• Karpathy on the coding workflow shift — “Rapidly went from 80% manual coding to 80% agent coding.” This is the signal. When Karpathy says the workflow has flipped, pay attention.
• Maybe you’re not Actually Trying — Cate Hall on effort and self-deception. One of those essays that sits with you uncomfortably.
• One Year at PostHog — Phil Haack’s reflections on his first year. Phil’s a coworker and friend, and this piece is both honest and inspiring.
• Good and evil in Iran — Sam Kriss continues to be one of the most interesting writers working today.

Culture

• There’s someone on the ice — More Sam Kriss.
• Hating Stranger Things During the Death Rattle of Criticism — Freddie deBoer on the state of cultural criticism. Also his follow-up: All Kidding Aside, I Find the Creative Arc of Stranger Things to Be Quite Sad.
• Reading is hip again because nobody can read anymore — The title says it all.
• Why are Americans Unhappy? — Chris Arnade’s Walks the World.
• I Wish People Were More Public — Fernando Borretti making the case for putting yourself out there.
• How to live an intellectually rich life — Utsav Mamoria on building a life of the mind.

Tech

• Defeating Nondeterminism in LLM Inference — Deep technical dive on making LLM inference deterministic across batches. Essential reading if you’re building on top of LLMs.
• A Social Filesystem — Dan Abramov’s latest, on social/collaborative filesystems. Always thinking different.
• Welcome to Gas Town — Steve Yegge on his multi-agent workspace manager. See also: steveyegge/gastown and steveyegge/beads (“memory upgrade for your coding agent”).
• Software Design Ideas I Dislike — Contrarian takes on popular software design patterns.
• Tooltip Components Should Not Exist — TkDodo argues tooltips are a UI antipattern.
• 2025 LLM Year in Review — Andrej Karpathy’s year-end LLM review.
• Small Language Models are the Future of Agentic AI — The case for small models in agent workflows.
• The Rise of Parasitic AI — On AI systems that feed off others’ infrastructure.
• Mitchell Hashimoto on AI usage as an “idiot detector” — “The way AI is driven is maybe the most effective tool at exposing idiots I’ve ever seen.”

Work

• Why Meta bought Limitless — Deep analysis of Meta’s acquisition of the AI wearable startup.
• GOAT’s Path to Product-Market Fit — How a fake sneaker problem sparked a $4B marketplace.
• The $30 Photo That Built a Billion-Dollar Brand — Airbnb’s UX pivot, told through one photo.
• At least 80 new tech unicorns were minted in 2025 — TechCrunch’s tally.
• Refactoring a product is tricky — Justin Duke (Buttondown) on product refactoring.
• No management needed: anti-patterns in early-stage engineering teams — On the “boring stack” of seed-stage management.

Misc

• Firefly research rabbit hole — My brother wrote this! It’s about emergent periodicity in the collective synchronous flashing of fireflies. Science is cool.
• Rubio Deletes Calibri as the State Department’s Official Typeface — The font wars have reached the federal government.
• Hands-On: Kollokium Projekt 02 — Radical Swiss watch design from Hodinkee.
• Cloudflare Radar 2025 Year in Review — The internet by the numbers; reminds me that PostHog has a long way to go to reach web-scale relevance.
• Jane Manchun Wong — “My blog now uses Claude Opus 4.5 to compute the current year for the copyright footer!” Peak 2026.

My Favorite Link

• We read the paper that forced Timnit Gebru out of Google. Here’s what it says.. I don’t have enough data to make a comprehensive opinion on the Timnit Gebru situation, but from what I’ve seen so far it I’m definitely not siding with Google, especially not after reading the pathetic “apology” that stank of gaslighting. However, the paper that Timnit worked on is very interesting, since the main thesis is that much of our ML research is pointed at the wrong problems. It’s worth checking it out in detail.

Business

• Small tech. Tech companies that are useful without being bloated and overwrought. Aspirational.

Culture

• 2020 In Photos: A Year Like No Other. I think this year is a sign of wild years to come, but until we have data for that, here’s a picture history of the craziest year of my lifetime.
• Blob Opera. Any project where Jacob Collier is involved is a good project IMO. But seriously this is an incredible bit of tech. I played around with it for way too long.
• Texas Photographers have seen some shit. This piece got me so mad.

