It’s been a couple of weeks since the release of OpenAI’s chatbot ChatGPT and enough and more has already been said about it. Within a span of five days ChatGPT users surpassed a million users. What can the bot not do? It can work with a scope wide enough to raise an argument for ChatGPT being a replacement for Google, it can write a poem, generate complex Python code, write college-level essays and offer life advice. 

But as the hype created around ChatGPT spread like wildfire, one didn’t have to look too far to find fault with it. ChatGPT wasn’t always accurate and appeared confident as ever while stating misleading ‘facts’. A couple of days back, the word came from the horse’s mouth — OpenAI chief Sam Altman cautioned people against the ‘reliance on ChatGPT for factual queries’. In brief, it turns out ChatGPT isn’t as smart as it initially promised. 

Also, the excitement surrounding ChatGPT isn’t isolated. Tech investors are pouring money into startups that specialise in the generative AI field. According to tech industry firm PitchBook, a section of early-stage investors have shifted their attention from cryptocurrency and related businesses to generative AI startups. So while ChatGPT is the latest entrant, a slew of AI-powered writing assistants had already cornered the content creation segment ahead of ChatGPT. And with so many competitors in the space, there are ones specially tailored for specific roles. 

As the fervour around AI text and image generators grew, a startup developing AI content called Jasper raised USD 125 million in October this year pegging the company’s valuation at USD 1.5 billion. 

Earlier this year, Jasper expanded its product range to include Jasper Art, an AI text-to-image generator and a browser extension to use the tool across websites and platforms. Jasper is also in the middle of acquiring AI startup Outwrite, a grammar and style-checking platform like Grammarly, with more than a million users worldwide. Besides having functions like the basic plagiarism check, Outwrite is also focused on making stylistic suggestions around structure to the piece. 

Since its inception in 2020, Jasper has brought in around 100,000 customers closing out 2022 with over 100% increase in Annual Recurring Revenue (ARR). 

Cerebras: Differentiating factor for Jasper

While ChatGPT uses Microsoft’s Azure AI supercomputing infrastructure to train the GPT-3.5 models, on November 29, JasperAI announced it would start training its models on Cerebras Systems. Cerebras is hosted on the Cirrascale Cloud Services and is built to enable training of the GPT-class models including GPT-3, GPT-J and GPT-NeoX. Cerebras AI Model Studio allows users to connect to the Cerebras Wafer-Scale cluster including Cerebras’ new Andromeda AI supercomputer through the cloud. 

Cerebras AI Model Studio can train the GPT models at half the cost of traditional cloud providers with only a few lines of code. The company stated that users can train models with greater sequence lengths of up to 50,000 tokens on a 16-node Cerebras Wafer-Scale cluster. These models can range from between 1.3 billion and 175 billion parameters and will be trained 8 times faster with greater accuracy than on an A100 GPU. 

Basically, Jasper “can train its profoundly computationally complex models in a fraction of the time and extend the reach of generative AI models to the people” using Cerebras’ Andromeda AI supercomputer. According to Andrew Feldman, CEO and co-founder of Cerebras Systems, the AI Model Studio is intended to democratize AI by providing “access to multi-billion parameter NLP models on our powerful CS-2 clusters, with predictable, competitive model-as-a-service pricing”.

Many in the AI writing assistant market

A report by Verified Market Research stated that the writing assistant software market was valued at USD 410.92 million last year and has been projected to reach USD 1.37 billion by 2030, growing at a compounded rate of 14.27% from between 2023 and 2030. 

Writer, an AI writing assistant used specifically by companies for forming content guidelines and brand compliance rules, raised USD 21 million in Series A funding from VC firm Insight Partners and other CEOs including Scott Belsky and Jack Altman. 

Last year in November, AI writing developer Grammarly raised more than USD 200 million in funding led by investment management firm Baillie Gifford, pegging Grammarly’s valuation at around USD 13 billion. The startup claims to have more than 30 million users and its AI gives more than 100 billion suggestions for improving users’ writing every month. 

Then there’s Notion AI, a popular note-taking and database app which now uses AI to create social media posts, scaffolds for blogs and other content pieces. 

Ryter AI has a variety of options on its dashboard to help users adapt to a variety of different tones including formal, inspirational and even worried. It also has a language translation engine that is close to perfect.