Marta 🥷

🥷 Data Science
🧠 Big Data
🤖 Machine Learning
| Data unicorn in progress 🦄 |

[Marta 🥷 @DatosNinja]

Are you thinking of building a career in Artificial Intelligence?

This is what you should learn according to @AndrewYNg himself

A thread 🧶 [↓]

[Marta 🥷 @DatosNinja]

👀

[Zach Wilson @EcZachly]

I'm launching a free data engineering boot camp on YouTube on 11/15!

It will be a new data engineering video every day from 11/15 until 12/31! I'm excited to share my hard work with you in a way that is accessible!

The launch video is here:
https://t.co/djB9bmxNDR

Please share this with your friends! I believe in each and every one of you!

[Zach Wilson @EcZachly]

There are so many different roles in the data domain. Deciding which role is best for you depends on your preferences.

One dimension you can look at these roles is what percent is spent building infrastructure versus digging into the data to find issues.

I personally love #dataengineering because I’m a builder at heart but love diagnosing things with SQL.

I’ve seen myself gravitating both directions towards #softwareengineering and #analyticsengineering depending on what I’m currently frustrated with.

Do you like building stuff? Do you like finding root causes of business issues? This core differentiator between these roles might help you understand where in the data value you chain you want to sit!

[Marta 🥷 @DatosNinja]

Gracias Rafa 🎾

[HeyGen @HeyGen]

We're big tennis fans here. For all the Rafa fans, here's Rafael Nadal's heartfelt retirement speech translated into English.

What a legend, @RafaelNadal 🎾🏆

[Marta 🥷 @DatosNinja]

The original gig is in Spanish:
𝗕𝗶𝗴 𝗗𝗮𝘁𝗮 𝗲 𝗜𝗻𝘁𝗲𝗹𝗶𝗴𝗲𝗻𝗰𝗶𝗮 𝗔𝗿𝘁𝗶𝗳𝗶𝗰𝗶𝗮𝗹
https://t.co/S3wimCPl5y

[Marta 🥷 @DatosNinja]

I've just lost my job to NotebookLM.

This is yesterday's episode with new English-speaking hosts on decision trees 🌳

[Marta 🥷 @DatosNinja]

@svpino Now that you all are talking about hardware... any recommendations on any laptop for ML, DL and the like 😇

[Marta 🥷 @DatosNinja]

🥇 @omarsar0 posts are always gold

every.single.time

[elvis @omarsar0]

Crawl4AI is an open-source web-crawling and data-extraction tool built to integrate well with LLM and AI applications.

It has a comprehensive list of features ranging from simultaneous URL crawling to advanced extraction strategies based on LLMs.

[Marta 🥷 @DatosNinja]

And most importantly, choose the model based on what you need.
👉 o1 is perfect for complex algorithms, coding, and planning.
👉 GPT-4 is better for quick answers and working with text and images.

[Marta 🥷 @DatosNinja]

⚠️ 𝗜𝗳 𝘆𝗼𝘂'𝗿𝗲 𝘂𝘀𝗶𝗻𝗴 𝗼𝟭 𝗹𝗶𝗸𝗲 𝗚𝗣𝗧-𝟰, you're wasting its potential.

Let me tell you how to fix that...

[Marta 🥷 @DatosNinja]

𝗞𝗲𝗲𝗽 𝗶𝗻 𝗺𝗶𝗻𝗱:
✅ Simplify. Short, direct prompts work best.
✅ Don't use "chain-of-thought". o1 already does the reasoning without being told.
✅ Limit context in RAG. Less is more.

[Marta 🥷 @DatosNinja]

“All models are wrong, but some are useful”

George Box, 1976

[Marta 🥷 @DatosNinja]

🔍 If you have complex challenges that require more analysis, try the o1 model and let me know how it goes.

And if you have more info about o1... drop the post in the comments, and we can make a thematic megathread 🍓🍓

[Marta 🥷 @DatosNinja]

What makes OpenAI's o1 model family special? 🍓

A thread 👇

[Marta 🥷 @DatosNinja]

⚠️ But beware, o1 isn't for everything.

If you're looking to generate or edit text, you won't notice much difference compared to GPT-4o. However, if you're dealing with mathematical or analytical problems… that's where o1 really shines. 🌟

[Andrej Karpathy @karpathy]

Haha we've all been there. I stumbled by this tweet earlier today and tried to write a little utility that auto-generates git commit message based on the git diff of staged changes. Gist:
https://t.co/1SbQsHSNwK

So just typing `gcm` (short for git commit -m) auto-generates a one-line commit message, lets you to accept, edit, regenerate or cancel. Might be fun to experiment with.

Uses the excellent `llm` CLI util from @simonw
https://t.co/LnHeCSfiHc

[Marta 🥷 @DatosNinja]

🍿🍿

[Santiago @svpino]

This is embarrassing. 🍿