I was asked in an interview about talks or public presentations I've given. Thought it might be useful to publish them here. Most of them are in Spanish, but here goes anyway:
1. Modern Data Stack @thedatapub
An attempt to demistify the word "modern" by going through the cycles technology has had for analytics
Se acuerdan cuando en un debate presidencial del 2018 Anaya le dijo a AMLO: “ el problema no es que seas viejo, el problema es que tus ideas son viejas, el problema no es que no entiendas inglés, el problema es que no entiendes el mundo”
A lo que AMLO contestó “Riqui, riquín, canallín”
Y todo México se cagó de risa.
¿Aún les parece gracioso a quién le dimos el poder?
It is great that the frontier labs want to support data analytics workflows. It is the #1 thing many enterprises want now. But enterprise data is difficult. Turns out most models are bad at it! They’re ok at writing single SQL queries or Python scripts, thanks to the plethora of text to SQL and data science benchmarks, but they struggle to query, clean, and make sense of data from multiple database systems (relational and non relational).
Fortunately, we released a new benchmark to help, the Data Agent Benchmark, with plans to get it into a super well-known benchmark very soon :-) stay tuned!
https://t.co/xZFcz6mUle
“AI slop is really a way of expressing that it’s difficult to identify the intent behind the form.”
This is by far the best definition of slop I’ve see
dude who complains model burns bajillion tokens and then prompts like “yeah uh, we should definitely check on that. tests should pass I suppose, I liked what you did the other time, let’s do that again”
.@ApolloAtomics builds the most compact nuclear reactors with the highest uptime and a deployment time of less than 24 months.
Apollo took the pressurized water reactor technology that already powers 80% of the world’s nuclear plants and flipped one part, the steam generator, to make the plant an order of magnitude smaller without compromising power.
Congrats on the launch, @AssilHalimi & Drew!
https://t.co/5lGDpZhmQ5
We've created the world's fastest PDF parser ⚡️
And it's more accurate than any other open-source, model-free PDF parser out there (pymupdf, pypdf, markitdown, pdftotext, opendataloader, pymupdf4llm)
Introducing LiteParse v2 - we rewrote the entire library into Rust and adapted it as native packages for Python and Node.
It supports 50+ different document types, can be triggered directly or installable directly within your favorite AI agent.
Blog: https://t.co/ckb0G73ESs
Repo: https://t.co/JNER0mVcB8
introducing howtoeval dot com. the no-bullshit guide to eval'ing AI agents.
from personal experience, and from working with the best companies in the world.
there's even a quiz. link below.
I've been a software engineer for years but I've never had to interview for a job before. I tried leetcode once out of curiosity and couldn't do it. Seemed really disconnected from what software engineers do.
For a decade, “streaming on Spark” meant micro-batches. Fine for ETL. A wall if your latency budget was under a second.
Spark 4.1 stops that. Real-Time Mode (SPARK-50708) 👇
Made a reader edition of Magnifica Humanitas, Pope Leo XIV's encyclical on AI bc the Vatican site is hard to nav.
It includes Tufte-style marginalia for the Vatican's footnotes + 78 Claude editorial annotations w/ Wikipedia links, and an optional silly little attention overlay.
rivers of ink will be spilled about the Luce design
but my biggest gripe with it is that it's clearly made by people with no rage in their heart, and no sex in their veins
ferrari is about love and anger in a way that wholesome california design bois will never understand
“A custom field on a user management page, a new permissions tier, a different webhook payload, an alternative way of grouping records. None of these are roadmap items anymore. They are afternoon tasks.”