Mateus Freira

NunDB Maintainer, Principal Software Architect at @vidalung, and Running my own SaaS (E-belle) for +10 years as a side project.
I use vim BTW

[Mateus Freira @mateusfreira]

I just dropped my annual post about 'Secrets for Becoming a Better Developer'! Packed with 16 years of #DevExperience, Dive into automation tips, pro coding tricks, and productivity hacks. Perfect for devs eager to excel!👨‍💻 🧵

[Mateus Freira @mateusfreira]

That’s why I recommend that all React devs learn @elmlang. The ones who did became better React developers as a side effect

[Cult.Repo @CultRepo]

Does learning Clojure make you a better developer? @ericnormand puts forward a hypothesis... 😉

https://t.co/MHKySrFhYT

Watch the full Clojure documentary here: https://t.co/wtLedKJfIU

[Mateus Freira @mateusfreira]

It looks like we devs are never happy, but this is so true. I’m also annoyed by PR descriptions.
Now LLMs explain every single line of change, which makes the description useless.
Small direct to the point, desc or comment is all we need!

[eileencodes @eileencodes]

There's some irony in how I wished devs would write comments on their code and now that Claude does, I'm annoyed that it feels the need to write a dissertation on everything. I don't need 8 sentences on why this method early returns.

[Mateus Freira @mateusfreira]

@jotadepaula Ate a resposta parece script the Chat GPT PQP ... "não é isso é aquilo" ... vergonha alheia

[DROID @droidbuilds]

"mom, how did we get so poor?"

"your father had Claude Max, ChatGPT Pro, Cursor Pro and shipped absolutely nothing"

[Valerio Capraro @ValerioCapraro]

Students without access to LLMs are 2 to 8 times more creative than students with access.

That is the finding of a new paper comparing 2,200 college admissions essays written by humans before ChatGPT with essays generated by GPT-4.

The key point is not individual creativity. GPT-4 can write well, sometimes better than individual students. The problem is collective creativity.

Each new human essay added new semantic territory. New ideas. New angles. New experiences. New combinations.

Each new GPT-4 essay added much less.

The authors call this the diversity growth rate: how much novelty each additional text contributes to the collective pool of ideas.

Humans kept expanding the pool. GPT-4 made the pool converge.

Even when the authors pushed GPT-4 to be more creative, changed parameters, or used chain-of-thought prompting, the homogenizing effect remained.

This is the real danger of AI in education.
Not that students will write worse.
That everyone will write the same.

*
Full paper in the first reply

[Mateus Freira @mateusfreira]

@bridgemindai So it looks like I am not the only one

[Mateus Freira @mateusfreira]

Is it just me, or does @opencode burn way more tokens than Claude Code?

Love the UX, but my tokens disappear much faster with @opencode.

Is this workflow difference, or is @claudeai over taxing me for not using their UI?

[Mateus Freira @mateusfreira]

For off-grid lovers this is music to the ears. You no longer have to choose between living in a forest and working with the smartest people on Earth.

Starlink is freedom at its peak, looking forward for the new speed

[Elon Musk @elonmusk]

Starlink V3 satellites have >10X bandwidth of V2 and there’ll be >10X launched, which means >100X more bandwidth.

Also, altitude will be 350km vs 550km, so min latency can be cut in half.

Light travels 300km/ms in space, so physics round trip min latency drops to <5ms.

[Mitchell Hashimoto @mitchellh]

I've got an agent in a loop optimizing a renderer with the goal to minimize frame times (and tests to measure). It got times down from 88ms to 2ms and allocations down from ~150K to 500. Sounds good, right? Wrong. This is exactly why agent psychosis is a big fucking problem.

As an experiment, I rewrote the Ghostty core render state in Go, with access to identically laid out data structures as Ghostty and the exact same validation tests. I made a purposely naive renderer (simple, correct, but slow). 88ms per frame with 150,000 allocations (horrendous, lol)!

I then kickstarted a Ralph loop to bring the frame times down. I told it it can't modify input data structures or the public API or tests (they're correct), but it can do anything else it wants. It got to work.

It has worked for about 4 hours. I've spent around $350 on this experiment so far. The results?

88ms => 1.5ms
150K allocs => ~500 allocs

Incredible right? Nope.

My hand-written renderer I ported has frame times (same benchmark) of ~20us (0.020ms) and 0 allocations in the update path.

This is the problem with psychosis and lacking systems understanding. If you don't understand the system, you're going to accept that this is an incredible result. If you understand the system, you'll see better solutions immediately and can do roughly 75x better on throughput.

The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity.

Standard disclaimer: I use AI all the time. I like AI. The point I'm making is to not blindly accept results. Think. Analyze. Learn.

[Loris Cro ⚡ @croloris]

Andrew's interview with JetBrains

https://t.co/7SsmfjksJP

[@levelsio @levelsio]

When you realize billionaires can't even buy luxury products

You realize luxury products are really just meant for poor people to fake they're rich

[Mateus Freira @mateusfreira]

That is a good idea ^, I did that +10 years with e-belle (Scala project), build still works today fast and 100% reliable.

Looks like it may became the norm, for the good of the dev community

[Mitchell Hashimoto @mitchellh]

Fork your dependencies, trim them to only your use case, never update unless it breaks for your users. I’ve been vocal about this for 10+ years. I’ve always said that updating is way riskier than latent bugs (which can be tracked and CVEs monitored).

If you are updating a dependency, it’s on you to analyze every single commit in the full transitive set of dependencies. If you dont see anything compelling, dont update!

I remember at HashiCorp once in awhile an engineer would try to update a dep or replace a DIY lib with an external one and id always ask “show me the commit we need.” Dont update for the sake of it.

Feeling pretty swell about this mentality with all the supply chain attacks happening.

[Mateus Freira @mateusfreira]

The age of AI coding will bring down more “senior engineers” than juniors.
What are the excuses now to deliver poor, untested code that barely works? “I have not had time” is not acceptable anymore!

[Mateus Freira @mateusfreira]

Man, I love open source. This is MBA level material out in the open for anyone to research and learn from.

Keep an eye on Bun, Rust, and the Zig community. I’ve wanted to write something big in Zig for a long time ... Maybe now is the time to finally give it a chance

[dax @thdxr]

we are impacted by bun stability more than almost anyone, anything they're doing to try and improve that is good

simultaneously we're going to wait a long time before considering moving a million daily users to the rust version

so we're stuck for now

[Mateus Freira @mateusfreira]

@jotadepaula vim FTW! Meu workflow mudou quase 0… só que agora tem uma tab a mais do tmux com o Claude do lado do nvim.

[Mateus Freira @mateusfreira]

Buy farmable land, ship a feature, leave the office, and go plant or water something! Repeat!

[Chris Power @typecraft_dev]

Buy an outdoor rug. Go into debt if you have to

[Mateus Freira @mateusfreira]

@govbr Uma vergonha, oportunismo eleitoral na cara dura. Se “o Governo do Brasil tá do lado do povo brasileiro” agora, então primeiro pede desculpa publicamente e assume que “o Governo do Brasil tava contra o povo brasileiro” quando criou essa taxa ridícula.

[Free Cities Foundation @freecitiesfound]

Brazil already has a Free City.
It's been operating quietly for 45 years.

@palomalecheta on the new Free Cities Podcast...

https://t.co/cs5iYbCW3C

[Mateus Freira @mateusfreira]

@DNAutics I don’t think you understand open source at all, @DNAutics. Sometimes it’s better to keep an opinion to yourself.

People contribute because they want to, whether for fun, to learn, or simply because they care about the project

[Evan You @evanyou]

No server components in Vue was intentional