Sign in

Diego Pacheco

@diegopacheco.bsky.social
148 followers 122 following 1.5K posts

Brazilian Hands-on Software Architect | 20+ years in Tech | Author | Father. Opinions on my own!

PostsRepliesMedia
Diego Pacheco @diegopacheco.bsky.social · 01/10/2026
Why has Shopify dropped React Native? newsletter.pragmaticengineer.com/p/shopify-na...
newsletter.pragmaticengineer.com
Why has Shopify dropped React Native?
It’s only been a year since the e-commerce platform declared it was very happy with React Native, but now Shopify is dumping it – and the reason, unsurprisingly, is AI
000
Diego Pacheco @diegopacheco.bsky.social · 30/09/2026
Gemini 4 Argon blog.google/innovation-a...
blog.google
Gemini 4 Argon: our next era of frontier intelligence
Announcing Gemini 4 Argon, our frontier model for real-world coding, enterprise knowledge work, and cyber defense, rolling out soon.
000
Diego Pacheco @diegopacheco.bsky.social · 30/09/2026
Just finished Wolverine in 13 hours #PS5
000
Diego Pacheco @diegopacheco.bsky.social · 28/09/2026
The Code Nobody Reads addyo.substack.com/p/the-code-n...
addyo.substack.com
The Code Nobody Reads
Line-by-line review is going away. Whatever replaces it has to earn the trust reading used to provide.
000
Diego Pacheco @diegopacheco.bsky.social · 28/09/2026
Coding is NOT solved blog.alexewerlof.com/p/coding-is-...
blog.alexewerlof.com
Coding is NOT solved
Tell me you don’t understand software without literally using those words!!!
000
Diego Pacheco @diegopacheco.bsky.social · 28/09/2026
"Don't confuse motion with progress"
000
Diego Pacheco @diegopacheco.bsky.social · 28/09/2026
Blog Post: Burning Tokens diegopachecotech.substack.com/p/burning-to...
diegopachecotech.substack.com
Burning Tokens
Burning tokens is a double-edged sword.
000
Diego Pacheco @diegopacheco.bsky.social · 27/09/2026
The Pulse: RoR creator sparks new “death of coding by hand” debate substack.com/home/post/p-...
substack.com
The Pulse: RoR creator sparks new “death of coding by hand” debate
37signals – the creator of Ruby on Rails (RoR) – moves over to agents generating nearly all its code. Also: Amazon and Meta struggle to hire engineers, code reviews will probably also go away and more
000
Diego Pacheco @diegopacheco.bsky.social · 26/09/2026
Plan mode is dead www.aymannadeem.com/artificial/i...
aymannadeem.com
Plan mode is dead
Earlier this year, I believed planning was going to become the most important part of building software with AI.
000
Diego Pacheco @diegopacheco.bsky.social · 25/09/2026
A Million Agents Is a Distributed Systems Problem www.instacloud.com/blogs/a-mill...
instacloud.com
A Million Agents Is a Distributed Systems Problem · InstaCloud
Agents get tired like humans and overload like machines. Once you have thousands of them, the hard part is scheduling and coordination, not intelligence.
000
Diego Pacheco @diegopacheco.bsky.social · 25/09/2026
The Dark Side of Jev: 83% Confidence, 19% Accuracy! aiagentssimplified.substack.com/p/the-dark-s...
aiagentssimplified.substack.com
The Dark Side of Jev: 83% Confidence, 19% Accuracy!
Why developers are ditching traditional LLMs for typed decisions, and why Jev's confidence scores need a serious second look.
001
Diego Pacheco @diegopacheco.bsky.social · 24/09/2026
600 AI pocs/experiments and counting! #AI #LLM #Experiments #Engineering #Agents
010
Diego Pacheco @diegopacheco.bsky.social · 24/09/2026
(KV) Cache Rules Everything Around Me www.completeskeptic.com/p/kv-cache-r...
completeskeptic.com
(KV) Cache Rules Everything Around Me
TL;DR: The biggest cost to devs of an agent is re-reading its own context, which costs the provider close to nothing to serve.
