If copyrighted works can be whitewashed / reimplemented via AI, then copyright is dead.
#AI #OpenSource
alexn.org/blog/2026/0...
If copyrighted works can be whitewashed / reimplemented via AI, then copyright is dead.
#AI #OpenSource
alexn.org/blog/2026/0...
"Boring Tech May Become Irrelevant Tech"
The best practices that your organization uses for picking the tech stack are probably obsolete.
#AI #Programming #Languages
alexn.org/blog/2026/0...
βTapir Server with Cats-Effect and Pekko HTTP (snippet)β
#CatsEffect #Akka #Pekko #Tapir #Scala #Snippet
alexn.org/blog/2026/0...
Released a new version of Monix Newtypes β cross-compiled #Scala library for defining newtypes with utilities.
It upgrades all dependencies to latest versions.
github.com/monix/newt...
Micro-library exposing an API for "delimited-continuations" for #Scala JVM, ScalaJS and Scala Native. Extracted from "lampepfl/gears", may be useful for "direct style" libraries that want to be cross-platform.
github.com/funfix/con...
Great Lent starts monday, so might as well do a digital detox. I'm getting off social media for 40 days π
Exceptions:
β£ I'm going to post about my attendance to #ScalarConf, a #Scala conference, March 26-27, see you there!
β£ I'll keep writing on my blog and links will automatically get posted here.
Functional programming (#FP) is programming with functions, where a function is a unique association from one set (the domain) to another set (the codomain) (β x1, x2, x1 = x2 => f(x1) = f(x2)).
Functions come from maths, and FP is not procedural programming.
Hope that makes it clear.
Thomas (on the left), my tomcat, tries kissing Zuzi, my black cat (on the right).
Zuzi, my black cat, looks startled and annoyed after Thomas approached her
Zuzi attacks Thomas with her front paw
Zuzi leaves Thomas behind, with Thomas looking as if he has a broken heart
I know you come here for professional, grown-up stuff, but you're taking all of this way too seriously.
So, here's a love story in 4 acts #cats #pic
Hahaha
Some of the best arguments I've seen against #LLM / #AI use:
- undermines experience
- redirects seniors away from teaching
- have limited & biased knowledge
- code has always been incidental for problem-solving
- not conducive to truth seeking
blog.fogus.me/meta/LL...
Meme: A cartoon character with short black hair and glasses is pointing forward with a surprised expression. The scene includes a window with a view outside and a butterfly flying near it. The text above reads βimplicit ec: ExecutionContextβ and below asks, βis this direct style?β This is related to the Scala programming language, of course.
Articles talking about βdirect styleβ in #Scala should definitely answer this question for `scala.concurrent.Future`
Cultural hegemony matters due to what people are actually exposed to. Hollywood movies, US-made games, US-made content on US platforms (YouTube, Netflix, Disney, Prime), US app stores.
People outside US may not care care about US's culture or politics, but we get the downstream effects anyway.
Also, censoring erotica but not the kind of content that has led people to suicide is stupid, showing the puritanical culture of the US (violence OK, nudity, not OK); and unfortunately, it also shows the reach of US's cultural hegemony (saying it despite me believing much of it is good).
Overall I got it done, but I relied on my experience to review everything that it did, and this was output from frontier models, Claude Sonnet 4.6 and GPT-5.3 Codex.
A lesson learned may be to estimate the complexity of the task and ask for less in a PR.
Yeah, and it wasn't the end of it. I continued work on my laptop with Codex, as I discovered other issues.
This is common IME: you go through multiple cycles of reviews (including automated ones) + fixes, and the PR gets huge. It can't be granular, since it expands from fixing what it got wrong.
GitHub Copilot Agent is very cost-effective, for now.
1 prompt = 1 session = 1 premium request = $0.04 (if not included in subscription), and I've seen sessions last for ~1 hour. For complex work you need more prompts, but you can have ~10 hours of work for $0.4 with Claude Sonnet 4.6.
#AI #LLM
I tested Claude Sonnet 4.6 via GitHub Copilot Agent.
π The good: resulting PR is much better than the previous PR, same topic (which was unusable)
ππ» The bad: I did a review with over 40 comments for changes (more than changed files), incl. bugs & security issues
#LLM #AI
github.com/alexandru/...
βTalk is cheap. Show me the code.β (a phrase once used by Linus Torvalds) can also be used to address online influencers peddling AI/LLM agents.
Folks, for all that orchestration and tokens spent, where are the projects and accomplishments you have to brag about?
Hybrid IO-driven Promise (Scala snippet)
#Scala #CatsEffect #Akka #Concurrency #FP #Programming
alexn.org/blog/2026/0...
π released sbt 1.12.3, featuring the end of Scala 2.13-3.x bilateral interoperability (aka Scala 2.13-3.x sandwich)
while it's popular to discuss the topography of hot dogs and tacos, an oft overlooked sandwich is Wawa hoagie, which is cut only on one-side
eed3si9n.com/sbt-1.12.3 #Scala
I reported a #Scala 3 bug, and it was fixed in one day π
And this isn't the first time I'm amazed of how fast some reported issues get solved in this repository.
github.com/scala/scal...
Project exposing a Delayed Queue, for #Java, #Kotlin and #Scala β¦ a database-powered message queue that can also delay messages, with support for your favourite RDBMS.
github.com/funfix/dat...
I wanted to make a dumb joke on soon seeing social networks designed exclusively for AI Agents, but it was already built.
Screenshot of a test showing `Currency.getInstance(new Locale.Builder().setRegion("BG")...)` returning `EUR` on the latest Java runtime.
A colleague investigated a unit-test that was mysteriously failing on some dev laptops, whereas it was passing in CI. It turns out it's because #Bulgaria's currency just changed to EUR β€οΈ so it depends on how up to date the #Java runtime is.
Happy dev/prod/CI divergence day π
"Beyond agentic coding"
I like VS Code's more recent inline hints, and coupled with the inevitable chat fatigue, I'm thinking the author is right.
haskellforall.com/202...
βAmericans will measure with anything but the metric systemβ π
I was looking at a piece of #Scala code today, written ~3 years ago for $work, and I'm in awe of its maturity and elegance.
Granted, it was my own code π
Hey everyone, apparently it takes $20,000 and zero developers to copy/paste code from GCC & LLVM. That code was already available, ofc, so that money is for whitewashing the OSS licensing and copyright away. Exciting times!
Small update for my Funfix Tasks project β I've added integration with Kotlin's Coroutines.
github.com/funfix/tasks...