The Generative AI newsletter for legal pros everywhere.

🚀 Hello! We have reached 715+ subscribers in 23 days! I cannot thank you enough for subscribing!! 🎉🔥

NEW SUBSCRIBERS, to read previous posts, go here.

My goal is to reach 1000 subscribers. Do you think I can make it? With your help, I can, so please share this newsletter with as many people as you can if you find it useful.

This is day 24 of 100 (our goal - 100 consecutive days!) and today we will:

1. ask if HAL is already here?

2. give you a crazy simple way to create educational programming

3. share 3 tools

4. review what ChatGPT is (via Y Combinator)

5. thank an awesome subscriber for referring folks

6. talk news you can use and lose

Here’s the thing . . .

Is OpenAI (the architect of ChatGPT) slow-walking the release of highly tuned and capable AI models? 

Are they closer to Artificial General Intelligence (AGI - you know, a fully ‘humanized’ computer) than they are letting on? Some think so.

This article lays it out with some compelling thoughts.

Here is a summary:

• OpenAI was initially met with skepticism but gained attention with the release of GPT-3 in 2020 and ChatGPT in 2022.

• The GPT-4 model seems to follow a "slow take-off" scenario, where AI capabilities increase gradually, allowing society to adapt.

• The GPT-4 technical report suggests that OpenAI selected performance numbers to limit disruption, rather than pushing the technology to its limits.

• OpenAI may be intentionally slowing down AI development to reduce social disruption and chaos while calling for democratic oversight.

• Evidence supporting this idea includes OpenAI's focus on predictable performance scaling and not releasing the parameter count for GPT-4.

• It is unclear how fast AI development could take off, and OpenAI might not be able to prevent sudden capability jumps from other companies.

Use Case: Educate like a Pro

I love it when these newsletters write themselves! No, I don’t mean using ChatGPT. I mean the idea and use case just resonate so deeply with me.

I know many of you reading this are knowledge workers. This means we deal with the intangible often but we have to communicate it to others. We give talks at events, prepare decks, put on a presentation, or otherwise have to engage an audience to educate them on something.

Over the years, I have been asked by many people how to do this effectively. It is always a challenge to describe because it typically involves a lot of work that most of us don’t have time for. 

You know your stuff, but you are unsure how to teach it!

Let’s solve that right here and now!

We are going to use ChatGPT to:

1. introduce a fundamental teaching construct

2. use that construct to help you focus on what and how to teach

3. structure your knowledge in a manner that makes it easy for you

4. create a complete annotated outline

5. develop simple assessment tactics

6. then turn that into a scripted outlined

7. laugh out loud with how awesome this is!

▶︎▶︎Prompt

🚨 Before you begin you should find some halfway-decent content you can use to feed ChatGPT. I used a prior edition of Brainyact (Edition #__) where I walked you through how to do market analysis and come up with a practice group strategy. I literally just copy and pasted that.

Do you have some notes, a blog post, or a whitepaper you have written before? It does not need to be complete or perfect.

I will include a link to the full ChatGPT session for all of the following so you have these prompts (compete with my typos 😂) and the full replies.

Step 1

Prime ChatGPT to get it in the right frame of mind to be highly effective for you.

You will also learn something about teaching from this prompt reply, so read it, please.

▶︎▶︎PROMPT

Please explain how to use Bloom's taxonomy in developing me to create educational content.

You now have learned one of the fundamental models to teach like a pro.

But let’s dig a bit deeper.

▶︎▶︎PROMPT

What are the components of the taxonomy?

Ok now you have a keen sense of what we are going to use going forward (and ChatGPT does too - which is critical)

Step 2

Now we are going to turn to the topic you are going to be teaching or talking about.

▶︎▶︎PROMPT

Great, thank you. I am going to be giving a presentation to [knowledge workers in the legal market]. I will be explaining [what generative ai is, how ChatGPT works, a few use cases, and how to use prompts to generate useful replies.] Can you provide specific examples for each of Bloom's levels?

