Board

Board

2272 bookmarks
Newest
How Kubernetes Broke Git
How Kubernetes Broke Git
How did Kubernetes push git to its limit? Some stories from my time (2016-2019) working on Kubernetes. No atomicity across subprojects – In 2016, Kubernetes was still a monorepo. Everything was developed in a single repository. This meant that developers could reuse CI infrastructure easily and ensure that changes to the different components (e.g., kube-proxy, kube-apiserver, kubelet) would work together. However, downstream projects needed to build on the API. That meant vendoring parts of Ku
·matt-rickard.com·
How Kubernetes Broke Git
What Comes After Git
What Comes After Git
Git was born from the collaboration problems in the Linux kernel. Nearly a decade later, new problems arose when Kubernetes (the operating system of the cloud) brought open-source collaboration to a new level. I saw the pain points of git (and GitHub) firsthand working on Kubernetes open-source. Will a new version control system (or something that solves similar problems) spring up? Some ideas on what a new version control system would look like. * Atomicity across projects –  GitHub is a de
·matt-rickard.com·
What Comes After Git
Shannon's Demon
Shannon's Demon
Or, how you can create positive returns out of a random walk simply by rebalancing a portfolio. Let's say you have $100 that you want to distribute 50% in stocks and 50% in cash. $200 in stocks, $200 in cash. On the first day, the stock halves in price, so you now only have $100 in stock. You rebalance your portfolio by buying another $50 in stock, bringing your assets to $150 in stock, $150 in cash (total $300). The next day, the stock doubles in price. Your stock is now worth $300. You sell
·matt-rickard.com·
Shannon's Demon
IaC: Strength or Weakness for Cloud Providers?
IaC: Strength or Weakness for Cloud Providers?
Infrastructure as code (IaC) will change the way that we consume infrastructure from cloud providers. IaC is a win for customers, but will it be long-term strategic for the cloud providers themselves? Or is it the start of the commoditization and abstraction of the cloud layer? A wedge for new entrants to compete on? * IaC turns cloud infrastructure from a GUI to an API layer. I believe this also changes the end-user of many of these services, disintermediating many purely operational roles (e
·matt-rickard.com·
IaC: Strength or Weakness for Cloud Providers?
The Story of CP/M
The Story of CP/M
Could microprocessors run full computers? This question seems obvious today, but in 1974, this was Gary Kidall's insight behind developing the operating system CP/M. It was an OS specifically built for small computers. Written in Kidall's programming language, PL/M. Kidall sold CP/M licenses through his company, Digital Research. CP/M was special because it separated the physical I/O system (now called BIOS – Basic I/O System) from the disk operating system (at the time, called the BDOS – Basic
·matt-rickard.com·
The Story of CP/M
Hotelling's Law and Differentiation
Hotelling's Law and Differentiation
Imagine there are two ice cream vendors at the beach. They sell exactly the same flavors at exactly the same price. Beachgoers are indifferent, so they walk to whatever vendor is closer. Where should each ice cream vendor park their stall to maximize the number of customers? One solution might be for each vendor to serve half of the beach, situating themselves 1/4 from the end (so that nobody has to walk more than 1/4 a mile). ----A----|----B---- This might be the "socially optimal" solution
·matt-rickard.com·
Hotelling's Law and Differentiation
Distribution in a Downturn
Distribution in a Downturn
When money is cheap, distribution is expensive. In the last two years, we saw companies look for distribution advantages: * Newsletter writers and podcast hosts raised venture funds (Packy McCormick's Not Boring Capital, Mario Gabriele's Generalist Capital, Harry Stebbings' 20VC, Sahil Bloom's SRB Fund) * Companies proprietary software advantages for distribution by open-sourcing more of their core product * Consumer social apps like TikTok and Instagram paid their creators just for posting
·matt-rickard.com·
Distribution in a Downturn
The Remix IDE
The Remix IDE