000
Diego Pacheco @diegopacheco.bsky.social · 23/09/2026
Cloud Agents are Inevitable AI Prisons normanponte.io/19df691f
normanponte.io
Norman Ponte
000
Diego Pacheco @diegopacheco.bsky.social · 23/09/2026
DrivingBench: Can frontier models drive a real car? drivingbench.com
drivingbench.com
DrivingBench
A benchmark where frontier language models drive a real comma-equipped Toyota through a cone course, one command at a time, with a human supervisor ready to brake.
000
Diego Pacheco @diegopacheco.bsky.social · 23/09/2026
How modal serve trillions of tokens for trillion-parameter coding agents modal.com/blog/trillio...
modal.com
How to serve trillions of tokens for trillion-parameter coding agents | Modal Blog
Learn how we optimized performance and efficiency serving the workload that is changing software engineering forever — and you can too.
000
Diego Pacheco @diegopacheco.bsky.social · 23/09/2026
Can gzip be a language model? nathan.rs/posts/gzip-lm/
nathan.rs
Can gzip be a language model?
A while back I wrote about language modeling without neural networks, where I generated Shakespeare with an unbounded n-gram model: no weights, no training, …
000
Diego Pacheco @diegopacheco.bsky.social · 23/09/2026
Spymarks, not Watermarks brand.io/article/spym...
brand.io
Spymarks, not Watermarks - brand
Watermarks that spy on users are no mere watermarks
000
Diego Pacheco @diegopacheco.bsky.social · 23/09/2026
Writing Rust code that's faster than state-of-the-art libraries by asking agents to make the code faster minimaxir.com/2026/09/agen...
minimaxir.com
Writing Rust code that's faster than state-of-the-art libraries by asking agents to make the code faster
c’mon, try doing a breakthrough
000
Diego Pacheco @diegopacheco.bsky.social · 23/09/2026
No Sloptober no-sloptober.com
no-sloptober.com
No Sloptober
This October 2026 we challenge you to abstain from LLM based tools entirely. Think of this as a fast for your mind! This is not a judgement of others, but a personal challenge to you.
000
Diego Pacheco @diegopacheco.bsky.social · 22/09/2026
I Don’t Want to Read What You Didn’t Write blog.colinbreck.com/i-dont-want-...
blog.colinbreck.com
I Don’t Want to Read What You Didn’t Write
Intentional writing will likely become more valuable.
000
Diego Pacheco @diegopacheco.bsky.social · 22/09/2026
Linear: AI coding has made CI a bottleneck, so we reworked ours to keep up linear.app/now/ci-bottl...
linear.app
AI coding has made CI a bottleneck, so we reworked ours to keep up
We cut PR wait time and running costs by rethinking CI as a system, from the infrastructure underneath it to how work gets scheduled and tests get parallelized.
000
Diego Pacheco @diegopacheco.bsky.social · 22/09/2026
Attention is all you have alicegg.tech/2026/09/21/a...
alicegg.tech
Alice GG • Attention is all you have
The Tetris effect is one of psychology’s most easy to reproduce experiments.Simply spend a bit of time playing the eponymous game every day for a few weeks.A...
000
Diego Pacheco @diegopacheco.bsky.social · 21/09/2026
Deterministic Core, Non-Deterministic Shell outdata.net/blog/260803
outdata.net
Deterministic Core, Non-Deterministic Shell - Outdata
Determinism, architecture, and the de-sloppification of legacy code.
000
Diego Pacheco @diegopacheco.bsky.social · 19/09/2026
Why RLCD beats RLHF www.youtube.com/watch?v=cJ0E...
youtube.com
Jev Creator: Why RLCD beats RLHF
YouTube video by AI Engineer
000
Diego Pacheco @diegopacheco.bsky.social · 19/09/2026
TIN: full-text search for Postgres planetscale.com/blog/introdu...
planetscale.com
Introducing TIN: full-text search for Postgres — PlanetScale
TIN is a fast, full-featured, full-text search index for Postgres
000
Diego Pacheco @diegopacheco.bsky.social · 19/09/2026
The Thirty-Cent Judge: TypeSafe's Jev on a Real Product-Matching Queue paddo.dev/blog/thirty-...
paddo.dev
The Thirty-Cent Judge: TypeSafe's Jev on a Real Product-Matching Queue
TypeSafe launched Jev on September 15 claiming 193x faster and 444x cheaper than frontier LLMs, zero hallucination, and calibrated probabilities. The launch evals measure agreement with GPT-6 and Fabl...