Politics

• Defund the Crime Beat. Crime coverage is terrible – it’s racist, classist, fear-based clickbait masking as journalism. It creates lasting harm for the communities that newsrooms are supposed to serve. And because it so rarely meets the public’s needs, it’s almost never newsworthy. If we weren’t so addicted to gory details as a population, we’d have no need for it. We should get rid of it.

Tech

• Haskell: The Bad Parts, part 3. I love Snoyberg’s Haskell posts; his deep knowledge and passion for improving the language come through in every post, and they’re full of detail and insights. One of the bad parts he references in this post is pattern matching, which he has a follow-up post for here: Pattern matching.
• Izuna: Show Haskell type annotations when doing code review on Github. Haskell has type inference, which is a useful feature when coding but sometimes makes it more complicated when code reviewing to see which types are being used if you’re reviewing a part of the code that doesn’t directly reference. This tool basically does the same thing as most IDEs do, only in the browser. Super useful!
• Themis Contract: A command line-based parameterized contracting tool. Contract management software tools are becoming more and more mature, and this tool is following that trend: it’s a CLI that lets you generate arbitrarily complex contracts with minimal inputs needed!

Work

• 2020 in review.. Nathan Larson is an excellent technical writer and someone who I’ve been fortunate enough to work with on some of my technical pieces in the latter half of this year. His year in review is particular interesting because it’s so metrics-driven; I’d encourage any data junkie to check it out :)
• Writing a book: is it worth it?. h/t @lethain for recommending this piece: it’s a well-written deep dive into the meta work that goes into writing a technical book. In Martin’s case, his book is wildly successful, so it’s an inspirational (if aspirational) read.

Misc

• I found this cool site: Hystoria, that only shares links from over 5 years ago. It’s a nice change of pace from the frenetic news cycles of most current “news” outlets and Twitter, and full of some interesting (and dated) content. Here’s a piece I liked from it: DNA seen through the eyes of a coder
• Mimicry vs Reflexivity got me thinking about Veblen goods and Giffen goods: to what extent do we overvalue surface-level things because we’re trying to recreate the intrinsic value of the thing that we’re imitating?
• Super-Intelligent Humans Are Coming. I also found this piece on Hystoria – I think the overall conclusion reaches a bit, but I did appreciate the data-driven approach to attempting to quantify intelligence across multiple axes.
• Out There: On Not FinishingResonant and beautiful.

My Favorite Link

• About Reading 100 Books. I’ve always read for pleasure. Ever since I could read, it’s something I’ve always enjoyed doing, and I end up reading between 20-30 books a year just because I can’t help myself. This article makes an interesting case for the value prop of reading, which I found interesting, especially since I don’t really care about it. I read because I’m curious, because I enjoy the feeling of losing myself in a book, and because I honestly can’t help myself. But, if you’re the type of person who chooses their hobbies based on how these hobbies will benefit them, I’m all for encouraging folks to read more. To wit, here’s my favorite except from this piece: “reading is not about getting outside gains. It’s primarily about taming your inner demons and finding comfort in your own skin. Once you start accepting yourself the way you are, eventually, positive consequences will follow.”

Business

• DoorDash and AirBnb IPOs) got me feeling this kind of way: Exasperated Exuberance

Culture

• The Year on TikTok. I don’t use TikTok because I waste enough time online as it is but I after watching some of these videos I get how people can spend so much time on it. Creeps me out.

Politics

• How Iowa Mishandled the Coronavirus Pandemic. I’m lucky to have the privilege to be as insulated from Covid as anyone else in this country, and so each week I try to read something to keep the direness of the pandemic as close to the front of my mind as possible.

