New Tool dropping in Mito: Scratchpad.
Problem:
- Agent needs to figure something out (explore filesystem, analyze data to create a mapping, etc.)
- Previously writing code in the notebook was the only way for the agent to learn more.
- Result: The agent leaves the notebook messy. This code serves no long term purpose.
Solution:
- Use the scratchpad tool! The new flow is: Scratchpad for exploration (hidden) → get additional context → hardcode results in the notebook
The scratchpad is fundamentally a tool for producing a cleaner notebook by doing exploration work invisibly.
Just shipped `mito-theme-light` a jupyter theme built for AI-assisted data science:
- Turns on line numbers + adds cell numbers so you can actually reference your code when talking to an AI
- Consolidates actions into one dropdown so new users discover things like "run all cells"
- Turns your dataframes into interactive tables
- Removes buttons nobody uses (when's the last time you clicked "go to line number"?)
🚨 X AI just made up that the @okcthunder beat the @HoustonRockets -- the game is only at halftime and the Rockets are ahead.
It also made up the box score for @KDTrey5 and @shaiglalex. KD has 14 points at halftime.
Continued hallucinated summaries could be trouble for bettors.
@X needs to improve it's live events coverage. These AI overviews are sometimes just wrong. The game isn't even over, and the @HoustonRockets are actually ahead at the time of this tweet.
This could actually be impactful for sports betting!
@NBA
By 2030, the data science field is projected to grow by 30% -- this is the tool I’d add to my learning path if I were re-skilling into data science:
𝐩𝐢𝐩 𝐢𝐧𝐬𝐭𝐚𝐥𝐥 𝐦𝐢𝐭𝐨-𝐚𝐢
It’s an open-source toolkit that adds AI to your Jupyter Notebooks, and it’s surprisingly helpful for learning by doing. Try now https://t.co/fzqyHvgQ5Y
It lets you:
• Use natural language to generate full data workflows
• Upload an Excel file and watch it convert to clean Python code
• Work with top models (ChatGPT, Claude, Gemini) right inside your notebook with no setup
One thing I really like about Mito AI: it shows you exactly what it’s doing.
(Which is super helpful for both beginners and debugging.)
-The agent explains its assumptions
- Every result links back to editable, transparent code
That makes it a great companion if you're:
• Transitioning from Excel
• Exploring data automation or dashboards
• Or just brushing up on how real data science pipelines work
Mito AI now lets you build custom visualizations inspired by any image 📊
Example: We help a lot of large banks use Python and AI.
A user can style their visualizations to be Citi colors just by uploading an image of the logo.
While I don't believe vibe data science exists yet, our team is getting close 🧙
#AI #Python #DataScience #LLM #Excel #Data
🚀 Upload your files and our agent will convert them to Python code.
This example is an Excel file with formulas and pivot tables.
As we continue to make data science accessible to everyone, we are building functionality that allows users to take their current workflows and bring them to Python.
Download Mito AI here: https://t.co/2gNgNURseb
@preetam_joshi Hey Preetam, we built an AI experience specifically for Notebooks, and see better results than with Cursor -- if you want to try: https://t.co/GmFoClOVdD
Whenever I get a call from an AI voice agent, I'm paranoid that they are recording my voice for training data, so I only talk to the agent in an accent.
Just doing my part to make AGI a little more french.