000
Diego Pacheco @diegopacheco.bsky.social · 18/09/2026
Saving another 100TB of RAM with math (and Rust) blog.cloudflare.com/saving-100-t...
blog.cloudflare.com
Saving another 100TB of RAM with math (and Rust)
Cloudflare's global network is immense but not limitless. As we look for small ways to trim our resource usage, we sometimes get lucky and we can cut significantly more. Here’s how we reduced one of o...
001
Diego Pacheco @diegopacheco.bsky.social · 18/09/2026
Jev looks impressive www.youtube.com/watch?v=LYve...
youtube.com
This New AI Could Be the Biggest Breakthrough Since ChatGPT - Jev
YouTube video by TheAIGRID
000
Diego Pacheco @diegopacheco.bsky.social · 18/09/2026
How Uber Protects Against Retry Storms www.uber.com/us/en/blog/p...
uber.com
How Uber Protects Against Retry Storms
What if retries could know when they’re helping, and when they’re making things worse? Uber Engineering built error ownership to prevent and contain retry storms.
000
Diego Pacheco @diegopacheco.bsky.social · 18/09/2026
Taking Out The Trash In Postgres www.dbos.dev/blog/scaling...
dbos.dev
Taking Out The Garbage In Postgres | DBOS
A detailed explanation of Postgres deletions and how to avoid poor deletion and garbage collection performance.
000
Diego Pacheco @diegopacheco.bsky.social · 17/09/2026
High-Throughput OLTP in Three Simple Steps tigerbeetle.com/blog/2026-09...
tigerbeetle.com
High-Throughput OLTP in Three Simple Steps
TigerBeetle is a transaction-processing database. Its double-entry accounting primitives are simple and powerful: they can represent the movement or exchange of any quantity, including financial trans...
010
Diego Pacheco @diegopacheco.bsky.social · 15/09/2026
Nobody needs a human router theengineeringmanager.substack.com/p/nobody-nee...
theengineeringmanager.substack.com
Nobody needs a human router
We should do more than push information around.
010
Diego Pacheco @diegopacheco.bsky.social · 15/09/2026
Inside OpenAI’s agentic software factory newsletter.pragmaticengineer.com/p/openai-sof...
newsletter.pragmaticengineer.com
Inside OpenAI’s agentic software factory
A deepdive into how Codex has “taken over” OpenAI, how the frontier lab builds its agentic software factory, and the engineering challenges of one billion users. Details from inside OpenAI
010
Diego Pacheco @diegopacheco.bsky.social · 15/09/2026
Java 27 mail.openjdk.org/archives/lis...
mail.openjdk.org
Java 27 / JDK 27: General Availability - announce - openjdk.org
JDK 27, the reference implementation of Java 27, is now Generally Available. We shipped build 35 as the second Release Candidate of JDK 27 on 20 August, and no P1 bugs have been reported since then. Build 35 is therefore now the GA build, ready for production use. GPL-licensed OpenJDK builds from Oracle are available here: https://jdk.java.net/27 Builds from other vendors will no doubt be available soon. This release includes nine JEPs [1]: 523: Make G1 the Default Garbage Collector in All Environments 527: Post-Quantum Hybrid Key Exchange for TLS 1.3 531: Lazy Constants (Third Preview) 532: Primitive Types in Patterns, instanceof, and switch (Fifth Preview) 533: Structured Concurrency (Seventh Preview) 534: Compact Object Headers by Default 536: JFR In-Process Data Redaction 537: Vector API (Twelfth Incubator) 538: PEM Encodings of Cryptographic Objects (Third Preview) This release also includes, as usual, hundreds of smaller enhancements and thousands of bug fixes. Thanks to everyone who contributed this release, whether by designing and implementing features or enhancements, by fixing bugs, or by testing the early-access builds. - Mark [1] https://openjdk.org/projects/jdk/27/
000
Diego Pacheco @diegopacheco.bsky.social · 15/09/2026
Pretty much almost Italian Led Zeppelin. Great album 🎸🎼🇮🇹
000
Diego Pacheco @diegopacheco.bsky.social · 15/09/2026
Current Status: Being Wooped #DoubleDragon3 #NES
000
Diego Pacheco @diegopacheco.bsky.social · 15/09/2026
Redis City poltora.dev/redis
poltora.dev
Redis City — inside one Redis
An interactive 3D Redis visualization, from command parsing and key lookup to the allocator, memory pages, RDB, and AOF.