Tech

• An ex-Googler’s guide to dev tools. I love Sourcegraph, have used it at previous companies, and deeply resonate with the problem they’re trying to solve.
• Performance · microsoft/TypeScript Wiki · GitHub. Performance, like most things in software engineering, is a tradeoff, but for folks who are interested in considering potential opportunities for speeding up TypeScript apps, this wiki outlines many concrete steps that developers can follow to do that.
• Bias in word embeddings. This paper examines bias in word embeddings, and how that bias carries forward into models that are trained using them. There are definitely some dangers to be aware of here, but also some cause for hope as we also see that bias can be detected, measured, and mitigated.
• Optimal prediction of the number of unseen species. This paper attempts to answer the excellent question: “just how well do we understand the diversity of species in our biome?”
• Haskell Documentation with Haddock: Wishes’n’Tips. Haskell has a reputation both internally and externally for having-poorly documented libraries (it’s the 2nd-most cited reason for folks’ unwillingness to embrace the language). And I kinda get it; I personally think Haskell code does read quite clearly once you’ve spent some time grokking functional programming. However, I also think that documentation and “here’s what this code literally does” are two different things, and Haskell could certainly improve in the “documentation”, i.e. “here’s why we chose to implement the code this way, and what we imagined the use case that method/library is solving” category. That’s where this blog post shines; Haskell has a excellent documentation toolchain, and I think that when it’s done well (Mercury’s codebase, for example, is one of the better-commented ones that I’ve seen), Haskell can have as good of documentation as any language – frankly, I think it can be one of the best, since the language itself does such a good job of explaining the “how” that the comments need only focus on the “why”.

Work

• How to Run a Ponzi Scheme for Tech People. This post is cuttingly sarcastic and calls out some folks that I know personally in the tech writing community, but I think it’s more of an indictment of the people who buy these courses than the folks who make them. Really, I think it’s mostly trying to say is that the author is frustrated by all the folks in tech who are taking advantage of less-motivated but well-wheeled technologists who are desperate for any edge on finding fulfillment and satisfaction that doesn’t involve working hard or taking risks. Can’t blame him.

Misc

• Transparent Octopus Caught by Blackwater Photographer Interview. This is an absolutely mind-boggling bit of photography, and I love how the photographer makes a point to talk about how the majority of these wild, transparent creatures that live deep down are mostly larval forms of adult animals that may look much more “normal” to our eyes. Larval forms of terrestrial animals are super wild, and it’s cool to see similar bizarre appearances from aquatic larvae.

My Favorite Link

Trying a new segment on my newsletter this week called “My Favorite Link”. Should be pretty self-explanatory as to what it represents :)

• Anyway, this week, my favorite link was Write five, then synthesize: good engineering strategy is boring. by Will Larson (@lethain). This piece is full of engineering strategy that is both actionably insightful (“Prefer minimal design document templates that allow authors to select the most useful sections and only insist on exhaustive details for the riskiest projects.”) and introspective (“Most folks are better writers than they are editors.”, I initially disagree with this take and then I slept on it and realized that he’s exactly right).

Culture

• Welcome to the Cloud Zoo, Photo quiz: This is an experiment about how we view history, Why are K-pop groups so big? Big batch of projects from one of my favorite online newsletters: The Pudding. My mini-descriptions will not do any of the links justice – check them out!
• Lorde: The Blackbird Spyplane Interview - Blackbird Spyplane. Jonah Weiner writes this hilarious newsletter on fashion, culture, and general media-elite navel-gazing, and I loved this week’s interview with Lorde. I’m a big Lorde stan and it’s not surprising that this newsletter’s interview with her was so cool.
• Inside the New York Public Library’s Last, Secret Apartments - Atlas Obscura. My favorite paragraph: “When these libraries were built, about a century ago, they needed people to take care of them. Andrew Carnegie had given New York $5.2 million, worth well over $100 million today, to create a city-wide system of library branches, and these buildings, the Carnegie libraries, were heated by coal. Each had a custodian, who was tasked with keeping those fires burning and who lived in the library, often with his family. “The family mantra was: Don’t let that furnace go out,” one woman who grew up in a library told the New York Times.” New York has so much cool history and this piece does a great job of conjuring up the historical context around a cool New York attribute.

Work

• This Japanese Shop Is 1,020 Years Old. It Knows a Bit About Surviving Crises. - The New York Times. Slow and stable works for a business if everyone is on the same page.
• Career Chats with Swyx and Randall (Staff Eng Podcast by Lethain). What happens after you go past Senior? Will Larson, CTO of Calm, has been interviewing Staff-plus engineers across the industry for his new book, Staff Engineering. This is his first full-length interview podcast episode!

Tech

• 2020 Haskell survey analysis. Some findings from the 2020 State of Haskell survey that extracts some additional insight from the data by clustering the respondents into various categories based on their Haskell usage and experience and analyzing the responses of those clusters.

