What We Learned from a Year of Building with LLMs (Part I)
Join the O'Reilly online learning platform. Get a free trial today and find answers on the fly, or master something new and useful. Learn more It’s an exciting time to build with large language models (LLMs). Over the past year, LLMs have become “good enough” for real-world applications. The pace of improvements in LLMs, coupled with a parade of demos on social media, will fuel an estimated $200B
The Scary Thing About Automating Deploys - Slack Engineering
Most of Slack runs on a monolithic service simply called “The Webapp”. It’s big – hundreds of developers create hundreds of changes every week. Deploying at this scale is a unique challenge. When people talk about continuous deployment, they’re often thinking about deploying to systems as soon as changes are ready. They talk about microservices and 2-pizza teams (~8 people). But what does continuo
=>htmz> a low power tool for html htmz is a minimalist HTML microframework for creating interactive and modular web user interfaces with the familiar simplicity of plain HTML. [GitHub] plain🍦 Use straight up HTML. No supersets. No hz- ng- hx- v- w- x-; no special attributes. No DSLs. No <custom-elements>. Just vanilla HTML. lightweight🪶 166 bytes in total. Zero dependencies. Zero JS bundles to l
Introducing Project IDX, An Experiment to Improve Full-stack, Multiplatform App Development
Introducing Project IDX, An Experiment to Improve Full-stack, Multiplatform App Development Posted by Bre Arder, UX Research Lead, Kirupa Chinnathambi, Product Lead, Ashwin Raghav Mohan Ganesh, Engineering Lead, Erin Kidwell, Director of Engineering, and Roman Nurik, Design Lead These days, getting an app from zero to production – especially one that works well across mobile, web, and desktop plat
GraphQL is for Backend Engineers | Apollo GraphQL Blog
Most articles explaining the benefits of GraphQL focus on advantages for the frontend: things like preventing overfetching, reducing round trips, and iterating faster. But GraphQL provides just as many advantages for backend developers, which is why I choose it by default for new APIs and why you should consider it, too. Improved communication The goal of building any API is to enable someone to u
Goodbye, Node.js Buffer
The Buffer type has been the cornerstone for binary data handling in Node.js since the beginning. However, these days we have Uint8Array, which is a native JavaScript type and works cross-platform. While Buffer is an instance of Uint8Array, it introduces numerous methods that are not available in other JavaScript environments. Consequently, code leveraging Buffer-specific methods needs polyfilling
記事の中で映画、ゲーム、漫画などのネタバレが含まれているかもしれません。気になるかたは注意してお読みください。 #ネタバレ 映画の核心部分ですが、あまり語られてないので、解説してみます。 映画の中でアインシュタインが出てくるシーンは4箇所ありますが、この記事ではあくまで時系列に沿って3つに分けて記載します。 ▼1943年：核の連鎖反応についてニューメキシコ州のロスアラモスのオフィスが準備できるまでカリフォルニア州バークレーのオッペンハイマーの研究室で、マンハッタン計画は進んでいました。そこでテラーが核連鎖反応理論を提唱します。すぐに数学者たちで理論が正しいのか計算に取り組みますが、オッペンハイマーは別行動でニュージャージー州のプリンストン高等研究所のアインシュタインに会いに行きます。部外者だからこそ意見をもらう価値があると思ったからです。 OPPENHEIMER Neutron smash
Posted on Monday, April 1, 2024. Updated Wednesday, April 3, 2024. Over a period of over two years, an attacker using the name “Jia Tan” worked as a diligent, effective contributor to the xz compression library, eventually being granted commit access and maintainership. Using that access, they installed a very subtle, carefully hidden backdoor into liblzma, a part of xz that also happens to be a d
We recently launched the JavaScript Registry - JSR. It’s a new registry for JavaScript and TypeScript designed to offer a significantly better experience than npm for both package authors and users: It natively supports publishing TypeScript source code, which is used to auto-generate documentation for your package It’s secure-by-default, supporting token-less publishing from GitHub Actions and pa
Using GitHub Copilot in your IDE: Tips, tricks and best practices
AI has become an integral part of my workflow these days, and with the assistance of GitHub Copilot, I move a lot faster when I’m building a project. Having used AI tools to increase my productivity over the past year, I’ve realized that similar to learning how to use a new framework or library, we can enhance our efficiency with AI tools by learning how to best use them. In this blog post, I’ll s
TL;DRWarpStream is an Apache Kafka® protocol compatible data streaming platform built directly on top of S3. It's delivered as a single, stateless Go binary so there are no local disks to manage, no brokers to rebalance, and no ZooKeeper to operate. WarpStream is 5-10x cheaper than Kafka in the cloud because data streams directly to and from S3 instead of using inter-zone networking, which can be
