Fabian

26 | ML Engineer | Software Consultant | Asperger | ADHD | Passions: AGI, Data Science, Distributed Systems, Event Sourcing, HPC, Research, Systems Programming

[Fabian @Fabian_Domenech]

I've been pondering on life a lot this year. I realized:

1. Retiring at age 40 is the same as retiring at 65
2. Retiring at age 30 is too early and will lead towards self-destructive hedonism
3. I should enjoy my journey to retirement
4. I should NOT make retirement a goal

[Fabian @Fabian_Domenech]

@_xjdr I'm interested 🤔

As a software architect with ML research experience, building intelligence infra in an ambitious project sounds like meaningful and exciting work, so I'm definitely interested

[HSVSphere @HSVSphere]

The ideals of Nix are almost perfect. The implementation sucks and it's only really usable if you know why it is the way it is (and why it's badly implemented).

That's why I'm working on Cab, and I'm taking my time thinking everything through before committing to an implementation.

It's not exactly a build language either. It doesn't specialize in any concept such as derivations, units, resources or whatever. It only lets you compose expressions with contexts—the stuff that makes Nix magical in the first place.

What are expression contexts? They're when a subset of an expression implicitly carries the whole expression with it, as a context. It's how you don't explicitly specify what a Nix derivation depends on—it just works!

The problem with Nix, ignoring all the QoL stuff that's missing [1] is that Nix contexts can *only* be used for derivations. They're not generic. derivationStrict is a builtin function, you cannot emulate it in the language itself.

That prevents you from using contextful expressions for other things, such as process management. Resource (like terraform) management. Literally anything that forms a graph - it cannot be done cleanly!

This is why a generic contextful-expression language is required. Cab will fix this, and that won't be the only thing it will fix. Cab has structural types enforced with a super novel system that works super well for a dynamic build system language. It has patterns, rather than hard coding for identifiers. I'd argue that its type system is going to be more powerful than typescript since there is no runtime-comptime difference (it's all the same, there's no IO either so it's simple).

Anyway, stay tuned for an MVP. I also plan on supporting using the Nixpkgs package set with the system I'm going to build on top of Cab (the Cull Build System) using the efforts the Ekala project has been going through, eventually.

However, I won't announce it until I have a working LSP, a unified documentation system that's flexible, a properly designed, hack-free "project" abstraction for the Cull Build System and a fast runtime overall (bye bye, NixOS module system and home-manager eval times).

[1] - Flakes are a bad solution to the purity problem, nix-* commands shouldn't exist, FODs are hard to use, nixConfig can pwn you, too many tunables, too little separation of concerns between the distro, module system, nix itself, the nix daemon, literal oddities in nix, much much much more

[public_intellectual @publicinte]

https://t.co/pvzdGbWkL0

[Fabian @Fabian_Domenech]

100%

My 1st Rust project was a WASM module that would parse GraphQL schemas for my npm package that would then use said info to codegen TS types

I knew offloading the parsing from TS to Rust was overkill and completely unnecessary 😅 but I went ahead anyways and I'm glad I did

[zack @zack_overflow]

If you actually want to learn new programming languages, you need to be way more ambitious with your first projects.

My 1st Rust project: compiler for Australian programming language

My 1st Zig project: Bytecode VM for my programming language

[Fabian @Fabian_Domenech]

@crypt0x_0 Yep 😅

With Trump putting tariffs on semiconductors from Taiwan, China is now in a more favourable position to invade Taiwan since the US is now less likely to defend it

[Fabian @Fabian_Domenech]

@tomieinlove Yep, sounds like tpot 😂

It's crazy how worked up some people get over what languages others like/dislike 😅

I love Rust, but it ain't perfect, so when tpot/moots trash on it, I usually ignore it or agree/disagree with nuance because lashing out and throwing insults is dumb af

[Fabian @Fabian_Domenech]

True

Regardless of how adamant you are that wfh is beneficial in your circumstance, your employer can simply ignore you until you quit or fall in line

However, your employer will definitely work something out with you if you become someone they deem worth making exceptions for

[dax @thdxr]

if you like wfh you're not going to keep it by whining about companies requiring in office

get good enough that companies will make an exception for you - this is how people did it before covid and it'll always be on the table

"but i like doing chores" is not going to cut it

[Fabian @Fabian_Domenech]

Since ChatGPT's first release, this has been my stance

It's simply illogical and expensive to build an AI startup fully reliant on other companies' LLM APIs