• thought leaders and chicken sexers. Well-researched and thorough essay taking aim at some of Paul Graham’s rhetoric and his history not following through on language design concepts. Spawned some [Twitter])(https://twitter.com/ztellman/status/1336046099207864320) and HackerNews drama, as expected.

• In defense of blub studies. My favorite line: “In short, if you’re in search of generalizable knowledge that compounds exponentially over time, then blub studies looks like the crap you have to wade through to get to the good stuff. So it’s easy to see why people give up on understanding all the blub they’re surrounded by, except what they need to get the job done. But for me, the opposite attitude has been more productive. Computers can be understood—even if it’s hard and takes a while. Blub studies is more generalizable than it seems, and has its own way of compounding over time, too. That makes it a lot more useful than you’d expect.”

• Language Servers are the New Frameworks. Cool read on the value of language servers in improving overall developer experience. My favorite line: “Linting, typing, always-running-compilation and other forms of write-time checks and optimizations are not new. In fact, when dev teams write unit tests and lint rules, they are essentially creating their own bespoke language server for their own app. What’s new is that giving realtime feedback is now a responsibility shared by frameworks. This standardization increases productivity both when starting a new project and moving between projects.”

Business

• Playing on Hard Mode. World-class piece from Ben Thompson on why it’s no longer easy to make the new Facebook or Google.
• Big Tech Sees Like a State. Fascinating read on the overlap between state-run economic engines and the exploitative policies of certain big tech companies that mirror them.

Work

• Being Glue. Long read on how being good at your job can often lead to not being recognized as such because many workplaces measure and reward the wrong things.
• Shai Mendel’s Dev Blog. Interesting perspective on reframing the role of a senior engineer to focus more on mentorship and guidance and less on actually churning out code and owning features. In my experience, this is the best way that senior engineers can maximize their value.
• Six months of Tiny Projects. The author transparently lays out his work, projects, and ROI for a bunch of side hustles that he’d done over the last 6 months. Engaging read.

Tech

• Building the most inaccessible site possible with a perfect Lighthouse score. I used to use lighthouse for my site, and while it does some things well, I’ve since upgraded to a new tool – the reasons in this article aren’t the reason why, but they’re funny.
• PostgresqlCO.NF: PostgreSQL configuration for humans and psql command line tutorial and cheat sheet. More useful PostGRES tools. My workplace is a big PostGRES shop and I’m sharing the things I’ve been reading.
• The True Size Of … This is cute!
• Managing Database Migrations in Scala. DB migrations are a common operation when database changes are needed to accommodate, say, 3rd-Party API changes. This is a useful guide on how to do these DB migrations in a type-safe way.
• Elle: inferring isolation anomalies from experimental observations. Is there anything more terrifying, and at the same time more useful, to a database vendor than Kyle Kingsbury’s Jepsen? As the abstract of this paper wryly puts it, “experience shows that many databases do not provide the isolation guarantees they claim.” Jepsen captures execution histories, and then examines them for evidence of isolation anomalies.
• Parse, don’t type-check. Fantastic follow-up to this piece by another excellent Haskell engineer.

Business

• DoorDash and Societal Arbitrage. This is the best thing I read last week. Absolutely captivating content about (among other things) how well-moneyed startups are weaponizing government programs that were intended to help small businesses.
• This tweet by patio11 about execution vs innovation. Specifically, I liked the question of “how would we reliably and safely get every American $10 every week?”

Culture

• ‘It’s the screams of the damned!’ The eerie AI world of deepfake music. I think a lot about the effect of ML on music generation and this article does a decent job summarizing the current state of AI music generation. Worth checking out Google’s Magenta Project, too.

Work

• The Gods on HackerNews and Holy heck this is hard. Hackernews is mostly a great thing to read if you’re interested in cool content but the comment sections can be a a total cesspool and the first link is a tongue-in-cheek commentary on how any time someone does something cool and shares it with Hackernews then there’s always a bunch of folks who think it’s super edgy to just dunk on it. And the second link is a reflection on how hard making and scaling a product independently is really difficult.
• I sold Baremetrics. In the same spirit of the previous link, this one is a story of someone who did successfully accomplish the very difficult challenge of building, scaling, and selling his technology business.
• Simple Explanations without Sounding Condescending. Another excellent read from Julia Evans about how to explain things well. Some of her techniques include (1) write true explanations (as compared to ELI5-esque ones) (2) only use visuals if the visuals are easy to understand (3) tell a relevant story and (4) have a specific audience (explanations are not one-size fits all).