If you're deploying applications on Ethereum, you might use the web-based Remix IDE. It bundles a working set of the different tools you need to write Solidity code, deploy it to a test environment, debug it, and eventually run it in production. Remix might be one of the first times a niche IDE has emerged and started browser-first. Some technicals: First, you can find the open-source code on GitHub here. It is Monaco-based (the same editor that powers VSCode). It uses its own plugin system ra
·matt-rickard.com·
The Remix IDE
SSH: Less Relevant In the Cloud
SSH: Less Relevant In the Cloud
The Lindy Effect is a heuristic that the future life expectancy of non-perishable entities is proportional to their age. The longer something has already lasted, the higher likelihood it will continue to endure. – Observations of the Lindy Effect Secure Shell or SSH has been the de facto way to connect to a server remotely. It's been around since 1995. It's simple on the surface – it uses public-key cryptography (or a password) to authenticate on a remote server. The connection protocol can ope
·matt-rickard.com·
SSH: Less Relevant In the Cloud
So Easy You Could Run It Yourself
So Easy You Could Run It Yourself
What happens when software becomes so easy to deploy that you could run it yourself? For instance, this blog could be reduced to probably ~100 lines of reusable AWS CDK that anyone could deploy. You wouldn't really have to do much maintenance – the static files are hosted on a CDN, and the dynamic parts have a small surface area. There are enough serverless cloud services to make sure you really only need to worry about application-level errors – not the mail server going down (AWS SES) or the
·matt-rickard.com·
So Easy You Could Run It Yourself
Commoditization of Large Language Models
Commoditization of Large Language Models
GPT-3 ushered in a new era of large language models (LLMs) that could generate human-realistic text output. But GPT-3 didn't come from a company with a large and proprietary dataset. Instead, the dataset consisted of: * 410 billion tokens from the public Common Crawl (60% weight) * 19 billion tokens from Reddit submissions with a minimum score of 3 (22%) * 12 billion tokens from "Books1" and 55 billion from "Books2", which are probably books downloaded from the Library Genesis archive (a pir
·matt-rickard.com·
Commoditization of Large Language Models
A Look Into My Development Stack
A Look Into My Development Stack
I've built a lot of developer tools (minikube and skaffold to name a few). But I'm often asked – what does my personal development stack look like? Hardware: The M1 Macbooks are fast. At the same time, I miss my Arch Linux and i3 tiling window manager, my day-to-day consists of a lot more than just coding (emails, spreadsheets, docs, calls). I don't want to spend time maintaining my snowflake Linux distro (although I keep around an Arch server in my closet with a few GPUs for training the occas
·matt-rickard.com·
A Look Into My Development Stack
Giving Up Decentralization: Scalability Trilemma
Giving Up Decentralization: Scalability Trilemma
The scalability trilemma states that blockchains must choose to optimize two of the three properties: * Scalability –transactions per second (throughput) and transaction finality (latency) * Security – the cost to mount of a 51% attack, cost to validate state * Decentralization – the cost and ability to run a node or participate in the network Many of the web3 call themselves decentralized de jure – the network must be decentralized due to the rules of the underlying technology: built on bl
·matt-rickard.com·
Giving Up Decentralization: Scalability Trilemma
Does Financial Policy Matter?
Does Financial Policy Matter?
You better cut the pizza into four slices because I'm not hungry enough to eat six – Yogi Berra Does it matter if a firm is financed with debt, equity, or a digital token? Does a firm's financial policy create value? The Miller-Mogdiliani ("MM" for short) Theorem1 says no – under certain conditions. The conditions for the MM theorem are: * Perfect markets – no arbitrage or no transaction costs * No taxes * No additional costs to bankruptcy * No conflicts of interests * Symmetric informat
·matt-rickard.com·
Does Financial Policy Matter?