In 2021 Ben Bridts published a highly inventive method for finding the AWS Account ID of a public S3 bucket. This post describes a technique to find the Account ID of any S3 bucket (both private and public). I'd highly recommend reading Ben's technique first as we will re-use a lot of concepts. S3 Bucket to AWS Account ID Shell output can be worth a thousand words, here's what our technique enable
Happy New Year everyone! In late December 2023 (Christmas Day to be precise), CPython core developer Brandt Bucher submitted a little pull-request to the Python 3.13 branch adding a JIT compiler. This change, once accepted would be one of the biggest changes to the CPython Interpreter since the Specializing Adaptive Interpreter added in Python 3.11 (which was also from Brandt along with Mark Shann
Get started with the latest updates for Dockerfile syntax (v1.7.0) | Docker
Dockerfiles are fundamental tools for developers working with Docker, serving as a blueprint for creating Docker images. These text documents contain all the commands a user could call on the command line to assemble an image. Understanding and effectively utilizing Dockerfiles can significantly streamline the development process, allowing for the automation of image creation and ensuring consiste
We want to empower you to experiment with LLM models, build your own applications, and discover untapped problem spaces. That’s why we sat down with GitHub’s Alireza Goudarzi, a senior machine learning researcher, and Albert Ziegler, a principal machine learning engineer, to discuss the emerging architecture of today’s LLMs. In this post, we’ll cover five major steps to building your own LLM app,
The New Internet
WireGuard is a registered trademark of Jason A. Donenfeld. Avery Pennarun is the CEO and co-founder of Tailscale. A version of this post was originally presented at a company all-hands. We don’t talk a lot in public about the big vision for Tailscale, why we’re really here. Usually I prefer to focus on what exists right now, and what we’re going to do in the next few months. The future can be dist
The ultimate goal of HTMX is to provide modern browser interactivity directly within HTML, without the need for JavaScript. Although relatively new, with its initial release in late 2020, this frontend library has quickly caught the attention of the IT web community. With 2nd place in the 2023 JavaScript Rising Stars “Front-end Frameworks” category (right behind React), a spot in the GitHub Accele
Introduction At the time of writing, HTTP/3 is supported by 30.4% of the top 10 million websites. This market penetration is astounding, but it seems like all of this progress has been possible almost exclusively by work on browsers, load balancers and CDN providers. What about the backend? How’s HTTP/3 doing there? The answer, sadly, is not as incredible. Because of this, I have been very interes
When Zed was open-sourced, someone on HackerNews commented that Sublime Text is faster when searching for all occurrences of the current word in a buffer. Zed takes 1s and Sublime somewhere around 200ms. Searching all occurrences means: you position your cursor over a word, you hit cmd-shift-l and all occurrences of that word in the current buffer are selected and you get a cursor at each occurren
Why, after 8 years, I still like GraphQL sometimes in the right context
A recent post, Why, after 6 years, I’m over GraphQL, made the rounds in the tech circle. The author argues that they would not recommend GraphQL anymore due to concerns like security, performance, and maintainability. In this post, I want to go over some interesting points made, and some points I think don't hold up to scrutiny. Always be Persistin' Ok, first of all, let's start with something may
I've found that one of the best ways to understand a new concept is to start from the very beginning. Start from a place where it doesn't exist yet and recreate it yourself, learning in the process not just how it works, but why it was designed the way it was. This isn't a practical guide to async, but hopefully some of the background knowledge it covers will help you think about asynchronous prob
Cloudflare is free of CAPTCHAs; Turnstile is free for everyone
Cloudflare is free of CAPTCHAs; Turnstile is free for everyone09/29/2023 This post is also available in 简体中文, 繁體中文, 日本語 and 한국어. For years, we’ve written that CAPTCHAs drive us crazy. Humans give up on CAPTCHA puzzles approximately 15% of the time and, maddeningly, CAPTCHAs are significantly easier for bots to solve than they are for humans. We’ve spent the past three and a half years working to b
Modules are published to JSR as TypeScript source code. API documentation generation, type declarations for Node-like environments, and transpilation are all handled by JSR. Module authors can focus on writing TypeScript only. Read on for more context on how and why we built JSR, how you can use it today, and how you can be involved in the project! Introducing JSR - the slightly longer versionJava
