29 November 2022 – The Survivor of Paradise
The sun was setting in the sky, painting the horizon in hues of orange and red. Sarah was trance-like, her mind focused on the moment’s beauty. Suddenly, the tranquility of the evening was shattered by the sound of alarms blaring from the cockpit of the small plane she was flying in. The aircraft was in trouble, and the pilot quickly tried to adjust the controls to level it out.
But it was too late. The plane was going down, and Sarah was the only one on board. She held tightly to her seat as the plane plummeted towards the ocean. As it skimmed the surface of the water, Sarah felt a sudden jolt as the plane slammed into the ground. She was thrown from her seat, and everything went black.
When Sarah awoke, she found herself lying on a beach in a tropical paradise. She had no idea where she was or what had happened. Sarah tried to piece it together in her mind, but she was still shocked by the crash. Looking around, she realized she was the only survivor. She was alone in this strange place, with no food, water, or way of getting home. She felt a wave of panic and fear wash over her. The story beginning was written by the newest text-completion AI model, text-davinci-003.
I hate doing two BLOGs on the same subject in a row, but this is a particular case. After writing the previouis BLOG on Writer’s Block, this came to my inbox yesterday from OpenAI.
We’re excited to announce the release of the newest addition to the GPT-3 model family: `text-davinci-003.` It includes the following improvements:
- It produces higher-quality writing. This will help your applications deliver clearer, more engaging, and more compelling content.
- It can handle more complex instructions, meaning you can get even more creative with how you use its capabilities now.
- It’s better at longer-form content generation, allowing you to take on tasks that would have previously been too difficult to achieve.
I decided to give it a run, and I was surprised at how much better it was. This little trial took five minutes. I opened the OpenAI Playground window and typed in these simple instructions.
Write the beginning paragraphs of a short story about a teenage girl named Sarah that is the only survivor of a plane crash on a tropical island. She is unprepared to cope with the situation and must move forward by human instinct alone.
It returned three paragraphs in fewer than two seconds. I loaded these paragraphs into Microsoft Word, where I have Grammarly installed, fixed a few little problems, and had what you read at the beginning of this post. You could (possibly) walk your way through a complete story by just giving the machine instructions for each scene. What a way to write the first draft!
Okay, just how much did this cost compared to Jasper, let’s ask. OpenAI charge 20 cents per 1000 tokens. This text, including my instructions, was 304 tokens. The cost was under seven pennies.
27 November 2022 – All work and no play make Jack a dull boy.
Perhaps the most dramatic case of writer’s block was a condition that afflicted Jack Nicholson in the 1980 psychological horror film The Shining. Nicholson played a “writer’s Blocked” author who takes his wife and son to the remote Overlook Hotel in the Rocky Mountains, which closes every winter season (no distractions there to interfere with writing). Once the doors are closed, and the snowpack prevents anyone from leaving, Nicolson’s character spirals into mental illness.
One of the most memorable scenes is when Nicholson’s wife sneaks into his writing room and discovers that her husband has repeatedly typed hundreds of pages with the same sentence. “All work and no play make Jack a dull boy.”
Writer’s block has a long history but has not always gone by that name. Lawrence R. Samuel wrote a brief history of writer’s block for Psychology Today.
While writer’s block has no doubt existed since the days of cave people, the much-dreaded psychological syndrome did not receive much attention within literary circles until the latter part of the 20th century. As the writing process was closely scrutinized after the flourishing of the self-help movement and the rise of a therapy-oriented culture in the 1970s, the notion of a block became a primary concern. Writers hesitated to even mention the “b-word,” afraid that it, like impotence, perhaps, would become a self-fulfilling prophecy merely by acknowledging it was a possibility. (“Dry spell” was a bit better.)