History of Version Control Systems: Part 3
History of Version Control Systems: Part 3
Part I: First-generation Part II: Client/Server Part III: Distributed My hatred of CVS means that I see SVN as the most pointless project ever started. — Linus Torvalds, creator of the Linux Kernel and Git The third generation of VCS was distributed. It's best to describe it through the story of Git. Larry McVoy had worked on a VCS called Sun WorkShop TeamWare in the 90s. TeamWare mirrored many of the features of Subversion and Perforce but built on SCCS. In 1998, McVoy saw the issues with th
·matt-rickard.com·
History of Version Control Systems: Part 3
Panic of 1907/2022
Panic of 1907/2022
The players are different, but the game is the same. F. Augustus Heinze was the founder of United Copper Company and one of the "Three Copper Kings" of Butte, Montana. United Copper produced 40 million pounds of copper a year. By 1906, Heinze was rich and set his sights on the financial markets. Heinze had two brothers, Otto and Arthur, who devised a "short squeeze," not unlike the one that happened to GameStop in January 2021. Otto had realized that United Copper had 105% short interest – i.e
·matt-rickard.com·
Panic of 1907/2022
How to Use GitHub Copilot Effectively
How to Use GitHub Copilot Effectively
I've been using GitHub Copilot for about a year now. I was initially skeptical. Are the suggestions actually good? Don't I still need to read the code for correctness? Isn't it distracting? Now, I couldn't imagine coding without it. For $10/mo, it's a great deal. I've never been more productive. But like any tool, you need to know how to use it. Here are five tips on how to use GitHub Copilot effectively. The more consistent you code is, the better Copilot's suggestions will be. Copilot is
·matt-rickard.com·
How to Use GitHub Copilot Effectively
Least Common Denominator APIs
Least Common Denominator APIs
Often, our instinct is to build for optionality. What if we change databases? What if we change clouds? We target the Least Common Denominator (LCD) interface to avoid vendor lock-in and make sure our software is portable – after all, Optimization is Fragile. LCD interfaces look like targeting the S3 API, a generic PubSub implementation, or vanilla ANSI SQL. LCD interfaces are good enough most of the time, but when we need to run a specialized workload, sometimes they don't perform how we'd lik
·matt-rickard.com·
Least Common Denominator APIs
Every Sufficiently Advanced Configuration Language is Wrong
Every Sufficiently Advanced Configuration Language is Wrong
Every sufficiently advanced configuration language is the wrong tool for the job. For basic configuration, YAML or JSON is usually good enough. It falls apart when you try to do more: * Template it with a templating engine * Use esoteric language features to reuse code (anchors and aliases) * Patch or modify it with something like JSONPatch * Type-check or schema validate These are anti-patterns and often cause more issues than they solve. So instead, we develop more advanced configuratio
·matt-rickard.com·
Every Sufficiently Advanced Configuration Language is Wrong
Is Snowflake a Platform?
Is Snowflake a Platform?
Salesforce built one of the world's most successful platforms around CRM data. Can Snowflake do the same for analytics? For platforms, I use the Gates definition, A platform is when the economic value of everybody that uses it exceeds the value of the company that creates it. Applications are already being built on Snowflake, powering the so-called Modern Data Stack. A cloud data warehouse enables the next generation of business intelligence, ETL, and workflow orchestration. At least the priv
·matt-rickard.com·
Is Snowflake a Platform?
The Power of Plaintext Protocols
The Power of Plaintext Protocols
HTTP is a plaintext protocol. Sending an HTTP request is as simple as writing GET / HTTP/1.1 Host: matt-rickard.com If you want to see for yourself, you can run this command on the command line to send the raw HTTP request and see the result. echo -en 'GET / HTTP/1.1\r\nHost: matt-rickard.com\r\n\r\n' | openssl s_client -ign_eof -connect matt-rickard.com:443 The response is plaintext as well HTTP/1.1 200 OK Connection: keep-alive Content-Length: 20729 Content-Type: text/html; charset=utf-8
·matt-rickard.com·
The Power of Plaintext Protocols
SSI News Digest
SSI News Digest
Cloudflare One Week in Review SSI Blog Post Digest Cloudflare One Week in Review Cloudflare held another innovation week from June 20 – 24, this time focused on Zero Trust solutions. Called Cloudflare One Week, the team introduced a number of new capabilities and aligned their marketing message around what is expected for a full-featured Zero Trust SASE (Secure Access Service Edge) platform.
·feedly.com·
SSI News Digest
Adding Top Down Sales: The “$20M to $500M” Question
Adding Top Down Sales: The “$20M to $500M” Question
For many enterprise startups, the predominant initial go to market (GTM) strategy has been “growth + sales”, which relies on a bottom up, product-led approach to acquiring and retaining users and customers. But after successfully executing on this …
·future.com·
Adding Top Down Sales: The “$20M to $500M” Question
The New Business of AI vs Traditional Software
The New Business of AI vs Traditional Software
At a technical level, artificial intelligence seems to be the future of software. AI is showing remarkable progress on a range of difficult computer science problems, and the job of software developers – who now work with data as much …
·future.com·
The New Business of AI vs Traditional Software
The Full-Stack Startup | Future
The Full-Stack Startup | Future
Q: So what’s a full stack startup? You’ve mentioned that it’s a new, important trend, and a pattern of startups we’ve been seeing over the past couple of years. Chris Dixon: The old approach startups took was to sell or …
·future.com·
The Full-Stack Startup | Future