Second, you heard about MCP, but how does it work? Where are the tools invoked? Dive in: virtuslab.com/blog/ai/how-...
Third: deep dives the latest trending Github AI-related repositories, delivered every week on virtuslab.com/blog/ai
If you're not yet fully immersed in agent-assisted programming, we've got some resources to get you up to speed!
First off, agent visualization: interact with a real LLM and see how and when are tools called virtuslab.com/agent-visual...
We also have special student tickets for a fraction of the usual price.
More: www.rustikon.dev
Are we going to rewrite everything in #Rust? Is Rust the best language to generate code using AI?
These would be my bets for the hallway track of @rustikon.dev - 19-20 March in #Warsaw, Poland.
If you don't have your ticket yet - don't wait any longer :)
Only a month left until @scalarconf! The agenda's there, the afterparty is planned - now you just need to get a ticket :)
Affordable as always, in the center of #Warsaw, easily reachable by plane, train, bus, car (by bike and on foot as well ;) ).
www.scalar-conf.com #Scala
#Scala's capture checking is useful not only when dealing with resources, but also when paired with other values that shouldn't leak syntactical scopes, such as Ox's concurrency scopes.
Here's a short breakdown of how this works:
youtube.com/shorts/ozoeD...
Latest Tapir (rapid development of self-documenting APIs) brings performance improvements when validating decoded request (e.g. JSON) bodies. Try it out!
github.com/softwaremill...
LLMs could guard that the human-maintained sources stay in sync with the now "low-level", compiled, C++/Rust/Java/whatever implementation.
With LLMs, can we now program using the abstraction level as presented in the book?
This new "source code" would be lower-level than a declarative specification ("I need to obtain a MST"), but higher-level than what we use currently, with liberal use of natural language.
On the left: Kruskal's minimum spanning tree algorithm as presented in CLR's "Introduction to Algorithms".
On the right: an implementation of the same algorithm, here in C++. Quite terse, but still much richer in technical detail, and the core idea is not that visible.
jox channels 1.1.2 (#Java virtual-thread based safe concurrency & streaming) introduce non-blocking trySend() and tryReceive() methods.
Thanks to Andrzej KobyliΕski for working on this!
github.com/softwaremill...
We're running the 2026 #Scala Survey!
Let us know how you are using Scala, #AI with Scala, and give hints to the Scala teams (both at VirtusLab and EPFL) on where the focus in driving Scala forward could be.
Thank you - your answers count :)
virtuslab.typeform.com/ScalaSurvey2...
After a couple of trial-and-errors, Sandcat finally works well for #Java & #Scala development, alongside IDEs such as Metals!
The first sandboxed autonomously-created PRs are already merged into our OSS projects
github.com/VirtusLab/sa...
FinBot: exposing data to a wider audience using natural language queries. Project profitability, employee metrics, quarterly comparisons - some examples of queries that our internal users asked.
How does it work? Read the full overview by @kgrajek.bsky.social : softwaremill.com/reducing-tim...
How does it work? See the full (written) guide: softwaremill.com/understandin...
A 01:46 introduction to what kind of problems capture checking in #Scala solves. Pushing forward what type systems might do!
youtube.com/shorts/TJofT...
Book cover of "Designing Data-Intensive Applications, 2nd edition". It has a similar wild boar on the cover as the first edition, but it uses O'Reilly's new cover design, and the boar is now slightly colourised.
The second edition of Designing Data-Intensive Applications, by myself and @chris.blue, is finished and sent off to the printers! Ebooks should be available in the next week, and print books in 3β4 weeks. Sigh of relief. π
(BTW, this is a good opportunity to support your favourite local bookshop!)
sttp-ai now supports #Scala Native!
You can use: OpenAI, Claude, Grok, Ollama, ...
With different flavours of Scala: direct, functional, Future-based
On various platforms: JVM & Native
(not all combinations are supported, though)
Enjoy :)
Prompt-to-install: yes or no?
(of course, especially since this is for a project which sets up sandboxing, such prompts should always be run in ask-permission-for-everything mode)
Based on your feedback (thanks!) we've updated our Agent Visualization!
What's new:
* better mobile view
* clearer labels - what's generated by the LLM vs. tool output
* improved prompt, so that it's easier to follow the reasoning
What do you think?
virtuslab.com/agent-visual...
adopt-tapir.softwaremill.com now uses the latest&greatest #Scala 3.8.1
Generate a one-file scala-cli project exposing a self-documenting HTTP API, or a full-fledged multi-module-ready sbt project with tests. All type-safe & LLM-friendly.
* Claude Code-specific mounts so that your `~/CLAUDE.md` is there, along with agents & commands
* "public" envs next to "secret" envs which are substituted, plus git config on startup
An update of Sandcat: a dev container setup for securely running AI agents in `--dangerous` mode
* no forwarding of SSH/GPG agents into the container
* restricted VS Code settings to lockdown the environment
We'll of course cover how to work with AI and on AI through Rust, but that's just the start - macros, async, web, concurrency,...: a great overview of how companies and people use Rust.
Tickets are just 350 euro, for 2-days w/ an after-party. Join us in #Warsaw, a great city to visit and live in :)
Are you into #Rust? Or hoping to get into the ecosystem?
Then @rustikon.dev is for you :) A friendly community, meeting for the 2nd edition of the conference. And we've got a great agenda!
Sandcat routes all container traffic through a transparent mitmproxy in WireGuard mode (deployed as side-containers). Heavily inspired by matchlock & agent-sandbox.
github.com/softwaremill...
Continuing the quest to lock down #AI agents in a sandbox: here's Sandcat, a dev container setup that intercepts all container network traffic, filters it & injects secrets.
Your API key values never even exist in the container. But you retain the convenience of reviewing & browsing code in an IDE
Learn Kafka essentials from a Confluent Elite Partner πͺ
Thereβs plenty of theory out there, but nothing beats production experience. Weβve seen exactly what works and what breaks in high-scale data streaming environments.
Download the ebook: softwaremill.com/ebooks/start...
A two-day conference ticket is 350 euro. As a bonus, find out why everybody's saying #Warsaw and #Poland are such great places to lives :)
Tickets here: www.scalar-conf.com
