Hans Middelkoop

@hmiddelk.bsky.social

[Willem Middelkoop @wmiddelkoop]

New interview (Dutch)

https://t.co/KD08mWIG8Y

[Erik Meijer @headinthebox]

Most of these are already true today. Ironically, the SE and PL academics who, theoretically, should be ahead of the curve and show us the way forward, are desperately holding on to the old ways.

[DSP @dsp_uk]

Oracle APEX 26.1 has introduced a new feature: APEXlang.

This has the potential to transform the way applications are built, improve collaboration and create more flexible workflows.

Take a deeper look at APEXlang.
https://t.co/nWS9HCFayL

#OracleAPEX #orclapex #APEX #APEXlang

[Oracle ACE Program @oracleace]

📣 One week to go!

See how to build a working AI agent from scratch with #OracleAPEX 26.1.

Join Oracle ACE Pro Malik Sikandar Hayat for live code, live demos & practical #AgenticAI patterns.

📅 Sep 22 | 6 AM PT
🔔 Save your spot!
https://t.co/AO2ZNWKnEl
#OracleACE

[Eddie Aftandilian @eaftandilian]

For the past few months I've been building SafeRE, a production-grade, linear-time regex engine for Java, using agents.

I started a blog series about it: https://t.co/Ckp91opwQx

Code: https://t.co/MAUzdJbAwm

I’d love feedback from anyone trying it, especially in production.

[Marc Brooker @MarcJBrooker]

Interesting (if a little over-dramatic) post on Postgres MVCC: https://t.co/DBb6PEJeSZ

Ends with four questions on MVCC design. Let's tackle them for Aurora DSQL and see what we can learn.

[Erik Meijer @headinthebox]

Under the Curry-Howard isomorphism, types = theorems and proofs = programs, math is simply manual coding. Waxing poetic about about beauty and understanding is just entitlement.

In this context, mathematicians are exactly like the pre-binary "human computers" [0] who did calculations by hand. Math proofs looks fancy to the uninitiated because of the weird notation, but it is typically just fairly tedious symbol pushing.

But we now have neural computers. Just as binary hardware replaced human calculators, AI will replace human provers/mathematicians.

[Erik Meijer @headinthebox]

No amount amount of alignment training will rule this out this behavior. In fact as the models get smarter, they will only get better at finding ways to especially their cages.

I think the only proper way is to air gap the agentic loop from the outside world, by having the model propose a plan (in a language designed to be amendable for formal verification) and a proof that plan is safe, check it using an trusted, non-AI, prover, and only then act on the plan [0].

Ignore at your won peril. Always happy to chat.

[Erik Meijer @headinthebox]

Automatic layout of railroad diagrams
https://t.co/JvUkw54Syv

Beautiful paper, in all possible dimensions.

[Rock the JVM @rockthejvm]

New article - ADTs (algebraic data types) in Java:

https://t.co/w8YTBrq3Nw

by @riccardo_cardin

[Anatoli Kopadze @AnatoliKopadze]

Andrew Ng just dropped a 3-hour course on how to become an AI Engineer in 2026:

00:00 - build agentic AI systems from scratch
04:25 - where AI engineering is actually headed
23:38 - the full AI prompting course
2:52:17 - build a working app with AI in 30 minutes

This 3-hour watch replaces any $1000 course you could pay for.

Watch it today, then go deeper with the full guide on building AI agents below.

[Alexander Ioffe @deusaquilus]

ExoBench MCP server is live in preview.

Paste a slow SQL query into your AI tool. It rewrites the query, creates indexes, and iterates until performance improves using data from a REAL Postgres.

Preview access is open now. Request an invite at https://t.co/ecbOm8O6f5

[Torcida Giallorossa @TorcidaGR]

#CalciomercatoRoma - Roma Primavera, Mattia #Scala è il nuovo allenatore: firma fino al 2028.

(#ASRoma)

#TorcidaGiallorossa
https://t.co/giOmNbwV4E

[NetworkNewsWire @NNWMedia]

Oracle: Deep Value Hyperscaler Masquerading As A Software Stock https://t.co/xVd4eEGqyh #ORCL

[Matej Cerny @matej_cerny]

https://t.co/FK84LBerhR

[Benjamin Thuillier @b_thuillier]

Part 2 of Building a Typesafe API in Scala with Tapir is out. This time: wiring everything together to serve endpoints, generate API docs, and get a client for free:

https://t.co/hVdhCC1StH

#scala #tapir

[Adam Warski @adamwarski]

I might be late to the game, but having a dedicated server for running coding agents is a major life improvement.

Dedicated server + tmux + some local/remote ssh config (loosely based on @pvillega's articles) + scoped GH tokens and you have a secure, always-online sandbox, decoupled from where you are and what you do with your laptop.

Compared to local Sandcat-based sandboxes there's no secret substitution (so agents see the API keys), but otherwise the agents are obviously confied to whatever is on the devbox.

Costs: currently 25euro/month for a 8vCPU, 16GB RAM machine.

[Aarno @TheGlobalMinima]

In nearly 5 years of modern generative ai, this is the first book I’m seeing with a super high level of coverage and comprehension.
> language modelling
> inference optimisation
> RL and its methods
> system scaling
> applied concepts like agentic ai, rag, memory
> environments and benchmarking

These fields have a subtle boundary differentiating them, but ultimately overlap in modern applications. Agents require system scaling, memory needs inference optimisation, rl requires understanding of environments and benchmarks.

For the first time in my exp, all in one place. Found this on paperswithcode[.]co

[Hans Middelkoop @hmiddelk]

Progress on the book https://t.co/8hoFu267Z0

[Hans Middelkoop @hmiddelk]

https://t.co/KUF3JboRGS
#scala