Replit ⠕ Profile picture
Mar 10 10 tweets 5 min read
Did you know you can build a ChatGPT powered personal tutor in just 3 minutes?

Here’s how 🧵
First, fork the GPT-3.5 template on Replit replit.com/@replit/OpenAI…

Then bring over your OpenAI API key to the Secrets Manager as OPEN_API_KEY.

If you click Run, you’ve already got some interaction going right out of the box, but let’s make it more useful!
The new API uses a messages array to deal with nearly everything. You can set up your starting prompt using the new `system` role and give context and instructions to your bot.
Using a prompt borrowed from @emollick we can set GPT-3 up to think of itself as an adaptive tutor, one that will adjust its teaching style and complexity to our level of understanding - that and multiple choice quizzing too! We’ve just got to customise the prompt.
@emollick Let’s set this array up, use an f-string to replace the learning topic with something user generated. Getting the response from the AI is as simple as sending a one-line command to the API. Initial code to set up the prompt and send it over to OpenAI
@emollick When we get back a message from the AI, we want to assign that to the `assistant` role and add it to the messages array. This means that we have a history of the conversation because…
@emollick Each time we go out to the API, we have to send it the full messages array, so it has all the instructions and the original context. If you don’t maintain your messages array no one else will! It’s not chat if it doesn’t remember what the user said.
@emollick The next step is to ask the user for their response and package this all back up for the API.

Nice and simple, we just need to append the dictionary containing the content back to the messages array before the end of the loop.
@emollick Hit run on your Repl and - boom - you’ve built a universal tutor bot, that can teach anyone about anything and adjust its language accordingly.

Can you imagine how much work this would have been six months ago? Now it’s a 3-minute job!
@emollick Want to try it yourself?

Just use this template and bring your own OpenAI API key.

replit.com/@replit/OpenAI…

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Replit ⠕

Replit ⠕ Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @Replit

Mar 10
1/ We launched Bounties 4 months ago. We thought the main use case would be normal coding projects such as integrations.

We were wrong!

Here are some of the most interesting Bounties posted over just the past month. ⬇️

2/ 👶🏽 🏊🏽 Build a webapp to send an SMS whenever a new baby swim class becomes available.

replit.com/bounties/@jisp…
3/ 🏡Build a webapp to calculate home affordability options based on user inputs like income and down payment.

replit.com/bounties/@Robe…
Read 7 tweets
Mar 8
In honor of #InternationalWomensDay, we’d like to shine a spotlight on some of the amazing women building on Replit.

From teen coders to professional developers, Replit is for everyone!

#IWD2023
VulcanWM is a top Replit creator and an accomplished Bounty Hunter.

Her mini games, like Escape the Punch, are addicting, and her portfolio demonstrates her amazing full-stack web development skills.

Check out her profile and give her a follow!
replit.com/@VulcanWM
.@alanaagoyal is a prominent Silicon Valley investor, and she’s also the creator of one of the most popular themes on Replit, used by over 26k people.

Install Alana’s theme, and try out her LangChain Quickstart Template to get started with AI!

replit.com/@alanagoyal Image
Read 5 tweets
Feb 16
Today we celebrate 20M developers on Replit.

And we are showing no signs of slowing down.

But most people don't know how we got here.

Here's our story.
Replit started as a side project to make it easy to run code in the browser.

It quickly grew and was used by startups like @Codeacademy to power their infrastructure.

It soon became clear that running code in the browser was a much bigger idea.
We wanted to make it easy for developers to create, collaborate, and share.

On one platform that just works.

To go from zero to hosting in 5 seconds or less.

Where you just have to press Run.

How did we get there?
Read 12 tweets
Feb 15
Announcing Ghostwriter Chat.

The first conversational AI built right into your IDE, complete with a proactive Debugger.

Generate, debug, refactor, and understand code faster than ever.

Available today in Beta.
Last fall, we announced Ghostwriter.

With Complete, Transform, Generate, and Explain Code, Ghostwriter is the AI code assistant for every developer.

But Ghostwriter Chat takes it to the next level The evolution of Ghostwriter
Debugging taking you out of flow?

Stop wasting time searching obscure error messages or finding that missing semicolon.

Ghostwriter Chat automatically detects errors in your code and suggests fixes in real-time.
Read 8 tweets
Feb 13
Last week, we kicked off our very first Machine Learning hackathon with @weights_biases. Here are some of the coolest projects we’ve seen so far:
Without any previous ML experience, @coding398 used 40,000 images to train a Convolutional Neural Network on detecting specific filters in new images.

You can upload your own image and put the model to the test here: replit.com/@codingMASTER3…
Generate synthetic kōans on-demand with @vertinski’s Zen Master GPT project. This fine-tuned GPT-2 model creates a new Zen kōan with every click.

Get your Zen on here: replit.com/@vertinski/Zen…
Read 4 tweets
Feb 11
What if GPT had internal context on your business?

What could you ask and learn?

@altcap and @paulinebhyang teamed up with Bounties to build a demo and find out.

Let's break it down below 🧵
Altimeter had an idea:

Build an AI chatbot that could use context from Wikipedia or enterprise data to answer internal business queries.

They posted a Bounty to find a builder to pull it off.

@thejessezhang was selected and got to work. Image
Jesse incorporated @LangChainAI and @gpt_index for this project.

These open source tools are doubling down on creating ways to connect external data to LLMs.

Jesse's other project, Llama Hub, is becoming a central library to share different data loaders.
Read 7 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

:(