Now you have each of Bloom’s levels broken out with specific learning objectives. Depending on your time, audience, and your level or expertise, you can now choose which learning outcomes are reasonable for you.

Step 3

But don’t stop with learning objectives. Any pro educator knows that you must evaluate or assess whether your audience “got it” or not. Even in a group setting or when you are on a stage, you can do some of this.

So let’s generate some assessment ideas. Choose one of Bloom’s levels to focus on. You can choose more than one for your teaching but run one prompt for each level. Also, don’t overstuff your talk with too many learning objectives!

▶︎▶︎PROMPT

In a presentation format, how might I be able to assess if these learning objectives have been achieved for each of the levels?

There you go. You have assessment ideas!

Step 4

Now we are going to turn to the content you have previously created. We want ChatGPT to craft clear learning objectives for it and a few assessment ideas. I am focusing on Bloom Level 2 for this (you will see it in my prompt).

▶︎▶︎PROMPT

A few weeks ago I wrote some content about using ChatGPT to create market analysis reports for lawyers. I want to use that as the topic for an upcoming lunch-n-learn. Using Level 2 of Bloom's, can you review the following content and create 1) clear learning objectives and 2) a few assessments? Here is the content: [paste your content]

In reviewing the reply, it suggested I use multiple-choice questions. I could use this as a mid-talk survey I send out to attendees (or use a QR code if I am on stage). I want to create 10 reasonable questions.

▶︎▶︎PROMPT

Elaborate on Number 2 "Multiple choice questions" - can you draft 10 multiple choice questions that are appropriate for this material?

Step 5

You psyched yet? But wait, there’s more!

Now let’s generate a complete outline and script to deliver a 45-minute talk.

Are you ready for some serious s#@%?

▶︎▶︎PROMPT

Excellent, we are almost done. Now I would like you to draft a thorough and comprehensive annotated outline for this 45 minutes session on this topic. I would like the outline to include the main topics, talking points for each topic, and an example script that I can use to deliver the whole program.

💣💥🤯 BOOM!!

There you go. You have a fully built program to start practicing, refining, and editing. You will look like a pro by using learning objectives and assessments.

🔗 Here is the link to my complete session.

AI Tools for Work

Trevorai: Plan your day, one task at a time. Access Deep Work. Basically gives you an easy drag-n-drop way to interact with your calendars so you can visually plan your day.

Poised: This looks really helpful for our distributed lives and using online video to communicate. It basically actively coaches you to become a better presenter and speaker. No more low energy, “ums”, or interrupting others.

Personal.ai: Supercharge your memory and relationships with your own, personal AI messenger. This looks interesting on the personal front. I have so many message apps, inboxes, data (music playlists, social threads, etc.) I can't possibly recall where I “saw” certain things, let alone remember them. This looks to address that. Got on the early waitlist.

News you can use : 

🚨 🚨 🇮🇹 ITALY BLOCKS CHATGPT. This is interesting and likely more GDPR jurisdictions will clamp down temporarily .

News you can lose: 

I just call this hype cycle phenomenon or chasing shiny objects. Most companies have zero AI strategy. That is why the likes of the huge consulting giants of Deloitte, Bain, etc., and supposedly other “big brain” organizations like Goldman Sachs are trying to capitalize on this - they want the big $$$ from companies to help them craft wickedly smart things to say (but will not help them actually do). Sorry, not sorry. 

Wall of Feedback - Big Kudos Sorcha-Mary Boyce

Sorcha-Mary Boyce is Senior Manager for Legal Technology Engagement at Allen & Overy in Singapore. No doubt she is working on quite interesting things. One of the early experimenters in the generative ai space.

That's a wrap for today. Stay thirsty & see ya next time! If you want more, be sure to follow me on Twitter and LinkedIn.

DISCLAIMER: None of this is legal advice. This newsletter is strictly educational and is not legal advice or a solicitation to buy or sell any assets or to make any legal decisions. Please be careful and do your own research.