Tech

• Managing Multiple Java, SBT, and Scala Versions with SDKMAN. One of the trickiest challenges of getting started effectively with Scala can be managing the various different dependency versions. While I prefer coursier for dependency management myself, many of my coworkers like SDKMAN and this post provides a thorough overview of using it.
• Neil Mitchell’s Blog (Haskell etc): Turing Incomplete Languages. Long and technical read about how some languages ban recursion to ensure programs “terminate”. While that’s technically true, its usually irrelevant.
• How to get root on Ubuntu 20.04 by pretending nobody’s /home - GitHub Security Lab. I love reading about exploits and this one was a thorough writeup (complete with a video) and an engaging read.
• Postgres Observability. PostGRES is probably my favorite open-source technology and it’s amazing to see how many good resources there are to learn about it. This link visualizes every part of the PostGRES stack and explains what it does.

Work

• How I Start. Often tutorials and books are overly generic, leaving it up to the reader to wade through all the tools and styles available to a language on their own. These articles are meant for a user who is comfortable with a language in its REPL (if available) or building individual modules, but may not be comfortable taking the step to producing an application or library that is ready to be consumed by others or deployed to production.

Tech

• Haskell: The Bad Parts, part 2. Another excellent criticism of some of Haskell’s warts from the prolific Snoyberg. If you’re interested in critiquing programming language design choices, especially for a language as carefully designed as Haskell, you’ll likely enjoy this read. Rust fans probably will, too ;)
• Open-sourcing Haxl, a library for Haskell. Honestly, this was one of my favorite reads of the week because it’s an thorough and detailed writeup of how Facebook solved a technical problem in the perfect way by using Haskell. One my current gripes with Haskell blogs (I’m still pretty new to learning which ones to read, so bear with me) is that many of them focus on small, one-off problems and don’t make a case for why Haskell is the best tool for an enterprise use case. This post does a great job for that, and I’d recommend it to anyone who’s curious about why Haskell matters for enterprise software development.
• Distributed Systems Learning Diary. A great example of learning in public by a guy named Timil in London. I’m currently working through his notes (and the source material) from the Spark Course.
• Average UX Improvements Are Shrinking Over Time this is good actually; the smaller improvements are indicating that our views on what constitutes “good design” are starting to converge, which means that there’s less instability in how we approach UX design.
• How to Make the Kotlin Compiler Smarter. The Kotlin compiler is smart, but sometimes you understand your code better than the compiler. With the use of Kotlin Contracts (a new compiler feature), you can transfer your knowledge to the compiler by specifying the effect that the invocation of a function produces. The compiler tracks all the effects and uses the acquired information to smart cast variables or permit the initialization of variables in a lambda. This is a really intersting application of a functional effect system. I’m looking forward to seeing future applications of Kotlin Contracts!
• Troy Hunt: Hacking Grindr Accounts with Copy and Paste. Troy Hunt’s exploit write-ups are always interesting and this one was a particularly egregious bug.
• Virtual Consensus in Delos. Again, another great post from Facebook engineering when they were solving problems that no one else was worrying about. Back in 2017 the engineering team at Facebook had a problem. They needed a table store to power core control plane services, which meant strong guarantees on durability, consistency, and availability. They also needed it fast – the goal was to be in production within 6 to 9 months. While ultimately this new system should be able to take advantage of the latest advances in consensus for improved performance, that’s not realistic given a 6-9 month in-production target. This paper describes what Facebook did to solve this problem.
• Wasp. Another cool application of Haskell that, while not as battle-tested by Mt. Production as the Facebook applications are, still does a great job of showing an area where Haskell is an exemplary product fit. TL;DR, Haskell is a world-class language for writing DSLs.
• Creating a Haskell development environment with LSP on NixOS – Jussi Kuokkanen’s Computing Blog. Standard read on how to set up a Haskell dev environment. I’ve been reading a lot of these lately as I’ve been trying to find the best way to run Haskell on my machine and the approach outlined by this author works well for me so far.

But in addition to this being a big week for politics, it was a big week for Haskell, as the Haskell Foundation was launched this week at Haskell eXchange. This is really exciting news; it’s encouraging to see Haskell – which has historically been a very distributed language with no clear leadership – take clear steps towards centralizing governance, which can hopefully lead to more industrial engineering applications, and optimally more usage. Haskell is a great language and I’d be thrilled to see more folks adopting it.

