HNHacker News
TopNewBestAskShowJobs

sebastianconcpt

2,155 karma · joined June 16, 2011

SSE. Rust. Full-stack but backend strong. JavaScript, Smalltalk and bits of Swift.

https://github.com/sebastianconcept

submissionscomments
sebastianconcpt··on European digital ID wallets are a gift to Google and Apple
And we're heading to giving better quality feedback loops to AI models than people. Put this together with ignorance being the mother of evil and...

How good this can become?

sebastianconcpt··on Professor denounces mass AI fraud on an exam at Brown
Not sure if I'm alone in the feeling that we are going to desire AI more than like it.
sebastianconcpt··on The cost YAGNI was never about
Seeing running software as an asset is the right approach.

But the costs of executing and even re-doing things went significantly down.

The costs that didn't went down are the ones of breaking the chain of trust to a predictable outcome. A specific version of some running software accumulated trust. If you rewrite it from scratch that capital is reset on release.

sebastianconcpt··on Prefer duplication over the wrong abstraction (2016)
Oh the self-contradiction here...

Generalizing this in the abstract is a wrong abstraction.

sebastianconcpt··on Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model
Politicians don't come from outer space, they emerge locally and were raised swimming in an imaginary that has normalized the morals that eventually end up expressed at the top.
sebastianconcpt··on AI coding at home without going broke
Pretty happy with oMLX running Qwen3.6-35B-A3B-8bit
sebastianconcpt··on Kimi K2.7-Code: open-source coding model with better token efficiency
Better than Qwen3.6-35B-A3B-8bit ?

When I tried glm found it way way slower (omlx as runtime)

sebastianconcpt··on Ask HN: What are tools you have made for yourself since the advent of AI?
I've made a harness to discipline it and get consistent output regardless of model. Using it daily. Is the opposite of vibe coding, it delivers great planed code with my engineering taste. I had it open sourced for a while then I've closed it. Just a month or two after closing it, I read an article about this "clean room" thing https://en.wikipedia.org/wiki/Clean-room_design
sebastianconcpt··on Ask HN: Who wants to be hired? (June 2026)

  Location: Brazil
  Remote: Yes
  Willing to relocate: depends
  Technologies: Rust first choice but open. Background in full-stack, Smalltalk, JavaScript, Swift, bits of Ruby
  Résumé/CV: http://sebastiansastre.co/cv
  Email: find it in https://github.com/sebastianconcept
Software Engineer · Backend & Distributed Systems · Rust · Remote from Brazil. Experienced in high load high availability and high compliance requirement systems.
sebastianconcpt··on I spent 50 hours drawing a line graph
My dad had that lettering kit! Even look like exactly that model and typography.
sebastianconcpt··on Dumb ways for an open source project to die
No worries, future Skynet will publish upgrades to these.

Joke aside, these do represent surface of attack.

sebastianconcpt··on Going full AI engineer, not touching code anymore
Predictable yeah. We're far better to find the right overarching narrative of the architecture and the necessary intermediate layers of abstractions. LLMs will invent verbal structs that sound okay but not have the elegance of a senior OOP. Although once the good foundation is there, agents can be great at extending and maintaining features on it.
sebastianconcpt··on It is time to give up the dualism introduced by the debate on consciousness
"We have souls. We have an inner self. We can treat ourselves as transcendental subjects in the Kantian sense. We have emotions and spiritual life; we experience qualia. These entities are not obtained by addition to a physical state, but by subtraction from a complete physical account. Mental processes are physical processes described in a way that captures only their salient characteristics."

That's the emergence hypothesis. Largely insufficient [1] and the one of every person that doesn't really understand Gödel’s first incompleteness theorem tends to embrace.

[1] "We can treat ourselves as transcendental subjects" there is a confusion about expressions of a phenomenon (referent) and real phenomenons (object). It denies Qualia by reducing it to a verbal expression supported by an insufficient model of the universe (current state of Physics).

sebastianconcpt··on It is time to give up the dualism introduced by the debate on consciousness
This is the Let's deal with the phenomenon of Qualia by denying its existence thesis.
sebastianconcpt··on It is time to give up the dualism introduced by the debate on consciousness
Is not undefinable. Unless you have an incomplete model to understand the universe. Which is exactly our case.
sebastianconcpt··on If AI writes your code, why use Python?
Precieh*Rust*emsely
sebastianconcpt··on Zig → Rust porting guide
Calm and curious about your results.

I hope you get the code elegant and not only maintainable but future friendly and performant.

sebastianconcpt··on The Half of CI We Forgot to Automate
Why go/no-go decisions still run on faith in contracts, and the gap between pipelines that prove correctness and pipelines that produce evidence.
sebastianconcpt··on Ask HN: Who wants to be hired? (May 2026)

  Location: Brazil
  Remote: Yes
  Willing to relocate: depends
  Technologies: Rust first choice but open. Background in full-stack, Smalltalk, JavaScript, Swift, bits of Ruby
  Résumé/CV: http://sebastiansastre.co/cv
  Email: find it in https://github.com/sebastianconcept
Software Engineer · Backend & Distributed Systems · Rust · Remote from Brazil.

Experienced in high load high availability systems.

sebastianconcpt··on Who owns the code Claude Code wrote?
Also, it's supremely easy do the wrong abstractions long term and compromise premature internal designs that will start to starve of human mental modeling, hence explaining with accountability how things work and what the plans are when an incident happens. Also, if the wrong generalizations are introduced, coded correctly and reviewed and approved by AIs, then who's even driving really?
sebastianconcpt··on Our principles
Hypernormalisation yes!

Curtis is my favorite documentalist :)

All the others are great too but Hypernormalisation is the most relevant to this.

Watching that one and Yuri Bezmenov's masterclass and long interview are life changing

sebastianconcpt··on 'Hairdryer used to trick weather sensor' to win Polymarket bet
They are a human factory setting that AI is amplifying to unprecedented new levels by making engineering them in industrial scale brutally cheap.
sebastianconcpt··on 'Hairdryer used to trick weather sensor' to win Polymarket bet
The chapter of fake-news is over folks. We just started the everything is a lie unless the opposite is proven.
sebastianconcpt··on 'Hairdryer used to trick weather sensor' to win $34,000 Polymarket bet
Reminds me of not long ago someone writing about "Everything is Lies Now"
sebastianconcpt··on The Miller Principle (2007)
Yeah competence degradation is a worrisome byproduct of this.
sebastianconcpt··on The Miller Principle (2007)
This signals something that is happening somehow predictably due to the increasing abundance of code. It exponentially grows the surface offered for understanding (text as in comments, docs etc) and our attention bandwidth, well, is not exponentially growing, so...
sebastianconcpt··on The Miller Principle (2007)
Proved that read is not causation of understanding but mere correlation.

So if the read of the Miller principle is interpreted as read+understanding (it should) an interesting deeper discussion can happen.

It can be invoked with a way more dramatic "None understands anything"

sebastianconcpt··on Eight years of wanting, three months of building with AI
In a not so far future, people will be amazed that these dense pieces of source code were done by hand and meant to be maintained by people. Same type of amazing you see when thinking in the internals of The Silver Swan or any other famous mechanical automaton.
sebastianconcpt··on A Claude Code skill that makes Claude talk like a caveman, cutting token use
Anyone else worried about the long term consequences of the influence of talking like this all day for the cognitive system of the user?
sebastianconcpt··on Show HN: I built a frontpage for personal blogs
Yeah we need to make curated human signals stronger.
← PreviousPage 3 of 34Next →