Bots That Overshare, Suck Up Less and Sync Mouths Like Pros

Meta launches a new Facebook-flavored chatbot, ChatGPT dials down the flattery and Tavus animates mouths better than most humans

May 05, 2025

Fellow Data Tinkerers!

Before we get to this week’s round up of all things data and AI, just wanted to remind you that you can unlock the following prizes by referring Data Tinkerer to others:

1 share - a PDF of top 15 articles on Data Tinkerer

2 shares - 100+ cheat sheets for all things data. From Python, SQL, Power BI to Tableau, Spark, Git and many more

4 shares - 100+ resources to learn all things data (science, engineering, analysis). It includes videos, courses, projects and can be filtered by skill or whether it’s free or paid

So if you enjoy reading Data Tinkerer, share with friends and earn referral rewards!

Refer a friend

Now, with that out of the way, let’s get to this week’s news round up on all things data and AI

The Buzz 🐝

Meta had its first ever LlamaCon last week and introduced products aiming to compete with OpenAI and others. Llama API, geared towards developers to fine-tune, deploy and evaluate models from the Llama series. And the stand alone Meta AI app, which has ChatGPT-like capabilities with a social twist. It has a ‘Discovery’ feed similar to Facebook where users can share and explore AI-generated content.

The early feedback has been mixed with some people mentioning LlamaCon missed the mark in comparison with competitors and others who tried the app likening it to an “oversharing ChatGPT”

OpenAI had two major updates as well. They rolled out a “Shopping” feature where it helps users shop by showing product visuals, prices, reviews, and direct links to the products without any ads

The second update was the rollback of the “sycophantic” ChatGPT-4o update that had even annoyed Sam Altman. They have provided more information about what went wrong and you can read it here if interested (tl:dr: they focused too much on the short-term feedback for the model)

Tavus has unveiled Hummingbird-0, a zero-shot lip sync model that delivers photorealistic, natural mouth movements from just an audio clip and a short reference video without any training or fine-tuning required. For The White Lotus fans, they even released a banger demo of Frank talking about fine-tuning AI models. You can check it below:

Data Science & AI

Using LLM to transcribe restaurant menu photos
Learn how DoorDash combines OCR and LLMs to auto-transcribe menus into clean, searchable formats.

How Walmart Automated 400+ Forecasts and Cut Runtime by Half
No model works for every SKU. Walmart's solution? An autotuning framework with rolling validation, automated feature engineering, and real-time model selection. If you want to learn more about it, check the article!
How Airbnb used Deep Learning to rank listings
Jose Parreño Garcia
dives into a paper published by Airbnb and goes through how they applied deep learning to rank listings

Data Engineering

How Canva Rebuilt Its Data Pipelines for Billions of Events per Month
Canva had to track billions of events to pay creators fairly and their old system couldn’t keep up. Curious how they rebuilt it? This article is for you

Building Pipelines for Automated Reconciliation with Near Zero Data Loss
Learn how Walmart built an automated reconciliation system that captures and compares expected vs. actual transactions across distributed different systems, ensuring near-zero data loss and timely seller payments.

Why You're Stuck at Senior Data Engineer (And How to Break Out)
SeattleDataGuy
shares why many data engineers plateau at the senior level (Hint: it’s not only because of the technical skills)

Data Analysis and Visualisation

The Dashboard Fallacy: When Seeing Everything Means Understanding Nothing
Dashboards look smart. Most don’t help you think. This post breaks down the dashboard trap and why more data ≠ better decisions.

How much plastic waste actually gets recycled globally?
An interesting visualisation by Nick Routley

What happens if you ask ChatGPT to create the exact replica of an image 100 times?

If you enjoyed reading this week’s round up, please give this post a like and share it with others. If you have any other feedback, feel free to leave a comment below or reply back to me. I read all of the replies and really appreciate them :)

Data Tinkerer