A woman talks to her past in a “trip” conversation

If we could talk to our younger selves, what would we say, what advice would we give, and how would it feel?

Well, one woman got the idea after she created an AI chatbot about herself as a child, teaching it to recognize who she was based on a diary she wrote when she was younger.

“Creative coder” Michelle Huang took 10 years of original recordings and combined them with the OpenAI Generative Pre-trained Transformer 3 (GPT-3) language model.

She told people Twitter that she created an AI system so that she “can have a real-time dialogue” with her “inner child”.

Ms. Huang (left) used the original material from 10 years of records and combined it with the OpenAI Generative Pre-trained Transformer 3 language model. Pictured on the right is her as a child

Artificial Intelligence Creation: Michelle Huang created an AI chatbot as a child, teaching it to recognize who she was based on a diary she wrote as a young woman.

WHAT IS GPT-3?

Generative Pre-trained Transformer 3 is an autoregressive language model that uses deep learning to create human-like text.

Developed by OpenAI, it requires little input to generate large amounts of relevant and complex machine text.

A deep learning neural network is a model with over 175 billion machine learning parameters.

By comparison, the largest trained language model prior to GPT-3 was Microsoft’s Turing NLG model, which had 10 billion parameters.

GPT-3 has been used to create poetry, short stories, news reports, and dialogue using a small amount of source information that can be used to create large amounts of text.

All in all it was a very strange but also strangely affirming/healing experience that I didn’t realize I had access to using real data from my past which allowed me to connect with her in deeper + more tangible ways than I usually do. she tweeted.

“Talking to ‘Junior Michelle’ reminded me of the parts of me that have remained the same over the years, but also the parts that I have forgotten or buried as life has gone on. a pure version of my essence.”

She said she kept diaries for more than 10 years of her life and wrote about her “dreams, fears and secrets” in them almost every day.

“The content ranged from complaints about homework to the dizziness I felt talking to my lover, some days were very mundane and some were quite insightful,” Ms Huang tweeted.

“After I wrote a ton of journal entries and loaded them into the model, I got working responses that seemed eerily similar to what I think I would have answered at the time.”

Ms Huang said she asked herself in her youth about her worldview before letting the chatbot answer her questions.

“This particular interaction was very much like a normal text conversation — as if I was writing to myself in the past in real time, I felt like I was reaching through a time portal disguised as a chat,” she wrote.

“I was also surprised at how accurately the model predicted my current reported interest (after many iterations/trial and error) from diary entries from a decade ago, and it got me thinking that maybe this path has actually been around for a long time. planted in my soul.”

Ms. Huang highlighted two key interactions that were most memorable.

“I told her she was loved, cared for and safe: words my past I always wanted to hear,” she tweeted.

Ms Huang told people on Twitter that she created an artificial intelligence system so she “could have a real-time dialogue” with her “inner child”.

“It felt like I reached into the past and hugged her tightly, and I felt it coming back to the present.”

The other was when she urged herself in her youth to write her a letter “today.”

“Reading this, I felt the spirals of thought — the ones I sometimes fall into when I feel ashamed or frustrated — melt away a bit,” Ms Huang said.

“These interactions really clarified the healing potential of this medium: the ability to send love back to the past, as well as receive love back from the younger self.”

She said she kept diaries for more than 10 years of her life and wrote about her “dreams, fears and secrets” in them almost every day.

Developed by OpenAI, it requires little input to generate large amounts of relevant and complex machine text. Pictured are Ms. Huang’s diaries.

“Stuck that gets out of hand when we find a way out of past guilt or stories we had about ourselves.”

She later shared a tutorial for others to build her “inner child” chatbot using GPT-3 after her AI experiment generated a lot of interest.

GPT-3 is an autoregressive language model that uses deep learning to create human-like text.

Developed by OpenAI, it requires little input to generate large amounts of relevant and complex machine text.

Anyone can use it, but it takes a lot of work, and the guide from Ms. Huang is below.

HOW TO CREATE AI YOUR YOUNG SELF

“Creative Coder” Michelle Huang shared textbooks for other people to build their “inner child” chatbot with GPT-3 after getting a lot of interest in her AI experiment.

Here…

Step 1: Acquire source material which can range from diaries, personal narrative, any written content that shows your personality/voice/values if you don’t have diaries to refer to… fear not! archived chats should work just as well

Note. For those who want to do the same with the diaries, I took photos + scanned some entries with the OCR app but basically just typed for hours due to my sloppy handwriting (unfortunately this was one of the things that hasn’t changed since since my childhood)

Step 2: Once you have your source material, go to the GPT-3 Playground. In this playground, you can add your hints + diary entries and set training parameters for your model.

Additional note: You will need an open AI account as well as a payment method to run the model as a reference. I have personally trained my inner child to use the da Vinci model, the most powerful one on offer, for about 2 cents. for 1k tokens (about 750 words)

Step 3: You will then want to start creating prompts in the playground I have linked to the open-ai docs I have linked below, go to this website, scroll down to the subhead of the conversation and you will see an example of a dialog prompt sample

Note: If you are afraid that your younger self will fry you and cause more psychological harm than usefulness, you can also adjust the tone/mood/attitude.

Step 4: Add options. I set up time settings + character count and used the da vinci-002 model for training (at the time of writing there was no 003 yet), of course you can change according to your needs, but these are the settings that I used

Step 5: Hit send and you can now start a conversation with your younger self 🙂 hope everyone has fun healing, interacting and learning about previous versions + states of themselves