Prompt engineering best practices, writing poems with AI, newsletter reaches 100 subscribers, AutoGPT, evolutionary tree of LLMs

Also Microsoft's new Athena AI chips, EU's AI Act, and GenAI increasing productivity by 14%

May 02, 2023

Welcome to my Effective AI newsletter, where I share my thoughts about AI, discuss latest digital news, as well as productivity and creativity tips for the AI community. If you missed my previous editions of this newsletter, you can check them out here. Additionally, you can visit my website for more of my articles, tips and tricks.

I also have this newsletter in the podcast form, so if you are interested in listening to it, please click on the player icon below.

🧠 What’s on my mind?

My very first poem for the lecture at Deakin University

I had a pleasure to speak in front of the masters class at Deakin University. I shared my personal views on how data and analytics can help large organisations gain competitive advantage. I shared my 5 pillar framework that I wrote about before.

I wanted to share my very first poem that I wrote in my life, created, of course, with the help of GPT-4. Here’s the prompt that I used to produce it.

And the output:

At Deakin, where minds intertwine,
Lubna's invite, my chance to shine,
We gather to learn, to seek, to find,
The Value of Information, knowledge combined.

Advanced analytics, a competitive edge,
In this lecture, a solemn pledge,
To delve, explore, unravel, and bind,
The secrets of data, for growth designed.

With case studies, we shall unfold,
The tales of triumphs, new and old,
Organisations' victories, vast and grand,
A testament to wisdom's guiding hand.

Components we'll study, one by one,
How advanced analytics has begun,
To shape the future, transform, and mold,
A path of success, brave and bold.

In Deakin's halls, where wisdom breeds,
We gather, share, and plant the seeds,
For a brighter tomorrow, built on today,
The Value of Information, leading the way.

This newsletter reached 100 subscribers!

This is the first milestone for my newsletter - I reached 100 subscribers last week, after launching it 2 months ago.

Here’s what I’ve learned during the last 2 months:

Events drive the engagement. When I look at my subscribers graph, the two events that I ran, generated big bumps of new subscribers. Another source of subscribers, although smaller but more consistent, is publishing to my LinkedIn network, where I have 1,500 followers.
Organic discovery is slow. The goal for me was not to grow it as quick as possible, but rather to learn the skills required to run the newsletter and a promote it and just to have fun sharing my knowledge. I also realised how important the discovery is and how slow is organic growth, without social media.
Even at small scale, publishing online can drive a lot of networking opportunities, sourcing great ideas and finding leads. I’ve never had so many people, who I know and strangers, reaching out with ideas, leads and interesting connections.
Social media consumes a lot of time, so prioritising it is vital when you don’t have a team to rely on.
Ask for feedback proactively. As with any product or service it is almost impossible to get it right first time. On the other hand, over-planning is also counterproductive. So, from the beginning I decided to prioritise action over planning, just give it a crack and just fix based on user feedback.
I learned heaps by just writing a newsletter every week
1. Writing with the help of GPT-4 is much easier but a bit of an art. You do need to provide your original ideas first.
2. Optimising my time to write weekly edition of newsletter based on my learning routine is important
3. Refreshed my skills on running and optimising a website in the cloud, as well as automating my workflow through scripts and storing content in git
4. SEO optimisation is an art
5. Audio editing skills and working with professional microphones is an entire new world for me
6. Voice training. By hearing your own voice gives a lot of room for improvement. I also discovered that voice training is a thing not only for actors but also for business professionals.
7. Understanding the social media ecosystem that I have never tapped into in the past

Next milestone is 500 subscribers!

🛠️ Practical tip - prompt engineering best practices

I’ve been playing with GPT-4 over the last few weeks and there are a few prompt engineering best practices that I developed through my exploration. For an extended version, visit article on my website.

Here are my 5 best practices for prompt engineering:
1. Write instructions clearly, as you would do for an intern
2. Add structure to your prompt
3. Develop prompts iteratively
4. Reduce the risk of prompt injection by using prompt delimiters
5. Never trust the model

🌏 What’s happening in the world of AI?

AutoGPT - GPT-4 based agent

AutoGPT is based on GPT-4, it runs recursively as an agent, search the internet in consistent way, and launch new agents. It also claims to have long-term and short term memory and file storage

Stanford and MIT: LLMs increase productivity by 14%

A study conducted by Stanford University and MIT researchers revealed that customer service employees at a Fortune 500 software companies experienced a 14% average productivity increase when provided with generative AI tools. The least-skilled workers gained the most advantages from these tools. This is the first time the impact of generative AI on workplace productivity has been measured outside a laboratory setting.

Microsoft is building its own AI chip based on 5nm process

Microsoft is joining the race with Meta and Google building their custom AI chip, code name Athena. This chip is specifically designed for LLMs. This may allow Microsoft to be even more competitive, as they won’t need to buy chips from third parties with markup.

EU passes a draft of AI act, which covers copyright rules for GenAI

Similar to GDPR, EU is leading the effort to pass an AI act that will cover laws of AI regulation.

One part of the act, covers rules that will require companies to make GenAI more transparent and disclose any copyrighted material used for training. The act also applies to countries outside of the EU that provide services to EU countries.

Harnessing the Power of LLMs in Practice: A Survey on ChatGPT and Beyond

An interesting paper from Jingfeng Yang et al. that covers LLMs from perspectives of models, data and downstream tasks. It also has a very nice graphic of evolutionary tree of LLMs