FreshWrite about AI (again)

AI is an extension of your brain. And like all tools, the brain needs to learn how to use them.

About a year ago, when I was a senior editor for Illumination online publication, we started to get more and more stories that smelled not quite right. It was the dawn of AI, and then in March this year, OpenAI released ChatGPT-3.5, and things went wild. Almost 2/3 of the stories were written mainly by AI, which was disgusting. We became rejectors shovelling shit instead of editors who could help, guide and promote good writers.

It was hard to resist being biased when the vast majority of the AI-pulp came from Indian guys who thought that AI would make them rich quickly. I remember looking with disbelief when Deepak (not his real name) posted daily dozens of half-baked articles from sex to sedatives and from how-to-make-money listicles to how-to-make-your-garden-green-again. And then he got abusive when we rejected them instead of listening to our feedback.

Since then, Medium (the online platform I use to publish my work) has tightened rules, and AI-generated content is verboten, as the Germans say.

The birth of the beast – and the beauty contest.

I wrote in my earlier newsletter about how AI has helped me streamline my work, proofread and fix my typos and support research materials for my stories. It has been a blessing.

But that blessing was a bit geeky and required a lot of workarounds and a more profound knowledge of how AI, in general, and specifically ChatGPT, work. We saw an avalanche of separate ChatGPT-based apps coming, and the promise was way more often sexier than the delivery. Everybody tried to put a subscription fee on those brain children who were conceived in God knows which garage in the high hope of sky-high earnings. 

The app developers got the semen packed in the API keys from the freezers of OpenAI, and the circus was on the road. None of those bastards was indeed the second coming of the holy dude. But people paid for the tickets, and some got rich.

The incremental improvement and breakthrough cycles seemed months instead of years. However, it took decades to come to this point of the overnight success of the ChatGPT. 

John McCarthy coined the term artificial intelligence in 1955. Coincidently, I was born in 1956, which obviously makes me an AI-native. But only the work of researchers in the 80s, 90s and 2000s, like Geoffrey Hinton, Yann LeCun and Yoshua Bengio, made it possible to create the neural networks and large language models (LLM) that paved the way for the AI we see now. 

The mother of AI is in those neural networks and LLMs, but the father is cloud computing and the increasing speed of processors. That marriage gave birth to the monster we are afraid of now – or trying to tame it for humane purposes. 

The jury is still out; who is the beauty who can tame the beast? There will be somebody for sure to do that.

Why this rant now?

Last week, OpenAI had its first developer conference in San Fransisco, OpenAI DevDay. It was a milestone defining the path for AI and how we think, use and misuse it. 

I saw the opening keynote by the CEO and founder of OpenAI, Sam Altman. He is a geek but growing into a formidable businessman. His public speaking skills are not precisely Jobsian, but his message is. He is rapidly making his dent in the universe next to Steve Jobs' supernova. 

Watch his talk. I can assure you that this speech is one of the most important in the industry this year. 

Here is the link to the YouTube (they didn’t allow me to embed it here)

And there was this one huge thing that Altman talked about amongst some more technical and equally important things. The opportunity to build our own GPTs is mindblowing for us ordinary mortals. That's what made me tick.

After I saw the keynote, I jumped on my ChatGPT-4. You must have a paid subscription to do this, but it pays back in minutes. I created (in less than half an hour) several GPTs: 

1) One for proofreading the materials I write for my customers (I uploaded each customer's style guides, reference stories, etc., as a knowledge base for my GPT to use as a sounding board). It saves hours of my time and also the nerves of my customers because everything strictly follows their style guidelines with every comma, period and em-dash in the right place – not to mention, for example, Te Reo Māori phrases.

2) One to combine some of my online resources into one GPT to speed up the research I need to do. Again, hours of saved time. My GPT does things quickly without me googling everything and having a gazillion browser tabs open.

3) One just for fun, source me some quotations.

4) One for building imagery using DALLE.

I use these personal GPTs every day. They are custom-made, and I can tweak them to do more detailed and specific tasks for me. And I don't need to code or scratch my head with convoluted user interfaces. And the best thing is that you can do it, too.

I sourced a handy, concise and comprehensive tutorial on how to make your life easier with the GPTs. Here is the link to the YouTube tutorial.

The future is bright. 

I think the AI industry is moving in the right direction. There will be healthy competition. Elon Musk just announced his Gork, which seems not to be as dork as some think. It even appears to have a sense of humour, which ChatGPT does not have. Google is in the game, and so is Apple. Microsoft is riding the wave like it always does with stolen surfboards, but I am not too worried as long as OpenAI has the majority of the shares. 

The main thing is that you can do something valuable with the AI tools out of the box. And yes, there are security worries and legislative challenges: how do we regulate, restrict and protect ourselves from this ejaculation of AI semen and make sure that it won't rape our minds and abuse our children? 

The answer is easy: by learning about it, testing it and becoming responsible adults who know their boundaries and ensure consent is there. The final frontier in this battle lies in ethics, morals and respect for privacy. 

We might see terrible applications (like AI-equipped war machines) and fraudulent behaviours, but I believe people ultimately win the battle. 

So, now we have the tools. It is time to use your brain to learn how to use them best. That's AI for today.

Have a great week.

Jussi

The book of the week is Dreams From My Father.

I was listening to Barack Obama's book 'Dreams from My Father' while having my bushwalks. The man himself was the narrator, adding intimacy to the experience. 

It was published in 1995 and explores Obama's life before his political career. It looks at his early years, his struggle with racial identity, and his journey of self-discovery. The narrative is divided into three parts: 

1. Origins: Obama describes his childhood in Hawaii and Indonesia and his time at Occidental College in Los Angeles. Born to a Kenyan father and an American mother, he grapples with his mixed-race heritage. 

2. Chicago: After transferring to Columbia University in New York, Obama eventually moved to Chicago. Here, he works as a community organiser, profoundly shaping his understanding of racial and social issues. 

3. Kenya: The memoir culminates in Obama's emotional journey to Kenya, where he confronts the complexities of his father's legacy and meets his Kenyan relatives. 

It is a heartwarming and touching tribute to the diversity that has made Obama who he is. His style and tone are eloquent in a way that resonates with the listener but is also filled with gentle humour and poignant observations. 

I might say, don't read it – listen to it, and you won't be disappointed. You can get the audiobook from Apple Bookshop.

My recent articles on Medium and Vocal