev, please help spead word out by refer to your friends. Buy Me a coffee would be great too In previous email,
You Might Also Like
Import AI 399: 1,000 samples to make a reasoning model; DeepSeek proliferation; Apple's self-driving car simulator
Friday, February 14, 2025
What came before the golem? ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
Defining Your Paranoia Level: Navigating Change Without the Overkill
Friday, February 14, 2025
We've all been there: trying to learn something new, only to find our old habits holding us back. We discussed today how our gut feelings about solving problems can sometimes be our own worst enemy
5 ways AI can help with taxes 🪄
Friday, February 14, 2025
Remotely control an iPhone; 💸 50+ early Presidents' Day deals -- ZDNET ZDNET Tech Today - US February 10, 2025 5 ways AI can help you with your taxes (and what not to use it for) 5 ways AI can help
Recurring Automations + Secret Updates
Friday, February 14, 2025
Smarter automations, better templates, and hidden updates to explore 👀 ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
The First Provable AI-Proof Game: Introducing Butterfly Wings 4
Friday, February 14, 2025
Top Tech Content sent at Noon! Boost Your Article on HackerNoon for $159.99! Read this email in your browser How are you, @newsletterest1? undefined The Market Today #01 Instagram (Meta) 714.52 -0.32%
GCP Newsletter #437
Friday, February 14, 2025
Welcome to issue #437 February 10th, 2025 News BigQuery Cloud Marketplace Official Blog Partners BigQuery datasets now available on Google Cloud Marketplace - Google Cloud Marketplace now offers
Charted | The 1%'s Share of U.S. Wealth Over Time (1989-2024) 💰
Friday, February 14, 2025
Discover how the share of US wealth held by the top 1% has evolved from 1989 to 2024 in this infographic. View Online | Subscribe | Download Our App Download our app to see thousands of new charts from
The Great Social Media Diaspora & Tapestry is here
Friday, February 14, 2025
Apple introduces new app called 'Apple Invites', The Iconfactory launches Tapestry, beyond the traditional portfolio, and more in this week's issue of Creativerly. Creativerly The Great
Daily Coding Problem: Problem #1689 [Medium]
Friday, February 14, 2025
Daily Coding Problem Good morning! Here's your coding interview problem for today. This problem was asked by Google. Given a linked list, sort it in O(n log n) time and constant space. For example,
📧 Stop Conflating CQRS and MediatR
Friday, February 14, 2025
Stop Conflating CQRS and MediatR Read on: my website / Read time: 4 minutes The .NET Weekly is brought to you by: Step right up to the Generative AI Use Cases Repository! See how MongoDB powers your