If you have never experienced writer’s block, you’re lucky. With my last novel, Through Both Eyes, I finished about half of the work in the first six months of writing, and then I seemed to burn out. I had to put the book aside and do other things. I decided that I ought to discover a new species before I die, an unattained childhood ambition. I planned my attack by researching which taxa likely contained the most undescribed species. Being in south Texas, I learned that very little work had been done on mushrooms. So, with a passion, I collected and identified all the mushrooms I could find. That was not easy. I resorted to taking tissue samples and sending them off for DNA analysis. The DNA sequences allowed me to compare my shrooms to officially identified specimens in a worldwide gene bank. The most naive mistake I made was to think I could be the one that introduced a new species to the scientific world. Not! I didn’t have the academic background to be taken seriously. I’d collected three specimens that had no match in the Genbank database and could be undescribed species. I finally had to turn these three specimens over to experts. They are studying them now. One day the mycologists might get around to writing a scientific paper on my discoveries, and my name might be mentioned as the collector. Or maybe they get writer’s block too.
This idea might help. The sophistication of artificial intelligence completion engines is reaching astonishing levels. Think of them as highly-evolved autocorrection tools. You’ve likely cursed these completion apps after noticing that the text you sent went something like this.
You meant to say, “I can’t wait to see those big beautiful dimples of yours.”
Your phone autocorrected your message to read, “I can’t wait to see those big beautiful nipples of yours.”
There is a growing number of AI options out there to help writers move their work along. Some will help with your written text, and others will even paint you a copyright-free cover image for your new book. To find most of them, do a Google search for Jasper Alternatives. These can get pricey, so I go straight to the source and use OpenAI Playground. It’s not as versatile, but it costs very little to use.
Let’s suppose that you are writing a simple Hero’s Journey story. You jot down the three parts.
Departure: the Hero leaves the familiar world behind.
Initiation: the Hero learns to navigate the unfamiliar world of adventure.
Return: the Hero returns to the familiar world.
Your Hero is an 18th-century frontier woodsman whose wife faces a problematic birthing, and the nearest midwife is eight miles away through rugged terrain and deep snow. He kisses his wife, calls his loyal huskie, and sets out in the dark of the night. You try to write the beginning but . . . . Oh No! Writer’s Block! You just can’t get started. You’ve tried a dozen first paragraphs, and they’re all duds. Now you can’t put a single word down. You’re locked up. Time to bring in the big guns. You navigate to the OpenAI playground, turn the temperature to point eight, and type in the instruction.
Write the beginning paragraphs for a story using the Hero’s journey model. The story is about an 18th-century frontier woodsman whose wife faces a problematic birthing, and the nearest midwife is eight miles away through rugged terrain and deep snow. He kisses his wife, calls his loyal huskie, and sets out in the dark of the night.
A second later, the AI completion engine gives you a possible start.
It was the middle of winter on the American frontier, and frontiersman John had been hunting all day. He eagerly awaited returning to his warm cabin and seeing his wife, Mary. However, when he arrived, he could tell something was wrong. Mary was in labor, and the nearest midwife was eight miles away through rugged terrain and deep snow. John knew he had to set out immediately if they were going to have any chance of saving the baby. He kissed his wife goodbye, called his loyal huskie, and set out into the night. John had no idea what was ahead of him, but he knew he had to try.
Not too bad. You can tweak it a little later. Let’s click the generation wheel to see what more we can get.
The blowing snow was so thick that John soon lost his way. He wandered for hours, fighting against the cold and the drifts of snow, but he couldn’t find his way back to the cabin. When he finally collapsed from exhaustion, he knew he wouldn’t make it much further.
This completion has a problem. It assumes that John is attempting to get back to his own cabin. Let’s change a couple of words and say it like this.
The blowing snow was so thick that John soon lost his way. He wandered for hours, fighting against the cold and the drifts of snow, but he couldn’t find his way to the midwife’s cabin. When he finally collapsed from exhaustion, he knew he wouldn’t make it much further.
We could go on, but let’s stop here. The AI completion engine could write you a completed story, but you get the idea. These simple suggestions written by artificial intellegence may get you past your writer’s block.