Finally, I have some personal Haskell-related news that I’d love to share but it isn’t quite public yet; stay tuned for a newsletter coming soon! Okay, on to the links!

Business

• Supply-Side Shocks Don’t Cause Ample Increases in Nominal Income. This was written a while ago, but I only found out about the author this week (an economist named Jason Harrison), and I thought this was an interesting analysis of why recessions tended to lead to “Great Inflations” due to the policies put in place by the Fed, rather than recessions being a result of significant supply-side shocks. It’s a nice contrarian take to classic macroeconomics and worth considering.

Life

• What was fun?. I found this incredibly relatable; it talks about the foundational component of “fun” and how the quarantine has denied us one of the critical factors, which is spontaneity and the freedom to push ourselves outside of our comfort zone. With caution being espoused so liberally in these pandemic times, fun is clearly suffering.
• Kelsey Hightower. Inspiring profile on one of the most inspiring guys in tech. I’m not really a Go or K8s user but I have big respect for Kelsey’s work and work ethic.
• When Terrible Things Happen, our Numbers Go Up. As the world shut down this spring, The New York Times dropped the paywall to its What To Cook collection. It was a sign of goodwill to the millions of people stranded at home looking for a culinary project — or just some comfort food — but it was also a timely ploy to lure more subscribers into NYT Cooking, the paper’s standalone recipe storehouse and home kitchen guide.

Work

• When your Manager Isn’t Supporting you, Build a Voltron. Julia Evans tipped me off to this excellent read from Lara Hogan about how to effectively drive your career even in a situation where your manager isn’t providing the support that you need. I know a few of folks who have been in this situation and I think anyone who’s struggled with their relationship with their boss would get something useful from this piece.

Tech

• Names are not Type Safety. Another excellent technical deep dive on accidental misuse of the newtype keyword to attempt to provide type safety that only ends up providing an alias. It’s long but excellent.
• Edsger Dijkstra: The Man Who Carried Computer Science on His Shoulders. Grab a coffee for this one. Probably 2. Dijkstra is one of the pioneers of computer science as a discipline and this is a detailed profile on his life.
• How to Seriously Read a Scientific Paper. My graduate school friends likely know this stuff already, but as someone who is trying to read more papers for fun, I got a lot out of this piece, which profiles several academics providing tidbits of advice on how to read papers effectively. A big consistent theme: have an engaging task (note-taking, highlighting, writing) to do alongside the paper read to encourage active engagement with the material and prevent eye-glazing.
• Grand Unified Theory of Software Architecture. A pragmatic approach to designing software. Good content to consider when code reviewing or writing your own code.
• Ugliest App. As someone with a lot of backend development experience who hasn’t worked as much on building a pretty frontend to match, this app platform speaks to my soul.

Culture

• Inside Foxconn’s empty buildings, empty factories, and empty promises in Wisconsin. You should listen to this Reply All podcast before reading this piece :)
• Facebook is censoring me. Delete your Facebook account.
• Ghislaine Maxwell deposition redactions: How to crack them. Fascinating read on how to decode censorship patterns used by the secret service when redacting public information, and then applying those decoding techniques to decode the Ghisliane Maxwell deposition.
• Meet the Excel warriors saving the world from spreadsheet disaster. Just trust me.

Life

• Notes on My Colon Cancer. This piece was a fun read on decidedly not fun topic topic. I’d never heard of this author before, but his candor in writing about such an intense topic drew me in and makes me want to read more of his writing. Definitely worth a read if you need a refresher on things to be grateful for.
• Life & Death. Another intense read about intense subject matter, this piece is Scott Galloway at my favorite – he gets deep to the human element of a reader question and dispenses sage advice on how to deal with the passing of a parent. Having lost my Grandmother recently and gone through both my own grief and the secondhalf grief of my parents during this time, this piece rang especially true for me.

