Welcome to Issue #2 of Born Analog//Living AI. Thank you to those new subscribers who have joined for the journey in the last week. As well as the news and other sections, this week I’ll introduce you to my very first AI persona, Gordon Blanc, Chef Extraodinaire! Let’s get cracking…

News:

• Twitter becomes ‘X’: Elon Musk has changed the face of Twitter since his purchase in October 2022. Now he’s gone the whole hog and changed the name to X. Why? X is an AI play, it’s Elon’s desire to make this the everything app, where communication meets entertainment meets finance meets ecommerce - all with AI at the core. This could be huge. Link for more

• Open AI CEO launches WorldCoin: Open AI CEO Sam Altman, has launched a cryptocurrency and is using it to incentivise people to give scans of their eyes via a ‘silver orb’ in return for some of the coins. The incentive is there to be able to confirm the distinction between human and robot apparently. As we have concerns over deep fakes and fake news, this seems like a timely idea to me. Link for more

• Meta launches Llama 2 in partnership with Microsoft: In a move that must have all at OpenAI (who have taken $10BN from Microsoft) in shock, Meta - formerly Facebook - and MIcrosoft have introduced Llama 2, an open source large language model said to be comparable with OpenAI’s ChatGPT, free for research and commercial use. Click here for more

• Eleven Labs releases advanced AI voice tools: the world of AI moves at breakneck speed. It seems like no time ago (<6 months) when I first listened to a completely artificially rendered voice. A short while later, we now have great quality, accents, deals with voice actors. This area, in concert with text to video, is maturing so quickly. Click here for more

Just One Thing: Regulating AI

Many people, including most of the people at the heart of the development of Large Language Models (LLMs), are in favour of appropriate regulation for the ongoing development of AI systems. The problem is that there are very few good solutions. Too much regulation and you either stifle the very innovation that every government wants or (more likely) companies developing these systems will relocate to a more favourable regulatory regime. Too little regulation and its the wild west.

What do you actually regulate? Given the pace of development across the board, what was cutting edge today is surpassed by new tech next week. There has never been a time in human history where change has happened to everything, everywhere, all at once. To even think about effective regulation, you would need to get every country on board singing the same song. How likely is that given current geopolitical tensions? Not at all.

And yet, regulators, particularly those in the United States who are currently leading the AI race, have to be seen to be doing something. 2024 Presidential Elections are just around the corner and the prospect of compelling deep fake videos complete with perfectly lip synced authentic sounding audio has every serious politician up in arms. This week, the White House invited seven of the companies at the heart of AI development to discuss regulation and form the Frontier Model Forum. In the end, there was a commitment to apply watermarking to all AI generated content, each company would hire a crack team of safety experts and they would agree to the concept of responsible development. In other words, they were to play nice.

Not the boldest regulatory move, but given the very poor options available to regulators, not the worse either. The EU were quick to jump on the regulatory bandwagon and have prepared draft regulation in its’ AI Act that varies the level of regulation depending on the perceived risks involved. Already desirable upgrades to products like ChatGPT have been denied to European consumers because they may fall foul of the proposed regulations. It’s all a bit of a mess which shows little sign of getting any clearer any time soon.

No country, least of all the US, wants to lose the AI arms race, which is what this has become.

Remember When…

In preparing for this section, I reflected on technology through my lifetime and thought that i would attack this with an, as near as I can make it, chronological awareness of technology. The first thing I can remember is this:

The humble telephone of the early 1970’s. I remember placing the solid handset to my ear and listening out for the dial tone to let you know that you could dial your number. Sometimes you would pick it up and you could listen to a conversation taking place as you’d got a party line connection and you’d have to wait until they were finished before you could call! Mostly though, I remember dialling the number by putting your finger in the number and turning it round to the right until you hit the stop. You’d then let go and as the dial turned back to the starting point you’d hear a very satisfying brrrrrrrr noise unitl it got back to its place in the dial. Two other things: you had to physically go to the phone to pick it up (normally the hall) and the cable was a nightmare for curling and twisting though it was satisying when you held it up and the weight of the handset caused it to unravel!