Reverend Sun Myung Moon gestures dramatically as he speaks at New York's Madison Square Garden. His chief associate, Col. Bo Hi Park, right, translates from Korean to English.Bettmann Archive/Getty Images It was 11.29 a.m. beneath pewter skies in Japan’s southern city of Nara when Shinzo Abe was handed the microphone. The nation’s former prime minister, wearing a navy blue jacket and crisp white s
WEELメディア事業部LLMリサーチャーの中田です。 ここ数日で、言語生成AIの「Groq」が話題になりました。 これにより、GPTよりも高速でテキストを生成できるんです、、、！ "GPT-3.5 class LLMs are too slow." Sure, that was true last week. Here is Groq (not the same as Musk's Grok) running Llama 2. Watch for the moment I click send. If you want to try: https://t.co/aZKkWVsamS pic.twitter.com/gOlB4cL5MC — Ethan Mollick (@emollick) February 19, 2024 XでのGroq関連の投稿のいいね数は、すで1100を超えており、
Today we’re excited to announce the release of TypeScript 5.2! If you’re not familiar with TypeScript, it’s a language that builds on top of JavaScript by making it possible to declare and describe types. Writing types in our code allows us to explain intent and have other tools check our code to catch mistakes like typos, issues with null and undefined, and more. Types also power TypeScript’s edi
Rust の hyper は何が嬉しいか
Rust でWebサーバーを書く時の技術選定をするときに調べていると hyper に必ず出会うと思う。これは黎明期から存在しているライブラリで、Webサーバーにしては珍しく version 1 まで到達している老舗だ(1に到達してたら安心って考え方が正しいかはさておき...)。このライブラリは actix-web や axum のような他のライブラリとは毛色が違い、かなり primitive だ。そのため axum のベースに使われてもいて、hyper はそのまま使わないライブラリなのかもしれない。 サンプルコードから存在意義がわかりにくい さて、そんな hyper だが公式の example はこのようになっている。 #[tokio::main] async fn main() -> Result<(), Box<dyn std::error::Error + Send + Sync>>
Note: This blog post is also available as a documentation page on Transformers. Large Language Models (LLMs) such as GPT3/4, Falcon, and LLama are rapidly advancing in their ability to tackle human-centric tasks, establishing themselves as essential tools in modern knowledge-based industries. Deploying these models in real-world tasks remains challenging, however: To exhibit near-human text unders
Elon Musk’s Shadow Rule
“We are living off his good graces,” a Pentagon official said of Musk’s role in the war in Ukraine. “That sucks.”Photo illustration by Matt Chase; Source photographs from Getty; Shutterstock Last October, Colin Kahl, then the Under-Secretary of Defense for Policy at the Pentagon, sat in a hotel in Paris and prepared to make a call to avert disaster in Ukraine. A staffer handed him an iPhone—in par
Source:imgflip.comAs I write this on the 25th of April, I am still reeling from the announcement of IBM’s acquisition of Hashicorp. When I first heard the rumours yesterday, I was concerned about the future of possibly my favourite Infrastructure-as-code (IaC) tool. It has long been obvious that Hashicorp has been struggling to make money, making a $274 million loss in 2023. This undoubtedly led t
HTTP/2 Rapid Reset: deconstructing the record-breaking attack
HTTP/2 Rapid Reset: deconstructing the record-breaking attack10/10/2023 This post is also available in 简体中文, 繁體中文, 日本語, 한국어, Deutsch, Français and Español. Starting on Aug 25, 2023, we started to notice some unusually big HTTP attacks hitting many of our customers. These attacks were detected and mitigated by our automated DDoS system. It was not long however, before they started to reach record b
The first step in building my text editor is to implement the core API. If you’re wondering why I want to do this, the original article is here. I researched several data types, and I tried to be language agnostic. I wanted my decision to not be influenced by any particular language, and first see if there was a “best way” out there, solely based on operations. Of course, a “best way” rarely exist
Don’t Build Useless Features
A guide to scaling product & engineering teams from $0 to past $100M ARR. © 2024. Stay SaaSy. As a product manager, it’s important to hone the minimum set of activities that allow you to keep a product line moving forward productively. One of the most important core product management skills: the ability to triage unsuccessful products and avoid spending unnecessary effort on products that are des
Using Zig in our incremental Turborepo migration from Go to Rust – Vercel
Using Zig in our incremental Turborepo migration from Go to RustLearn how we're cross-compiling a Rust-Go-Rust sandwich to six platforms. We’ve been porting Turborepo, the high-performance build system for JavaScript and TypeScript, from Go to Rust. We talked about how we started the porting process, so now let’s talk about how we began porting our two main commands: run and prune. Since last time
We've added JavaScript-native RPC to Cloudflare Workers
We've added JavaScript-native RPC to Cloudflare Workers04/05/2024 Cloudflare Workers now features a built-in RPC (Remote Procedure Call) system enabling seamless Worker-to-Worker and Worker-to-Durable Object communication, with almost no boilerplate. You just define a class: export class MyService extends WorkerEntrypoint { sum(a, b) { return a + b; } } And then you call it: let three = await env.
