Watching the tech industry unleash unpredictable software into production in the name of AI adoption, then spend twice as much time and money trying to limit the blast radius, proves humans should be investing in our own intelligence before allowing the machines to mimic us.
Exciting day!!! The Early Release of our new O’Reilly book is out! First two chapters. More to come soon! Grab it while it’s hot!
learning.oreilly.com/library/view...
#Java #OReillyMedia #Agents
Anecdotal evidence that despite all the great progress and the undisputed usefulness this is what we're all (still) dealing with:
🤯
> ... **to avoid the work of fixing them** ... I'll be more careful about **accepting test failures as "by design"** ...
Let that sink in before the weekend. 🙃
In the first ages of AI, writing a program to play checkers or chess or Jeopardy or Go was seen as an analog to human intelligence.
In this current age, language is the proxy for human intelligence.
Now I have a clear picture. 🤡
The International AI Safety Report 2026 concludes that fully autonomous attacks aren’t here yet…the experiences of teams deploying real-world autonomous offense tell a different story.
Explore the gap between assumption and operational reality in our latest blog: https://bit.ly/4sjKrnN
A new #Jepsen report, on MariaDB Galera Cluster 12.1.2. Two cases of write loss, along with P4 (Lost Update) and Stale Read in healthy clusters.
jepsen.io/analyses/mariadb-galera-...
Apache Iggy's migration journey to thread-per-core architecture powered by io_uring
iggy.apache.org/blogs/2026/0...
Reactive Loom with Francesco Nigro.
Join us next Monday, March 16th, for Quarkus Insights Ep. 239 as Francesco Nigro joins us to show how a custom Loom scheduler built on top of Netty brings together the performance of event-driven I/O & virtual-thread-friendly blocking APIs.
bit.ly/quarkusinsig...
#quarkusinsights
Goldsky was hitting the limits of Kafka at scale with tens of thousands of partitions and petabytes of blockchain data. By moving to WarpStream, they cut costs by 90%, removed bottlenecks, and now scale seamlessly toward 100 PiBs and beyond 👉 buff.ly/pbkiUMJ
"Further, the sources claim that Palantir can see far more information than the government realises. Palantir can aggregate data from across different government datasets to generate top secret information, the Nerve has been told."
Love, that this is the quote @martinfowler.com picked.
(btw, pi has an rpc.mode, should be trivial to integrate with the one true editor)
Awesome. And well deserved highlight!
📝 BLOGGED ✏️
Here is the #FlinkCDC #JSON #UDF #deepdive I hinted at lately.
Hopefully useful for anyone trying to extend the processing capabilities of Flink CDC by means of UDFs and ideally saves you a few precious hours if you have similar requirements.
🔗 dev.to/hpgrahsl/fli...
It's rare that I get to reflect back on my entire tech career and my philosophy towards work and life, but this interview captures it perfectly. One of the few recordings of myself I've watched end to end. www.youtube.com/watch?v=HdUb...
A new #Jepsen report: we demonstrate data loss and persistent split-brain in the NATS streaming system, in response to simulated power failures/OS crashes.
https://jepsen.io/analyses/nats-2.12.1
New blog post! ✨ I argue that AI will make formal verification go mainstream. martin.kleppmann.com/2025/12/08/a...
Three reasons for this:
1. LLMs are getting increasingly good at writing proofs using proof assistants. This will make formal verification vastly cheaper than it's been to date.
“AI models not only point some users to false sources but also cause problems for researchers and librarians, who end up wasting their time looking for requested nonexistent records…”
Wait what's this?!
I've managed to pull my finger out to ship a list of **a whole lot of interesting links about stuff going on in the world of data for November**, a whole three days left until the end of the month!
rmoff.net/2025/11/26/i...
Docling #Java is the official Java client and tooling for #Docling — a suite that simplifies document processing and parsing across diverse formats (with advanced PDF understanding) and integrates seamlessly with #GenAI frameworks.
Good news for #MongoDB / #ApacheKafka users who need to feed #Iceberg tables... 🧊 $iceberg now in private preview in #Atlas #StreamProcessing 👇
medium.com/towards-data...
Runden die normalerweise nicht immer auf ganze 10 Tausender? Irgendwas stimmt da nicht sag ich dir 😜
Kind of the #final #boss for your favorite #LLMs... Try anything related to #Wingdings and just have fun in the conversation. 😂
At Netflix, @alex.bertails.org and his team have adopted #RDF to build a Unified Data Architecture (UDA) and a #domainModeling language, Upper, that let them quickly and efficiently share complex #data with their internal engineering customers.
knowledgegraphinsights.com/alex-bertails/
Ja wäre wohl vergeudet für dich. Ich hab es auszugsweise angeschaut und ohne auch nur ansatzweise so detailliert bescheid zu wissen wie du, hat es dennoch sehr weh getan phasenweise :)
Und ich hab mich ernsthaft gefragt, ob du dir das angesehen hast, wie es dir ergangen ist und was dir geholfen hat, es bis zum Ende durchzustehen???
This is why we can’t have nice things
Bist du eigentlich jemals "in Serie" gegangen damit? Ich frage für einen Freund, ob man die Dinger wo erwerben könnte?
