ATL #02 Does your AI gulp WATER? 🌊

Plus: Amazon's FREE CodeWhisperer to aid 25m+ coders

Reading time 5min; Characters 6041; Words 1024; Sentences 77

Hi there Thought Leaders aka humans.

First of all, an AI 🤖 salute to all of you for supporting and subscribing this newsletter. Humans like you make this AI world real. A big THANK YOU!

Let’s see what’s whispering in this week’s AI world:

• Your AI secretly gulps lot of water 🌊

• Amazon CodeWhisperer to help you code ✍️

• Quick AI BYTES for snacking 🍟

• Learning of the week: WEB3 🧑‍💻

• Meme Corner 😂

🌊Your AI secretly gulps lot of water

A new study by universities (University of Colorado Riverside and University of Texas) in the United States looked at how much water is used to train big computer programs like OpenAI's ChatGPT.

Results? 🧐

Not so good. Studies showed the AI models are super “thirsty”. Yes you heard it right. ChatGPT alone can ‘drink’ 500ml of water to answer 20-50 questions! Thats not large you say? Imagine the millions of active users asking billions of questions and the amount of ‘water’ required for the same + tons of other AI engines from other tech companies. Mind blowing amount of water being used out there. All gone with AI!

What’s going on? 🤯

okay, okay.. Not drinking water literally. The water is used to cool the data centres & generating electricity, which is essential to keep them running. It often uses fresh water to prevent corrosion and bacterial development.

Give me more context please 🤔

According to the research paper, Microsoft used a lot of water to cool its data centers in the US when they were training GPT-3. They used so much water that they could have made 370 BMW cars or 320 Tesla electric cars. If they were training the program in their even bigger data centers in Asia, the number would be three times higher! That's a lot of cars, enough to start a whole new taxi service for Microsoft🚕🚕🚕🚕🚕

What next? 😑

The paper raises a lot of questions like - where does the water go? Does it disappear into thin air? Well, the water comes from rivers, lakes, and other freshwater sources. And there's a difference between "withdrawal" (physically taking water from a source) and "consumption" (when the water is lost through evaporation in the data center).

The study is around consumption, which is where most of the water use comes from. But on the brighter side, the water isn't gone forever. Instead, it gets released into the air through cooling towers, although it takes some time to return as rain.

Research puts focus on collective efforts from all tech giants to combat global water challenges and make their data centers more efficient and innovation

Bonus for the research Geeks 👻

You can read the entire research paper here Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Model

✍️ Amazon CodeWhisperer to help you code

A good news for 25m+ growing software coder community in the world! Amazon makes CodeWhisperer FREE.

Wait a min, do you know what is CodeWhisperer? Let me start from the basics.

CodeWhisperer is an AI coding companion that generates real-time, single-line or full-function code suggestions in your Integrated Development Environment (IDE) to help you quickly build software.

Amazon launched this AI tool in June 2022 however from 13 Apr 2023, now this is available for free for individual users.

How to access?

It’s as simple as sign up with an AWS Builder ID based on your email address. Signup link is HERE.

Amazon also stated, they have a paid version of the whisperer for professionals at $19/month with enterprise administration capabilities.

What does it mean for coders?

Amazon says as per their productivity tests, developers who used CodeWhisperer completed tasks 57% faster.

Thus more efficient and more productivity from a coder. Pleasing the ears of many $ hungry organisations 🤑🤑🤑

Thoughts?

Will you use it? or have you already used it? Do let me know your thoughts in comments.

🍟 Quick AI BYTES for snacking

• Elon Musk launches OpenAI rival named X.AI

• Alibaba unveils ChatGPT rival and custom LLMs

• Google unveils never-before-seen text-to-video AI tool

• Adept AI raises $350m, trained to use every software tool and API

• AI to spot Alzheimer's Early, Using Folks' Speech Patterns

🧑‍💻 Learning of the week: WEB3

WEB3 explanation in a simple language

Before Web3, lets find out Web2.

Web2 is today’s internet the so called www.

Web2 is like a library where you can read books and find information. But you have to ask the librarian for help and follow their rules.

Web3 is like a playground where you can play games, buy and sell things, and make friends. Everyone gets to decide the rules together and no one person is in charge. It's like having a big group of friends who all agree to play by the same rules. Web3 is more fun and fair because everyone gets to participate and have a say.

Today, web3 exists in more technical manner in the form of cryptocurrency, NFT, blockchain etc. Let’s go through these terminologies in the upcoming weeks.

😂 Meme Corner

AI taking over the jobs is a hot topic these days amount humans.

I am sure you might have discussed & debated about this among your circles.

One such discussion or rather a short one among two friends caught AI’s eyes 😂

Last week’s inaugural episode was loved by many of you and shared some valuable feedback. Thanks a ton!

Ben (Singapore) “Lovely newsletter on AI, just the one I need for the moment”

Rajoli (Germany) “The way you put things together is really great”

Mayur (Norway) “subscribed to the newsletter! It looks great 👌”

Kit (Hong Kong) “This will add value to a lot of people, crisp & clear newsletter”

Nadya (Indonesia) “This initiative sounds really interesting and exciting, glad to be a subscriber to get regular updates”

Abenav (India) “It looks interesting. I would recommend to include content around AI ethics and governance too”

That's all we have for this week! Tell us how you feel either by hitting a reply or comment.

We love feedbacks 🫶

See you next week 23 Apr at 14:00 HKT / 06:00 GMT