CloudNative and DevOps enthusiast 🚀 👨🏻‍💻 Engineer @elastic #o11y - Opinions expressed here are my own and do not represent my employer

Joined February 2010
172 Photos and videos
Pinned Tweet
The most important things to consider when building #software: make it reversible, easy to adapt to future needs or change completely. Velocity is the first metric of code quality.
1
3
10
Francesco Gualazzi retweeted
Second-to-last stop: TigerBeetle in Milano! Join us for a talk by @errebino, followed by Zig-focused lightning talks and demos, and an aperitivo near the venue. Grazie mille to our hosts, @inge4pres and @croloris! ​Dec 6, 5pm luma.com/fbt5g0kx
And Rio! @rbatiati will talk about TigerBeetle’s development process and philosophy, @joymachs will walk through his creative process, and Kalebi Serafim, CTO of OKTO Payments, will share insights on how to balance technology, regulations, and business. luma.com/ejwcicei
4
17
8,517
Published 2 new libraries to aid #Zig ⚡️adoption in #CloudNative ecosystem: 1. yaml.zig: github.com/inge4pres/yaml.zi… YAML 1.2.2 parser, needed in k8s/OTel land to parse all sorts of configurations 2. k8s.zig: github.com/inge4pres/k8s.zig Very barebone #kubernetes client
1
2
76
Francesco Gualazzi retweeted
Excited to announce the TigerBeetle 1000x World Tour! 13 Cities in 6 Days, December 1-6 From AMS, BLR and CPT to VIE—join us! tigerbeetle.com/event/1000x
11
12
61
41,620
New blog post is out: "#Zig is great for Observability" 4pr.es/qH_EoHNh My thesis: it is time for O11y ecosystem to consider Zig as relevant due to its characteristics. Shameless plug at the end of the post: we're building an #OpenTelemetry SDK for Zig

49
Never in the history of programming it’s been so important to know when to use Shift Enter. #software #engineering #ai
49
Francesco Gualazzi retweeted
30 Sep 2025
Actually managed to find time to write a blog this week! Lots of shiny new features rolled out this month! CPU GPU metrics, abs CPU %, Python 3.13 support, Azul Zing unwinding, and pretty massive performance tweaks in the profiler -- among other things! Link in thread.
1
1
2
174
Francesco Gualazzi retweeted
25 Sep 2025
Paving the way for unlimited context windows. Introducing @zml_ai /attnd: sparse logarithmic attention, on CPU, over UDP, faster than GPU.
33
100
763
121,890
I just published the first version of Zig #opentelemetry semantic conventions github.com/zig-o11y/opentele… this will be paired with the OTel SDK we're developing in parallel to provide Zig users the full spectrum of OTel attributes to populate data points, logs/events and traces.
2
3
2,306
To the Zig community here: we are building an #OpenTelemetry SDK for Zig, and we're stuck on a couple of hurdles... Join forces with us to make Zig a first-class citizen of OpenTelemetry and Observability! github.com/zig-o11y/opentele…

4
10
5,533
Interesting times ahead
The debate around AI browser security is heating up. As @simonw points out, LLMs browsers create structural security risks We explore why building a lightweight browser can better contain those risks ⤵️ lightpanda.io/blog/posts/bro…
1
63
Francesco Gualazzi retweeted
22 Aug 2025
Brief info and POC for this week's Apple 0click iOS 18.6.1 RCE bug CVE-2025-43300 github.com/b1n4r1b01/n-days/…
15
217
806
131,356
Francesco Gualazzi retweeted
We were able to reproduce the strong findings of the HRM paper on ARC-AGI-1. Further, we ran a series of ablation experiments to get to the bottom of what's behind it. Key findings: 1. The HRM model architecture itself (the centerpiece of the paper) is not an important factor. 2. The outer refinement loop (barely mentioned in the paper) is the main driver of performance. 3. Cross-task transfer learning is not very helpful. What matters is training on the tasks you will test on. 4. You can use much fewer data augmentations, especially at inference time. Finding 2 & 3 mean that this approach is a case of *zero-pretraining test-time training*, similar to the recently published "ARC-AGI without pretraining" paper by Liao et al.
45
294
2,584
368,255
I am not an #AI expert, but I stumbled on this paper and it seems like a nice breakthrough. arxiv.org/pdf/2506.21734 Essentially, throwing away the current state of the art based on transformers and modeling a different type of interaction between the layers of the NN.

1
73
#Opentelemetry protobuf definitions for Zig are available as a standalone package in github.com/zig-o11y/opentele… 🎉 I will try to backport prior releases, for now the latest v1.7.0 is available for use in all Zig projects
2
1
9
3,187
Francesco Gualazzi retweeted
TigerStyle existed before TigerBeetle was created, and was intended for software development in general, as a methodology for “higher quality software in less time”. Enjoyed this post from @LewisCTech, which gets it. lewiscampbell.tech/blog/2507…
4
20
141
13,990
Francesco Gualazzi retweeted
The core idea: @ziglang's new async I/O is an injectable interface. You inject the I/O implementation and your choice propagates to all dependencies. Simple. Composable. x.com/DominikTornow/status/1…
Asynchronicity is not concurrency @ziglang New Async I/O x.com/croloris/status/194416…
2
3
56
10,602
Francesco Gualazzi retweeted
26 Jun 2025
Elastic Cloud Serverless on @Azure is now GA! Powered by the same Elasticsearch you love, reimagined for the cloud - it's the fastest way to start and scale security, observability, and search hassle-free. Learn more in our latest blog: go.es.io/4lpLKOc
9
21
1,875