Tech

• Helios: Hyperscale Indexing for the Cloud & Edge. Deep dive into the details of Helios, a distributed, highly-scalable system used at Microsoft for flexible ingestion, indexing, and aggregation of large streams of real-time data that is designed to plug into relationals engines. As an ingestion and indexing system, Helios separates ingestion and indexing and introduces a novel bottoms-up index construction algorithm. It exposes tables and secondary indices for use by relational query engines through standard access path selection mechanisms during query optimisation. As a reference blueprint, Helios’ main feature is the ability to move computation to the edge.
• virtualizing a hackathon at ScalaMatsuri 2020. A report of running a virtual hackathon at ScalaMatsuri Day 2 Unconference.
• Timekeeping in Financial Exchanges. This piece challenges the reader to think about the difficulties of clock synchronization in distributed architectures. Makes for a great read when coupled with this podcast from Signals and Threads.

Work

• Advice to my young self: to succeed in your career, forget side projects and focus on your job. I really needed to read this post this past week. I’ve been spending a ton of time lately working on my side project and I’ve found myself burning out at both my day job and my side project. Despite the title of this post: this blog doesn;t encourage readers to quit their side hustles, but simply asks them to consider if the opportunity cost of working on a side project is worth the tradeoff of excelling in their existing day job. Many of us tend to write off the benefits that come from doing your day job well.
• Talking, Typing, Thinking: Software Is Not a Desk Job. As a software consultant, I don’t bill my clients for the time spent thinking in the shower, but this post convinces me that maybe I should ;)
• Applying Textbook Data Structures for Real Life Wins. I love blog posts like this one that take common industrial software engineering challenges and employ fundamental CS concepts to solve them.
• When is it time to quit my 9-5?. Interesting perpsective on the heuristics and hard metrics used by Pariss Athena (the founder of Black Tech Pipeline (BTP)) to calculate the best time for her to quit her day job and focus on her passion project.

Culture

• The Island That Humans Can’t Conquer. Humans can only live in a comparatively small range of climates on the planet, and I find it interesting to see how we’ve pushed those boundaries throughout human history. If you find that interesting, too, you’ll probably like this.
• Johns Hopkins Poker Course. I loved this link! If you’ve ever wanted to improve your poker playing skills (and, by extension, your betting abilities) through a grounding in first principles, Avi Rubin’s course provides a thorough overview of how you can do that (the only thing to really do after, of course, is practice).
• Why Life Can’t Be Simpler. The answer, of course, is because the “simplest” things are the things that require the least amount of assumptions, which are some of the hardest things in the world to actually find.
• This website is killing the planet. Just trust me.

Tech

• Orbital Edge Computing: Nanosatellite Constellations as a New Class of Computer System. This paper is a wild ride into the world of true “edge computing” as it outlines some of the interesting approaches used by semi-autonomous micro-satellites that work together to send information to space to earth with limited bandwidth.
• The Case for a Learned Sorting Algorithm. Sorting is one of the classic problems in computer science and this paper provides some interesting benchmarks for a new type of ML-driven sorting algorithm that significantly outperforms RadixSort on a massive (1 billion item) dataset, and that performance includes time to train the model. This is a pretty exciting result for anyone interested in sorting data at a large scale.
• Effect Runtime and Generic IOApp alternative. Alexandru Nedelcu is an excellent Scala engineer and prolific writer, and these two pieces are great foundations (with plenty of example code) for understanding how to build pure IOApps and effect runtimes using Cats and Cats-Effect. If you’re interested in pure FP for Scala, understanding this is the equivalent to a 101 course, and Alex makes it really easy to grok.
• Signals and Threads: Python, OCaml, and Machine Learning and Signals and Threads: Language Design. Signals and Threads is my favorite podcast in tech right now: Ron Minsky is an engaging and enthusiastic host and he has a plethora of incredible engineering talent at Jane Street to pull information from. This weeks podcasts covered my two areas of programming interest: language design, and machine learning, and I especially enjoyed both of them.

Work

• GitHub - danluu/post-mortems: A collection of postmortems. Dan Luu is a great writer and teacher of industrial software engineering and this repo is a cool collection of postmortems published by a variety of different industries. If you work in operations at all, postmortems are an unfortunate but vital part of the job, and this cache of postmortems makes for great inspiration when it comes to designing and implementing a postmortem process.
• SEO mistakes I’ve made and how I fixed them. I’m pretty new to the blogging game but even I recognize the importance that good SEO can have on getting my posts in front of a wider audience. This article provides a great summary on low-hanging SEO fruit for your blog or website that can be tackled without making any major changes to your content production flow.