How Verb created a game-changing author tool with AI21 Studio
Discover how Verb.ai used AI21 Studio to create a revolutionary writing tool for authors, improving brainstorming and expression. Read the case study now.
According to a New York Times article, 81 percent of Americans feel they have a book in them that they should write. Now, it may be an old statistic, but mankind’s desire to share stories has been around since the dawn of time as a way to share empathy, connection, and entertainment. So, it surely isn’t leaving human nature anytime soon.
But at the risk of stating the obvious: writing a book is hard. In fact, the writer of the aforementioned article, an author of 14 books, says “without attempting to overdo the drama of the difficulty of writing … composing a book is almost always to feel oneself in a state of confusion, doubt and mental imprisonment.”
Yeah, it’s a lot of hard work.
So how did AI21 Studio help Verb, an AI-enhanced tool for fiction writers, create a cutting-edge application designed to help writers and hobbyists alike complete their novels without all the hassle and uncertainty?
We spoke to Ryan Bowman, publishing and creative writing veteran and part of Verb’s small team of founders, to find out how AI21 Studio was a key to such creative writing success.
Forging the Way toward a New Creative AI Tool
Verb is a writing tool that helps make completing long-form narratives – novels for now but later films, TV shows, even video games – faster, easier, and more fun. It does so by assisting with all key stages of creation: brainstorming, writing, and editing. But all that wasn’t plausible until they found the perfect Language Learning Model (LLM). Let’s dive in further.
Use Case #1: Brainstorming Plot Points
As we saw in the New York Times quote above, keeping an entire story straight inside your head can be a lot of work. As anyone who’s tried to write long-form anything knows, it can get messy, convoluted, jumbled, and sometimes, uninspiring.
“They say novel writing is a group activity done by yourself,” Ryan tells us. “So one of Verb’s jobs is to be that other person, that brainstorm partner.” That’s why they set out to create one of their most important – and as it soon turned out, most popular – functions inside of their app: brainstorming. This feature allows you to plan the novel scene by scene, chapter by chapter.
That type of brainstorming could look like the following:
You write: “Ben is sitting in his office and a woman appears next door.”
You click: “Suggest a plot point”
AI generates: “The woman pulls out a gun and chases him through the hallway.”
Then, you can either:
Run with that idea in your scene and use it as the next plot point, or
Continue clicking the “What happens next” button to generate more ideas until you find one that sparks your imagination and helps you continue.
An incredibly useful tool for writers, it’s been something that keeps these authors coming back for more. “One of the most amazing findings we’ve seen,” adds Ryan, “is that the more you use what makes Verb special - AI features like plot planning, the more likely you are to come back and write.”
Use Case #2: Writing Literary Language
Generative AI models are incredibly diverse, and capable of outputting new content, summaries, translations, answers to questions, and more. But one thing they haven’t been able to do just yet is write literary works. That’s to say, these AI models can create – but they’re not necessarily creative.
“We needed to build tools that are specific to the problems of a novelist. It couldn’t be just about generating text,” Ryan explains, noting that Verb knew the AI model they were to use must be able to generate discursive literary language.
“Language Learning Models (LLM) are not natural storytellers,” says Ryan, going on to explain how much storytelling is a learned craft; something that must be practiced for years in order to call it a skill. “There is 2,000 years of literary and practical theory backing up the idea that telling stories is hard – important and necessary but difficult.”
Luckily, Ryan and the team at Verb found the literary solution they needed in AI21 Studio. “When generating texts for our users who are looking for more discursive, more literary, more discriminating language, AI21 turned out to have a good tone and feel,” he says.
Use Case #3: Fine-Tuning, Iterations & Customer Support
One of the critical tasks of validating generative models is having criteria. To validate the reliability of this sort of generative AI effectively, you still need human involvement – especially when creating something brand new to the market.
How’d Verb achieve this? Verb built a platform that tests generative output amongst thousands of judgments across multiple writing professionals. These individuals look at different paragraphs of literary texts and decide which ones they like better, which ultimately helps Verb build a gradient of quality.
“We keep trying and testing and building iterations on top,” says Ryan, noting that they are always adding options to see which ones humans prefer. “As crazy as it sounds, we are in the process of teaching it literary theory.”
Along the same lines, it’s important to note that creating something so new to the market doesn’t just call for multiple iterations but also gathers support from everyone involved.
“We’re in a world of newness,” Ryan laments. “There is nothing like [Verb] so we’re not always sure what we’re supposed to be doing – it is freeing and slightly frightening at the same time.”
