Srini

ChatGPT o1-preview failed this relatively simple test, and our Enterprise h2oGPTe system nails it, much faster:

One of my favorite conversations I ever had with LLMs. Not backing down, but getting funnier with each reply. Way better than replying "I cannot answer this" - loving it!

A macOS calculator clone in ~50 lines of #Python. No HTML/CSS/Javascript. github.com/h2oai/nitro/bl… #madewithnitro #100DaysOfCode

youtube.com/watch?v=Aiitmp… Get a sneak preview of the new Wizard for H2O Driverless AI #automl built with @h2o_nitro @h2oai - avoid common pitfalls in #DataScience and #MachineLearning and get value out of #AI!

Join us tomorrow for Make with H2O.ai! The session will cover how to get started with H2O Driverless AI with Chief Technology Officer, @ArnoCandel. ✅ Register: bit.ly/3PDirIm There will be a rerun for APAC with live Q&A on August 25 at 1pm SGT!

The Nitro docs include a cheatsheet now. Despite having only two functions, view() and box(), the cheatsheet has 100+ ways you can use them to produce variations in functionality and behavior - a kind of grammar for interactive user interfaces. nitro.h2o.ai/cheatsheet/

H2O Wave is now a critical project on PyPI (based on downloads?). Thanks to everyone who made it happen: @unusual_code, @MichelleTanco, @aalencarv, @cyberbaggage, @srini27even, Marek, the @h2oai cloud team, and everyone who provided feedback and helped shape Wave.

Here's H2O Drive - the file storage/sharing system for H2O AI Cloud, written by @srini27even. This is authored in 100% Python 🐍 using H2O Wave. No JavaScript, HTML, CSS used! Wave is open-source: github.com/h2oai/wave 100+ widgets. #Python #DataScience #apps #webdev

📢 H2O Nitro 0.11 is out! It's now possible to run Nitro apps entirely in-browser, thanks to @pyodide. You can embed Python code directly in HTML, or dynamically load external packages, wheel files, or modules at run time. nitro.h2o.ai/wasm/ #Python #DataScience

📢Nitro 0.11 will ship with Pyodide support this week! - Run Nitro apps entirely in-browser using Python-in-Wasm. - Embed Python code in HTML using "text/python" script tags. - Load external ".py" modules. - Tiny (~13KB) wheel. #Python #DataScience #Coding

This is neat: @srini27even and @unusual_code made the H2O Cloud Drive app react to the system theme 😎 Seems like it's already possible with stock Wave and wave.emit(), which was a pleasant surprise. #madewithwave

Happy Father’s Day

Following last week's introduction to @h2oai this week we're taking you through a tutorial of how to use H2O Wave to build a simple app to display #data in various forms including plots, tables and graphics using #python 🐍 jumpingrivers.com/blog/displayin…

Dev on Nitro v0.10 is going at a breakneck pace. Shipping this week: - Make partial updates to the UI (ins, swap, del) - w/ low-latency. - Bookmarkable screen flows - Chromeless mode (@projectjupyter, here we come!) - Automatic busy spinner (animated 🚀!) #Python #DataScience

Nitro v0.9 now has a CLI that lets you "nitro run" a URL pointing to a #Python file or gist on Github or anywhere else. Makes it easy to run examples: just copy/paste commands from the docs into your terminal nitro.h2o.ai/cli/ #DataScience #apps @h2oai

One of the advantages of having a really fast programmer try out your new library is that you get bug reports really fast. #ShowerThoughts @ArnoCandel

Rough prototype: "nitro run URL" fetches a .py file from a gist, sets up a local virtual env, downloads additional datasets or files, and finally launches the app. A dockerfile-like spec is required somewhere in the comments. No Docker involved. Just need #Python. 1/4

Wave 0.22 is out! #Rstats web apps, data table enhancements (markdown support, in-row menus), a shiny new image annotator. Most exciting: the tour now lets you live-code in the browser, with autocomplete and code snippets! wave.h2o.ai/blog/release-0… #Python #DataScience @h2oai

A short guide on deploying H2O Wave apps to Heroku using Docker. The dockerfile could be useful for other cloud envs as well. Via Srini (our resident infrastructure/tooling mad scientist). #Python #DataScience @heroku @h2oai @srini27even @srini.x/deploy-your-wave-app-in-5-minutes-bd440d00917f" target="_blank" rel="nofollow noopener">medium.com/@srini.x/deplo…

Make with H2O: Getting Started with Wave for the Wildfire Challenge twitter.com/i/broadcasts/1…