BlonDev

software | data | content | ai

[BlonDev @blonTweets]

I unlocked $38 Rush tickets for Babe on @TodayTix! https://t.co/o6ijtBiaDm

[Omar Khattab @lateinteraction]

Can we teach small, local LMs to *use* large, remote LMs as *tools*? Papillon with @Sylvia_Sparkle & team shows that this can be very consequential for privacy.

Using DSPy optimizers, we can teach Llama3-8B to reach 86% of frontier LLMs' quality while hiding your private data!

[BlonDev @blonTweets]

I unlocked $49 Rush tickets for A Wonderful World: The Louis Armstrong Musical on Broadway on @TodayTix! https://t.co/e7jqi2jeoX

[BlonDev @blonTweets]

I unlocked $45 Rush tickets for Sunset Boulevard on Broadway on @TodayTix! https://t.co/prLVEA7U69

[BlonDev @blonTweets]

I unlocked $49 Rush tickets for A Wonderful World: The Louis Armstrong Musical on Broadway on @TodayTix! https://t.co/e7jqi2jeoX

[BlonDev @blonTweets]

I unlocked $45 Rush tickets for Sunset Boulevard on Broadway on @TodayTix! https://t.co/prLVEA8rVH

[BlonDev @blonTweets]

I unlocked $45 Rush tickets for Illinoise on Broadway on @TodayTix! https://t.co/wdyN4EqR7d

[BlonDev @blonTweets]

I unlocked $45 Rush tickets for Illinoise on Broadway on @TodayTix! https://t.co/wdyN4EroWL

[BlonDev @blonTweets]

I unlocked $50 Rush tickets for I Ought to be in Pictures on @TodayTix! https://t.co/T1np9eYjLe

[BlonDev @blonTweets]

I entered the Here We Are Lottery for $25 tickets on @TodayTix! Tap here to enter: https://t.co/5LbIonFecQ

[BlonDev @blonTweets]

I unlocked $40 Rush tickets for Here We Are on @TodayTix! https://t.co/NtutzxqwPo

[BlonDev @blonTweets]

I unlocked $30 Rush tickets for The Night of the Iguana on @TodayTix! https://t.co/cGsm6SEsQA

[BlonDev @blonTweets]

I entered the Hell's Kitchen Lottery for $40 tickets on @TodayTix! Tap here to enter: https://t.co/fqV5ylfYAK

[BlonDev @blonTweets]

I unlocked $33 Rush tickets for Walk On Through: Confessions of a Museum Novice on @TodayTix! https://t.co/EMY9jRHERX

[Nick St. Pierre @nickfloats]

I'm shook @tldraw

[OpenAI @OpenAI]

OpenAI announces leadership transition https://t.co/fFYDLwGXQz

[BlonDev @blonTweets]

https://t.co/iOl55ZXjR3

[Joanne Jang @joannejang]

🌱 excited to share small updates to ChatGPT (while we also work on bigger things)!

Here are some B-tier visuals for each new thing.

1. Prompt examples (let us know if you have suggestions on what should go here!)

[Nelson Djalo | Amigoscode @AmigosCode]

Awesome Illustration for Software Architecture Styles

🚨 Join Waiting List for New 10+ Hour Linux Course - https://t.co/m8gcGjW5vP

Image Creds: @bytebytego

#systemdesign #softwareengineering #coding

[Santiago @svpino]

StackOverflow should deploy its own AI model to automatically answer questions.

The community will still vote for the best questions. People coming to the site will get the best possible answers (regardless of whether a human or an AI model produced it.)

(The voting system will be fantastic feedback to improve the model!)

They shouldn't stop there.

They can have another model that improves human answers by including missing context, fixing issues, and removing toxicity.

The last part is probably the most crucial idea of this post: StackOverflow's problem is not AI; it's how toxic it is, especially for newcomers!

All of this is doable.

But they went the opposite way and banned AI-generated answers.

That was a mistake.

The ball is in their court.