iiinova — where your digital self keeps evolving

Pedro Madruga @pedromadruga.com · 23/09/2026

There was a moment in time but it is premature to say that EU models are done. The EU is cooking.

Pedro Madruga @pedromadruga.com · 19/09/2026

I believe the same principle can be applied to most of the abstractions out there, especially the ones branded as "build once run everywhere".

001

Reposted by Pedro Madruga

Peter Ullrich @peterullrich.com · 08/09/2026

🚨 Announcing Pushin 🚨 Now officially: I'm announcing Pushin.eu, a git hosting platform built and run entirely in Europe. Think GitHub/GitLab but European, better, and faster. Built with #ElixirLang and #RustLang. peterullrich.com/announcing-p...

peterullrich.com

Announcing Pushin

I'm launching Pushin.eu, a git hosting platform built and run entirely in Europe.

1114449

Pedro Madruga @pedromadruga.com · 08/09/2026

There has never been a better time to learn software engineering.

000

Reposted by Pedro Madruga

Micah Lee @micahflee.com · 02/09/2026

Very much worth a read

03614

Reposted by Pedro Madruga

Doug Turnbull @softwaredoug.bsky.social · 29/08/2026

As a search person, I get a front row seat to AI team problems. Honestly they look like search team mistakes from 10 years ago * Ignoring evals * Thinking retrieval is 'one thing' (in this case RAG) * Thinking in chunks, not metadata softwaredoug.com/blog/2026/08...

softwaredoug.com

Three mistakes of new AI teams

My biased search expert perspective on where new AI teams run into expensive and unfortunate mistakes

042

Pedro Madruga @pedromadruga.com · 29/08/2026

I've had the Fairphone 6 for a year now and I'm in general quite satisfied. It took some time to go from iOS to Android, and I still want to try e/OS. www.theguardian.com/technology/2...

theguardian.com

Fairphone 6+ review: the most repairable, ethical phone gets faster

Dutch sustainable smartphone revamped with new chip and more RAM without sacrificing modular design

000

Pedro Madruga @pedromadruga.com · 28/08/2026

When the price per token of proprietary models stops being underpriced, then the edge is sitting on the side of the ones who are running local models.

100

Pedro Madruga @pedromadruga.com · 27/08/2026

pedromadruga.com/blog/domain-... Domain Driven AI

000

Pedro Madruga @pedromadruga.com · 26/08/2026

pedromadruga.com/blog/domain-... Domain Driven AI

000

Reposted by Pedro Madruga

Tom Aarsen @tomaarsen.com · 18/08/2026

🚨I've just released Sentence Transformers v6.0! MultiVectorEncoder joins the family: ColBERT-style late interaction models are now a first-class model type, for training, inference & interpretation, alongside dense, sparse & reranker models. Big thread 🧵

1246

Reposted by Pedro Madruga

Nathan Lambert @natolambert.bsky.social · 24/08/2026

Check out our latest open model data -- which models are mentioned in every arXiv ML paper since ChatGPT. A very fun weekend project with Codex! From idea is full dataset in <72hours. ~500K papers processed 😎 Plot: dashboard.interconnects.ai?axrange=all#...

5317

Pedro Madruga @pedromadruga.com · 24/08/2026

Just found out about this and it's such an interesting concept to read it together. Following the group closely now and sharing for wider reach. Awesome stuff! 👍

110

Pedro Madruga @pedromadruga.com · 23/08/2026

A better way of explaining GitHub's outages. Source: @benjdicken on twitter

001

Pedro Madruga @pedromadruga.com · 22/08/2026

AI development should follow domain knowledge. I'm calling it Domain Driven AI and I wrote about it. (No AI was used to write this blogpost which basically took me 4 weeks to write it 🥲) pedromadruga.com/blog/domain-...

000

Reposted by Pedro Madruga

Ethan Mollick @emollick.bsky.social · 21/08/2026

Its going to be an era of contradictions. Polls will show everyone hates AI overall but also everyone will secretly use AI all the time for lots of stuff. AI companies will be underwater in public opinion but also many people will feel strongly attached to their own favorite model & care about it.

1113012

Pedro Madruga @pedromadruga.com · 18/08/2026

Opencode VS PI VS claude code VS ... reminds me of the coices between Sublime VS vscode VS neovim VS ...

010

Pedro Madruga @pedromadruga.com · 17/08/2026

The twitter bubble of silicon valley is saying that Sol has surpassed Fable but I'd really like a better source of information about it.

010

Pedro Madruga @pedromadruga.com · 16/08/2026

Opencode looks good so far, although I have only been using it for a couple of days. I chose it over PI because I don't need a barebone option. Most importantly, I can easily switch from different providers and also between cloud and local models easily. Anthropic models seem to be lagging...

000

Pedro Madruga @pedromadruga.com · 15/08/2026

Luddites are at it against AI coding, but it feels they're less by the day. I wonder what's going to be their next victim.

