Librarian

Engineering Interoperability of Financial Meaning. Knowledge Governance as Code.

[Librarian @PeculiarLi34455]

We provide an executable, governed semantic state in which data, entities, provenance, rules, missions, and agents are first-class participants and can be discovered and compared through a common ontological namespace.

[Librarian @PeculiarLi34455]

The need to have your models running on your own computer locally has never been greater.

[Rafa Schwinger 🇻🇦 @Rafa_Schwinger]

After OpenAI stole the navier stokes work it is obvious that big orgs and universities will need to have their own AI centers running their private IP just like they have their own supercomputers.

Imagine being a big pharma company using ChatGPT rn.

That is urgent.

[Librarian @PeculiarLi34455]

@JyiscS @bubbleboi Not sure how they train on code. All I know is code is a vehicle that wraps and transport the actual knowledge, mostly write in turtle format for easy execution. This is far from procedural code.

[Librarian @PeculiarLi34455]

@JyiscS @bubbleboi Code isn't data, code is merely a vehicle to transport actual data.

[Librarian @PeculiarLi34455]

@millatevin What else could it mean?

[Librarian @PeculiarLi34455]

@JyiscS @bubbleboi Data and knowledge part is entirely a Librarian field of expertise. Computer science sits as a layer above data and knowledge engineering. CS provide the machinery to execute what a library curates. That's why in most institutions these courses are CS and Informatics.

[Librarian @PeculiarLi34455]

@mattpocockuk Knowledge management demands governance as first class citizen. You cannot hallucinate a fact. Anything not supported by verifiable source assertion is treated as noise.

[Librarian @PeculiarLi34455]

Your immediate concern should be, how to organize, verify, and retrieve human knowledge without losing its context.

[Librarian @PeculiarLi34455]

Your immediate concern should be, how to organize, verify, and retrieve human knowledge without losing its context.

[Librarian @PeculiarLi34455]

If you ask an LLM a question, it predicts the most likely next token based on training weights. Its "truth" is an illusion born of interpolation. It can hallucinate a corporate financial figure because, statistically, it sounds correct.

[Librarian @PeculiarLi34455]

At its core, epistemology asks three fundamental questions:

1. What is knowledge?

2. How is it justified?

3. How do we distinguish truth from belief?

[Librarian @PeculiarLi34455]

A Peculiar Librarian says:

-here is an ontological schema,

- a SHACL validation gate,

- and a 1:1 fact contract.

The fact is verified, cross-referenced, and absolute.

[Librarian @PeculiarLi34455]

Machine Learning says:

Here are ten million pictures of a corporate report, based on probabilistic weights, I guess this number might be revenue.

[Librarian @PeculiarLi34455]

A Peculiar Librarian says:

-here is an ontological schema,

- a SHACL validation gate,

- and a 1:1 fact contract.

The fact is verified, cross-referenced, and absolute.

[khalif kairo @KhalifKairo]

People think the hardest part about building tech products is writing code , i might be wrong but i tend to believe the hardest bit is actually engineering/ architecting a system.
Code is a language, engineering is a state of mind ( ideas, philosophy , logic )

[Librarian @PeculiarLi34455]

@zulu_code_tech @Akintola_steve A system is useless if it's data is garbage.

[Akintola Steve @Akintola_steve]

I think if I was to advise a newbie coming into tech now, I’d advise them to go into anything that relates to data.

Data science, data analytics, data engineering, ML engineering, etc.

But specifically, data engineering.

Why?

The truth about the matter is that we are in that phase where most of the applications being built are becoming highly data-centric.

We are generating massive amounts of data from users, transactions, logs, events, IoT devices, AI systems, APIs, you name it.

And somebody has to make sure all that data can actually be collected, cleaned, transformed, moved, stored and made available to the systems that need it.

That is where data engineering comes in.

You start getting into things like ETL/ELT, data warehouses, data lakes, streaming, Kafka, Spark, Airflow, distributed systems, cloud infrastructure and real-time pipelines.

And with AI becoming even more prominent, this is becoming even more important.

Models are only as useful as the data you can reliably get into them.

So if I were starting tech from scratch today, I’d seriously consider building my career around data.

Not because other areas are dying.

Far from it.

But because almost every serious technology company is eventually going to have a data problem.

[Librarian @PeculiarLi34455]

What most call "Machine Learning" is mostly just probabilistic guesswork wrapped in massive marketing budgets. They throw compute at messy data and hope a neural network finds patterns but what they get in return is pure hallucinations.

[Librarian @PeculiarLi34455]

@inf_sanju @KhalifKairo Very.

[Librarian @PeculiarLi34455]

@sattyyouneed We don't have to be ignorant just because somebody else is.