Investing in the hardware and infra to self host your own LLM allows you to more easily innovate while keeping costs down

[kache @yacineMTB]

if you want to do anything interesting with an LLM you have to self host

[Fabian @Fabian_Domenech]

@0xluffy Bruh, for almost 2 years, Docker Desktop on my Mac would often randomly reset my Docker settings which would make pulling Docker images not work and would make the Docker Desktop dashboard crash when opened

That's just one of the billion issues I've had using that garbage 🙃

[Fabian @Fabian_Domenech]

They're basically like an underdog, so they receive less hate by default

OpenAI being known as the "number one" company in relation to AI will naturally draw everyone's warranted/unwarranted negative sentiments with anything related to AI affecting them

That said, Anthropic is much more open with their service and the research projects they conduct, which directly contrasts OpenAI's corporate vibes with their extreme secrecy and overhyping

Basically, regardless of whether Anthropic is truly better than OpenAI or not, OpenAI has a bad reputation similar to Meta or Google and they operate with extreme secrecy, so people will opt to hate OpenAI more

[Fabian @Fabian_Domenech]

@ludwigABAP Yep

Since college I've been coding shit up to see how stuff works, to stay up to date and to learn new stuff I find interesting

Only now in hindsight do I realize how much my skills have grown by the slow incremental knowledge buildup and skill acquisition of constant practice

[Fabian @Fabian_Domenech]

@yacineMTB That amount of growth in such a short time is insane

People will cope and chalk it up to luck and constant shitposting, but the truth is that reaching such heights is more achievable than it seems with enough dedication, life experience and delusional levels of confidence

[Nous Research @NousResearch]

You can now experiment with Psyche’s DisTrO-enabled training code on our GitHub, and the larger open-sourced distributed training stack will be released alongside testnet.

https://t.co/7W9gLqtiap

[Nous Research @NousResearch]

Recent AI breakthroughs challenge the status quo narrative that only closed, mega labs have the ability to push the frontier of superintelligence.

Today we announce Nous Psyche built on @Solana - a cooperative training network for generative AI. Psyche coordinates heterogeneous hardware to join a run and train open-source models.

We retell the myth of Psyche — a mortal’s quest for retribution against divine odds:

[Fabian @Fabian_Domenech]

Went MIA for 2 months as I had to lock in for a contract I secured on November

I got to experiment a lot with Nix in dev shells and CI/CD pipelines using NixOS VPSs and Colmena (which is great btw)

Like Rust, Nix has a HUGE learning curve, but it's worth the time investment imo

[Fabian @Fabian_Domenech]

@HSVSphere With how fast the field evolves, how opinionated people are and the trillion ways software projects can be built, it's just inevitable that all code will end up sucking in some way after some time

It's how we adapt to that fact that might make our code suck less than average 😅

[Fabian @Fabian_Domenech]

@wordgrammer Probably

The key to relevancy is to continually learn and improve how you solve problems through software

Programming has been the go to method for decades, but it doesn't mean it will always be that way

Regardless of whether it becomes obsolete or not, I'm excited to adapt

[Fabian @Fabian_Domenech]

Very true

Most people don't realize how much ego they attach to their opinions and how avoidant they naturally are to things they subconsciously know will challenge them

Acknowledging that ego is step one towards being able to share ideas to seek feedback and not confirmation

[sarah @sarahtonin]

The goal isn't to appear knowledgeable; it's to actually become knowledgeable. A remarkably effective way to expedite learning is by speaking up, sharing your ideas honestly, and seeking feedback - rather than shielding yourself from criticism through silence.

[Fabian @Fabian_Domenech]

Furthermore, you gotta glaze your devs too in lead roles so they feel valued and build confidence in themselves and trust in you as a lead, but you gotta know how to glaze

Empty platitudes? They'll feel undervalued
Too technical? They'll feel one-upped

Study how & what to glaze

[tuna🍣 @tunahorse21]

Glazing in a professional context is a must-have skill. Eventually, you reach a level where you don't need to, but the reality is you need the right amount of glaze.

Too much? Cringe
Too little? Get passed up for a promo

Study the right amount of glaze

[Fabian @Fabian_Domenech]

@wordgrammer Those steps are basically Lean & Agile ideals

Although some companies have had success following them, which usually resembles Continuous Delivery when applied irl using engineering principles, the fact that they're so open to interpretation usually leads to misguided execution