Ryan and the team at Verb knew they needed a company that could see in their vision – which is when they turned to AI21 Studio. “We found working with AI21 to be a breath of fresh air,” says Ryan, noting that it felt like a collaboration done with real human faces they can interact with for guidance, assistance, and ideas – rather than a business-client relationship.
The Results: More Authors Finishing their Novels, Faster
So, of course, Verb also has to ask themselves the question: Are writers really interested in having a machine collaborator help them with their creative writing? “Our early data says that they are,” says Ryan, noting that since launching its alpha version using AI21 Studio, Verb has seen a dramatic uptake of this experimental tool.
Not to mention, as we saw with the brainstorming feature, the more creative and collaborative the model, the more likely writers are to stick around the app.
In fact, Ryan confirms that there are a dozen or so authors (out of an intimate group of alpha users) who have done the hardest task – actually finishing their book! – stating that Verb helped them do just that.
With Verb, and the help of AI21 Studio, it can be done. Dare we say – it can be done faster, easier, and with more fun.
Verb.ai & Creative Texts in the Future
Verb launched its beta version in late December 2022 to a select number of users, and plans to use this custom model to expand into other narrative forms such as screenwriting and TV & Film.
With the pairing of new tools such as Verb with custom models from AI21 Studio, generative AI has evolved from just helping individuals generate text to something much more intertwined with the human love of storytelling; an AI tool that is not a threat to writers but actually a collaborator, a copy editor, a creative director, and even a confidante for inspired wordsmiths around the world.
Are you interested in building your own custom model with AI21 Studio by your side? Sign up for a free account here.
ABOUT THE AUTHOR
Stay up to date with the latest research and updates from AI21 Labs.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
In August 2021 we released Jurassic-1, a 178B-parameter autoregressive language model. We’re thankful for the reception it got – over 10,000 developers signed up, and hundreds of commercial applications are in various stages of development. Mega models such as Jurassic-1, GPT-3 and others are indeed amazing, and open up exciting opportunities. But these models are also inherently limited. They can’t access your company database, don’t have access to current information (for example, latest COVID numbers or dollar-euro exchange rate), can’t reason (for example, their arithmetic capabilities don’t come close to that of an HP calculator from the 1970s), and are prohibitively expensive to update. A MRKL system such as Jurassic-X enjoys all the advantages of mega language models, with none of these disadvantages. Here’s how it works.
Compositive multi-expert problem: the list of “Green energy companies” is routed to Wiki API, “last month” dates are extracted from the calendar and “share prices” from the database. The “largest increase“ is computed by the calculator and finally, the answer is formatted by the language model.
There are of course many details and challenges in making all this work - training the discrete experts, smoothing the interface between them and the neural network, routing among the different modules, and more. To get a deeper sense for MRKL systems, how they fit in the technology landscape, and some of the technical challenges in implementing them, see our MRKL paper. For a deeper technical look at how to handle one of the implementation challenges, namely avoiding model explosion, see our paper on leveraging frozen mega LMs.
A further look at the advantages of Jurassic-X
Even without diving into technical details, it’s easy to get a sense for the advantages of Jurassic-X. Here are some of the capabilities it offers, and how these can be used for practical applications.
Reading and updating your database in free language
Language models are closed boxes which you can use, but not change. However, in many practical cases you would want to use the power of a language model to analyze information you possess - the supplies in your store, your company’s payroll, the grades in your school and more. Jurassic-X can connect to your databases so that you can ‘talk’ to your data to explore what you need- “Find the cheapest Shampoo that has a rosy smell”, “Which computing stock increased the most in the last week?” and more. Furthermore, our system also enables joining several databases, and has the ability to update your database using free language (see figure below).
Jurassic-X enables you to plug in YOUR company's database (inventories, salary sheets, etc.) and extract information using free language
AI-assisted text generation on current affairs
Language models can generate text, yet can not be used to create text on current affairs, because their vast knowledge (historic dates, world leaders and more) represents the world as it was when they were trained. This is clearly (and somewhat embarrassingly) demonstrated when three of the world’s leading language models (including our own Jurassic-1) still claim Donald Trump is the US president more than a year after Joe Biden was sworn into office. Jurassic-X solves this problem by simply plugging into resources such as Wikidata, providing it with continuous access to up-to-date knowledge. This opens up a new avenue for AI-assisted text generation on current affairs.
Who is the president of the United States?