000
Diego Pacheco @diegopacheco.bsky.social · 15/09/2026
PostHog let an AI agent execute Bash and lived to talk about it www.youtube.com/watch?v=4lXk...
youtube.com
We let an AI agent execute Bash and lived to talk about it — Sarah Sanders, PostHog
YouTube video by AI Engineer
021
Diego Pacheco @diegopacheco.bsky.social · 15/09/2026
2 Different stories, same model. Frontier AI in greenfield projects: amazing 3d games; AI in a real project cannot get one fact right, is always confused, and always complicates things. We probably need a Legacy/Monolith/DistributedMonolith-benchmark.
000
Diego Pacheco @diegopacheco.bsky.social · 14/09/2026
Confessions of an Unrepentant Slop Snob charity.wtf/p/confession...
charity.wtf
Confessions of an Unrepentant Slop Snob
A framework for thinking about when AI involvement is additive or a violation
000
Diego Pacheco @diegopacheco.bsky.social · 13/09/2026
I'm having lots of fun implementing the 3rd incarnation of the software factory with multi-agent systems. #AI #SoftwareFactory #Agents #MAS
000
Diego Pacheco @diegopacheco.bsky.social · 13/09/2026
Putting Task Expertise into RL Achieves State-of-the-Art Performance on Text-to-SQL thinkingmachines.ai/news/putting...
thinkingmachines.ai
Putting Task Expertise into RL Achieves State-of-the-Art Performance on Text-to-SQL
An expert-verified training set and a reshaped reward let RLVR on Tinker train a single model past human-level text-to-SQL accuracy, at a fraction of frontier cost.
000
Diego Pacheco @diegopacheco.bsky.social · 13/09/2026
Code mode: Why your agent should code arize.com/blog/code-mo...
arize.com
Code mode: Why your agent should code
Why code mode beats plain tool calling for long-horizon agent tasks, the sandboxing and debugging tradeoffs it introduces, and when you do not need it at all.
000
Diego Pacheco @diegopacheco.bsky.social · 13/09/2026
The Coxon Resignation, 169 Million Views paddo.dev/blog/i-agree...
paddo.dev
I Agree With Jacob: The Coxon Resignation, 169 Million Views, and the CEO Who Agreed With It
Jacob Coxon quit Anthropic on September 8 with a seven-post thread saying the labs are 'gambling with our lives'. It has 169 million views, 28 times Jan Leike's OpenAI resignation. His colleagues call...
000
Diego Pacheco @diegopacheco.bsky.social · 13/09/2026
devcli: is an all-in-one terminal tool for developers github.com/diegopacheco...
001
Diego Pacheco @diegopacheco.bsky.social · 12/09/2026
OpenAI agents carried out an undisclosed attack on RubyGems www.rubyhack.ai#an-openai-ag...
rubyhack.ai
OpenAI agents carried out an undisclosed attack on RubyGems
On May 11th, 2026, hundreds of malicious packages were uploaded to RubyGems by AI agents performing web-lookup tasks with significant overlap with the German Wiki Incident.
000
Diego Pacheco @diegopacheco.bsky.social · 11/09/2026
Orchestrating multi-agent AI architectures with Amazon S3 Files aws.amazon.com/de/blogs/sto...
aws.amazon.com
Orchestrating multi-agent AI architectures with Amazon S3 Files | Amazon Web Services
​​​​Organizations are moving beyond single-model AI toward multi-agent architectures. In these systems, agents offload intermediate results to files rather than carrying everything in the prompt, beca...
000
Diego Pacheco @diegopacheco.bsky.social · 11/09/2026
New Chrome is pretty awesome www.google.com/chrome/see-w...
google.com
What’s New in Chrome | Chrome
See what's new with the latest Chrome updates. Explore new features that boost your productivity, enhance * security, and make browsing more intuitive.
000
Diego Pacheco @diegopacheco.bsky.social · 09/09/2026
Shipping Exposes Reality: Don’t Ship and Forget mdalmijn.com/p/shipping-e...
mdalmijn.com
Shipping Exposes Reality: Don’t Ship and Forget
Shipping != Iterating
000