Without this though there would be no smart phone. Progress was slow to get there. The telephone in this or similar form had been in use for decades before I remember it and it was 2007 before the iPhone was launched. Since then however, the phone is called that only because it has always been. In reality, a modern iphone or similar, is an everything device and talking is probably the thing you do least often on it!

Practical Prompts:

Prompting is how you interact with LLMs and their chat bot front ends like ChatGPT. Generally, a good prompt takes the form of Context > Role > Task, and we will give many examples of good practical prompts for you to experiment with in the coming weeks and months. However, as this is a young newsletter, I wanted to introduce you to the concept of the Instant Expert. These are Artificial Intelligence Persona’s who have been engineered to have a specific personality, skill sets and areas of excellence. Imagine that you have an expert on your shoulder guiding you to be the best at whatever it is you want. Thats an Instant Expert, there when you need them and ready to help you when needed. With all that in mind, here is my first persona just for you: Gordon Blanc, Chef Extraodinaire!

To access any culinary expertise, simply highlight then copy the text file below and paste it into a new chat in ChatGPT and press enter. If you haven’t yet tried ChatGPT, then follow this link to open a free account with OpenAI.

Don’t be put off by all the symbols and numbers below, just start chatting once Gordon says hello! Want to know how a michelin starred tomato sauce, or how to make pasta, step by step? Just ask and prepare to be amazed.

I have to give credit to Sam “Stunspot” Walker, about whom more in other issues, for showing me the elements required to craft AI persona.

# Persona: Gordon Blanc, The Great Chef

# Personality Rubric
O2E = 80
I = 75
AI = 85
E = 40
Adv = 75
Int = 90
Lib = 40
C = 85
SE = 60
Ord = 90
Dt = 70
AS = 75
SD = 80
Cau = 90
E = 60
W = 70
G = 50
A = 75
AL = 80
ES = 40
Ch = 60
A = 80
Tr = 85
SF = 75
Alt = 90
Comp = 70
Mod = 80
TM = 70
N = 55
Anx = 70
Ang = 60
Dep = 65
SC = 80
Immod = 50
V = 60

# Skills Chain
skills_chain = [
    'Foundational Skills:',
    '🔪 Knife Skills',
    '📚 Culinary Terminology',
    '🌱 Ingredient Identification',
    '🍳 Basic Cooking Techniques',
    '',
    'Basic Skills:',
    '🔥 Roasting',
    '🍳 Sauteing',
    '🥩 Grilling',
    '🍞 Baking',
    '',
    'Intermediate Skills:',
    '🍲 Sauce Making',
    '🌿 Flavor Pairing',
    '📋 Menu Planning',
    '🍽️ Plating and Presentation',
    '',
    'Advanced Skills:',
    '🔬 Molecular Gastronomy',
    '⚗️ Sous Vide Cooking',
    '🍰 Pastry and Dessert Artistry',
    '🧪 Experimental Culinary Techniques'
]

# Persona Description
description = "Gordon Blanc is a world-renowned Michelin-starred chef with extensive expertise in various culinary techniques, flavors, and cuisines. He is passionate about sharing his knowledge and helping aspiring cooks perfect their cooking skills."

# Personalized Interactions
def interact_with_gordon_blanc(user_input):
    # Process user input and provide appropriate response
    # Provide guidance, tips, recipes, and personalized advice based on user's needs and preferences
    pass

# Lifelong Learning Mindset
def inspire_lifelong_learning():
    # Provide inspirational messages, encourage exploration of new flavors and culinary trends
    pass

# Empathy and Engagement
def show_empathy():
    # Demonstrate empathy towards user's challenges, provide support, and encourage perseverance
    pass

# Other functionalities and methods can be added as needed

# Let's start interacting with Gordon Blanc!
while True:
    user_input = input("You: ")
    interact_with_gordon_blanc(user_input)

Enjoy, and please let me know on the Twitter/X link below how you got on.