000

Pedro Madruga @pedromadruga.com · 14/08/2026

For the past year my setup includes a combination of github worktrees, tmux, kitty tabs, with a configuration that syncs all of these together so I can plan, build and review in parallel. Similar to what herdr is, which I have tried now. Recommended.

github.com

GitHub - herdrdev/herdr: the runtime your coding agents live on

the runtime your coding agents live on. Contribute to herdrdev/herdr development by creating an account on GitHub.

000

Pedro Madruga @pedromadruga.com · 12/08/2026

If we.all agree that a lot of our learnings come from side projects and experiments, why not have teams increase doing this research? This would make people more knowledgeable and a benefit to any organization. Research and experiments are essential in teams work on AI-based products.

000

Pedro Madruga @pedromadruga.com · 11/08/2026

Something that I've stayed completely away from when using LLMs is analysing data from a dataset.

000

Pedro Madruga @pedromadruga.com · 03/08/2026

Stifling innovation will very rarely be a positive thing.

000

Pedro Madruga @pedromadruga.com · 17/07/2026

After SIGIR2025, I selected the best talks (imo) and summarized them using the Feynman Technique. The output was an overview explained in an incredibly straightforward way. I shared them all here github.com/pmadruga/sig... This year, I'm skipping SIGIR2026 but will probably do the same.

github.com

GitHub - pmadruga/sigir2025: Explained (relevant) papers from SIGIR 2025

Explained (relevant) papers from SIGIR 2025 . Contribute to pmadruga/sigir2025 development by creating an account on GitHub.

000

Reposted by Pedro Madruga

Olivia Guest · Ολίβια Γκεστ @olivia.science · 09/07/2026

you 👏🏼 cannot 👏🏼 automatically 👏🏼 detect 👏🏼 plagiarism > the available detection tools are neither accurate nor reliable and have a main bias towards classifying the output as human-written rather than detecting AI-generated text link.springer.com/article/10.1... 1/

link.springer.com

Testing of detection tools for AI-generated text - International Journal for Educational Integrity

Recent advances in generative pre-trained transformer large language models have emphasised the potential risks of unfair use of artificial intelligence (AI) generated content in an academic environme...

518862

Pedro Madruga @pedromadruga.com · 05/07/2026

Software engineering without innovation is just technical complacency.

000

Pedro Madruga @pedromadruga.com · 18/05/2026

MTP is wild. Just a few days after the llamacpp support.

000

Pedro Madruga @pedromadruga.com · 07/03/2026

Wait, what? source: www.anthropic.com/engineering/...

000

Pedro Madruga @pedromadruga.com · 02/03/2026

I need to know the inference speed on a raspberry pi because these sizes look great.

000

Reposted by Pedro Madruga

Sung Kim @sungkim.bsky.social · 01/03/2026

Rust implementation for Speech-to-Text based on Qwen3 models by Michael Yuan * Self-contained binary build — no external dependencies * Uses libtorch on Linux with optional Nvidia GPU support * Uses MLX on MacOS with Apple GPU/NPU support github.com/second-state...

github.com

GitHub - second-state/qwen3_asr_rs: Rust implementation of Qwen3-ASR automatic speech recognition

Rust implementation of Qwen3-ASR automatic speech recognition - second-state/qwen3_asr_rs

0353

Reposted by Pedro Madruga

Rasmus Aagaard @rasgaard.com · 20/02/2026

LLM cloud inference dominates usage, but should it? Local models and accelerators have improved massively over recent years. Perfect routing to best local model "reduce energy consumption by 80.4%, compute by 77.3%, and cost by 73.8% versus cloud-only deployment" arxiv.org/pdf/2511.07885

393

Pedro Madruga @pedromadruga.com · 20/02/2026

I went from Obsidian after a few years, to Notesnook, to Standard Notes. I am back using Obsidian. I can configure it enough to still keep it minimal while addressing all my use cases. I do notice that the Android version is slower than the iOS one. But manageable.

000

Pedro Madruga @pedromadruga.com · 19/02/2026

These are the engineers I like to work with. Boris is (again) on point.

000

Pedro Madruga @pedromadruga.com · 18/02/2026

A company's "AI Winter" starts when its AI-based product only has focus on engineering and zero on AI innovation. However, trying to avoid engineering debt creates AI debt. And AI debt has deeper (& longer) consequences due to rapid changes happening in the field. Aim for a balance, always.

pedromadruga.com

The art of (AI) science and engineering - an intro

A few thoughts about SIGIR 2025 and how to apply academic research by experimenting in the industry.

000

Pedro Madruga @pedromadruga.com · 17/02/2026

The claude code agent teams feature seems great in theory but there's an upper limit on the cognitive load of reviewing every change each agent makes. I'm still reviewing every single change, yes.

110