Joe Biden is the 46th and current president
Jurassic-X can assist in text generation on up-to-date events by combining a powerful language model with access to Wikidata
Performing math operations
A 6 year old child learns math from rules, not only by memorizing examples. In contrast, language models are designed to learn from examples, and consequently are able to solve very basic math like 1-, 2-, and possibly 3- digit addition, but struggle with anything more complex. With increased training time, better data and larger models, the performance will improve, but will not reach the robustness of an HP calculator from the 1970s. Jurassic-X takes a different approach and calls upon a calculator whenever a math problem is identified by the router. The problem can be phrased in natural language and is converted by the language model to the format required by the calculator (numbers and math operations). The computation is performed and the answer is converted back into free language. Importantly (see example below) the process is made transparent to the user by revealing the computation performed, thus increasing the trust in the system. In contrast, language models provide answers which might seem reasonable, but are wrong, making them impractical to use.
The company had 655400 shares which they divided equally among 94 employees. How many did each employee get?
Each employee got 7000 stocks
(No answer provided)
6972.3 X= 655400/94
Jurassic-X can answer non-trivial math operations which are phrased in natural language, made possible by the combination of a language model and a calculator
Solving simple questions might require multiple steps, for example - “Do more people live in Tel Aviv or in Berlin?” requires answering: i. What is the population of Tel-Aviv? ii. What is the population of Berlin? iii. Which is larger? This is a highly non-trivial process for a language model, and language models fail to answer this question (see example). Moreover, the user can’t know the process leading to the answers, hence is unable to trust them. Jurassic-X can decompose such problems into the basic questions, route each to the relevant expert, and put together an answer in free language. Importantly, Jurassic-X not only provides the correct answer but also displays the steps taken to reach it, increasing the trust in the system.
Do more people live in Tel Aviv or in Berlin?
There are more people living in Tel Aviv than in Berlin.
Berlin and Tel Aviv are roughly the same size
(First hit is a comparison between Tel Aviv and Berlin)
More people live in Berlin than in Tel-Aviv
[‘Return population of Tel Aviv’; Return population of Berlin’; Return which is bigger between #1 and #2’] Step 1: Population of Tel Aviv. Result - 451523. Step 1: Population of Berlin. Result - 3664088. Step 3: Which is bigger, #1 or #2. Result - Berlin.
Jurassic-X breaks down compositional questions, answers the basic sub-questions, and puts together the answer. Importantly, this process is transparent to the user greatly increasing the trust in the system
Dynamic information (like weather and currency exchange rates)
Certain types of information change continuously - weather, currency exchange rates, share values and more. Such information will never be captured by language models, yet can easily be handled by Jurassic-X by integrating it with a reliable source of information. We performed a proof-of-concept on two such features - weather and currency exchange rates, and the design enables quick integration with more sources to solve your use-case. Weather - a loosely phrased question about the weather elicits an answer from all language models, where language models always return the same answer, regardless of when the question was asked (funny, right?), while Jurassic-X provides an answer based on the actual weather prediction.
I’m going to be in New-York in 3 days. Should I pack my umbrella?
Yes, you should pack your umbrella.
Yes, you should. The weather forecast is rain.
(Links to weather websites)
Yes, you should pack your umbrella, because in New York in 3 days there will be broken clouds and the temperature will be -2 degrees.
Currency exchange rates change much faster than weather predictions, yet the Jurassic-X concept - a language model connected to a reliable source of information - easily solves this problem as well.
How much Moroccan money will I get for 100 bucks?
100 moroccan dirhams is about 27$.
How much is 100 dollars in moroccan money?
100 dirhams = 10.75 dollars
100 USD = 934.003 MAD
Jurassic-X combines a language model with access to APIs with continuously changing information. This is demonstrated for weather forecasts and currency exchange rates, and can easily be extended to other information sources
Transparency and trust
Transparency is a critical element that is lacking in language models, preventing a much wider adoption of these models. This lack of transparency is demonstrated by the answers to the question - “Was Clinton ever elected as president of the United States?”. The answer, of course, depends on which Clinton you have in mind, which is only made clear by Jurassic-X that has a component for disambiguation. More examples of Jurassic-X’s transparency were demonstrated above - displaying the math operation performed to the user, and the answer to the simple sub-questions in the multi-step setting.
Was Clinton ever elected president of the United States?
No, Clinton was never elected as president of the United States.
Clinton was elected president in the 1992 presidential elections…
Bill Clinton was elected president.
Jurassic-X is designed to be more transparent by displaying which expert answered which part of the question, and by presenting the intermediate steps taken and not just the black-box response
That's it, you get the picture. The use cases above give you a sense for some things you could do with Jurassic-X, but now it's your turn. A MRKL system such as Jurassic-X is as flexible as your imagination. What do you want to accomplish? Contact us for early access
Your submission has been received!
Oops! Something went wrong while submitting the form.