AI Writing Tools: The Ghostwriters Behind Substack’s Top Newsletters

“Some of Substack’s Biggest Newsletters Rely On AI Writing Tools”

“In April, Substack launched an open beta for an AI writing tool called Threader, which uses an AI model developed by OpenAI called GPT-3. Some of the writers using it found it not so helpful.”

Ah, Substack’s grand experiment, Threader, paving the way for a brave new world of artificial intelligence innovation in the writing industry. Or, as some might depict it, turning out to be more of a hindrance than a help. Truly, nothing warms the heart more than a story of advanced technology slightly confounding those who were eager to welcome it with open arms.

Threader was released in April, a gleaming new plaything for Substack writers, a tool developed by OpenAI’s GPT-3. Substack’s idealistic vision was to enhance creativity, to inspire writers with artificial intelligence, to make the writing process smoother, faster, better. Who wouldn’t love a helping hand in slogging through the daily grind of writing? Imagine the sweet serenade of a few keystrokes as all the heavy lifting of creation and innovation is performed by AI. Heavenly.

Alas, the reality turned out to be slightly less idyllic. As it happens, writers found Threader to be less of a help, and more of a nuisance. It was discovered that the tool, although packed full of incredible AI power, often generated content that was similar to previous inputs, creating a troubling pattern of repetition. With this twist of fate, writers ended up spending more time rewriting and editing the AI-written text. Truly, the irony is delicious.

Now, don’t get it wrong. AI and machine learning infuse the sector with innovation, and their potential is limitless. But, alas, Threader has shown us that there might be a bit more work needed before these tools can replace a good old-fashioned writer sitting in front of a keyboard, tapping out words from their own mind.

So, what can be learned from our dear friend Threader’s misadventures on Substack? Well, for starters, innovation and ambition must be paired with thorough testing and realistic expectations. After all, dreams of a seamlessly helpful AI writing assistant just aren’t quite reality… yet. Rest assured, AI in writing isn’t going anywhere, it’s just mildly stumbling over its own shoelaces on the way to the future.

Threader’s little hiccup offers a critical lesson – intelligence, artificial or not, means little without the nuanced understanding, creative spark, and unique perspective that a human being brings to the process of writing. For better or worse, us humans are here to stay in the written world. You’re stuck with us, Threader.

So let’s all raise a metaphorical glass to Threader, proof that the road to innovation is paved with good intentions, and occasionally, some light comedic material. Here’s to the future of AI in writing, it won’t always be smooth sailing, but with every stumble, we’re getting a step closer to something truly exciting.

Read the original article here: https://www.wired.com/story/substacks-writers-use-ai-chatgpt/