HTTP/2 Zero-Day vulnerability results in record-breaking DDoS attacks
HTTP/2 Zero-Day vulnerability results in record-breaking DDoS attacks Loading... This post is also available in 简体中文, 繁體中文, 日本語, 한국어, Deutsch, Français and Español. Earlier today, Cloudflare, along with Google and Amazon AWS, disclosed the existence of a novel zero-day vulnerability dubbed the “HTTP/2 Rapid Reset” attack. This attack exploits a weakness in the HTTP/2 protocol to generate enormous,
Piccolo - A Stackless Lua Interpreter 2024-05-01 History of piccolo A "Stackless" Interpreter Design Benefits of Stackless Cancellation Pre-emptive Concurrency Fuel, Pacing, and Custom Scheduling "Symmetric" Coroutines and coroutine.yieldto The "Big Lie" Rust Coroutines, Lua Coroutines, and Snarfing Zooming Out piccolo is an interpreter for the Lua language written in pure, mostly safe Rust with a
Highlights from Git 2.44
Open SourceHighlights from Git 2.44The first Git release of 2024 is here! Take a look at some of our highlights on what's new in Git 2.44. The open source Git project just released Git 2.44 with features and bug fixes from over 85 contributors, 34 of them new. We last caught up with you on the latest in Git back when 2.43 was released. To celebrate this most recent release, here is GitHub’s look a
Every so often I read an essay that I end up thinking about, and citing in conversation, over and over again. Here’s my index of all the ones of those I can remember! I’ll try to keep it up to date as I think of more. There's a lot in here! If you'd like, I can email you one essay per week, so you have more time to digest each one: Nelson Elhage, Computers can be understood. The attitude embodied
Back to articlesOnyx, a new programming language powered by WebAssemblyLearn about Onyx, a new imperative programming language that leverages WebAssembly and Wasmer for seamless cross-platform support What is Onyx? Onyx is a new programming language featuring a modern, expressive syntax, strict type safety, blazingly-fast build times, and out-of-the-box cross platform support thanks to WebAssembly
j次のブックマーク
k前のブックマーク
lあとで読む
eコメント一覧を開く
oページを開く
htmz - a low power tool for html
【オッペンハイマー】徹底解説：アインシュタインとの会話｜まいるず
research!rsc: Timeline of the xz open source attack
How we built JSR
Kafka is dead, long live Kafka
How to find the AWS Account ID of any S3 Bucket
Python 3.13 gets a JIT
The architecture of today's LLM applications
HTMX vs React: A Complete Comparison - Semaphore
gRPC Over HTTP/3
From 1s to 4ms
Learning Async Rust With Entirely Too Many Web Servers
Introducing JSR - the JavaScript Registry
The Unification Church's Global Grasp for Power
【Groq】一秒で500トークン、GPT-4の25倍のスピードで出力できるAIを使ってみた | WEEL
Announcing TypeScript 5.2 - TypeScript
Optimizing your LLM in production
The beginning of the end for Terraform?
Text Editor: Data Structures
kyju.org - Piccolo - A Stackless Lua Interpreter
Essays on programming I think about a lot
Onyx, a new programming language powered by WebAssembly