As life science marketers, it's essential for us all to stay on top of the latest trends in AI to ensure we're reaching our target audiences in the most effective way possible. However, with so much information out there, it can be challenging to keep up. To help, BioStrata's co-founder Paul Avery has teamed up with his industry friend and colleague, Martin Broadhurst, to do all the reading, research, and tool testing... so that you don't have to. They cover everything you need to know in a weekly podcast released every Friday: Artificially Intelligent Marketing.
In the latest episode, Paul and Martin discuss topics including:
- The launch of ChatGPT plugins
- Nvidia Foundations
- Google announcing Bard
- New creative AI tools including Canva AI, Adobe Firefly, and Runway ML's new text-to-video tool
- Recent papers including Microsoft's paper on GPT-4's advancements towards artificial general intelligence (AGI), and OpenAI's paper discussing the impact of AI on jobs
- Tool of the week: AskYourPDF and ChatShape
- Showcase segment on Ninja marketing and sales application of AI.
You can listen to the episode below, or by subscribing on your favourite podcast platform. We've also briefly summarised the podcast in the rest of this blog post. Use the links above to skip ahead to specific sections of the blog.
OpenAI launches ChatGPT plugins
OpenAI has revealed an innovative plugins platform for ChatGPT, which allows third-party developers to create and integrate plugins on top of the ChatGPT system. This marks a significant shift in the way users interact with the AI, as it now has the ability to perform more advanced and diverse tasks through these plugins.
OpenAI has already built a few plugins as proof of concept, one of which is a web browser. This enables ChatGPT to visit websites, gather relevant information related to user queries, and present it back to the user as part of the conversational interface of ChatGPT. In particular, this new web browser plugin addresses one of the major criticisms of ChatGPT, which was that it was only trained on data up to mid-2021. By allowing the AI to access websites and retrieve current information, this limitation is significantly reduced.
OpenAI has also partnered with around 20 initial providers. One example is the Kayak plugin, which allows users to ask ChatGPT to find hotels, flights, and other travel-related information based on specific criteria. Another significant partnership is with Zapier, a platform that connects ChatGPT to over 5,000 other apps, enabling users to make requests in natural language for various tasks, such as adding contacts to a CRM, searching for information in a spreadsheet or sending emails based on natural language-driven requests in the ChatGPT interface.
Perhaps even more excitingly, the introduction of this platform opens up a world of possibilities for others to build plugins, potentially leading to the development of an app marketplace specifically for ChatGPT. The value of the ChatGPT marketplace could be immense, with some predicting that users may end up spending a significant portion of their work time interacting with a ChatGPT-style interface. This would fundamentally change the way people work and create substantial value for those who develop the most successful and sought-after applications, just as happened during the smartphone era.
The ease of creating plugins for ChatGPT is another notable aspect of this development. Developers only need to write an OpenAPI manifest for their API, using human language descriptions for everything, and ChatGPT will handle the rest. This simplifies the complex process of building APIs and allows ChatGPT to do the heavy lifting when it comes to interpreting and processing data.
For marketers, the introduction of ChatGPT plugins could lead to a plethora of new tools that can help with content creation, data querying, and automating tasks that would otherwise require multiple clicks or manual work. The ease of building custom plugins also opens up the possibility of affordable, personalised solutions tailored to individual use cases (so start drawing up your wish list now!).
However, this development also raises important questions and dilemmas for marketers, especially about the future of search engine optimisation (SEO) and the way people interact with online content. For example, if ChatGPT becomes the primary method of accessing information, traditional channels like Google may experience a decline in usage. This could force marketers and SEO professionals to adapt their strategies and optimise content for visibility within ChatGPT's search results, rather than search engines like Google. In itself, this may still not be an ideal solution, as ChatGPT's ability to access websites and retrieve information without necessarily crediting the source might limit the impact of content-driven marketing to send traffic back to your website. Some have speculated that this could lead to an increase in gated content and paywalls as publishers and content creators seek to protect their intellectual property. It's worth considering how your content marketing and/or SEO strategy would need to change if you could no longer rely on creating text-based content to drive traffic from search engines.
Also note, as the ChatGPT plugins platform gains traction, it may spur competition among major tech companies, such as Google, Amazon, and Microsoft, to develop their own AI-driven platforms that offer similar functionality. This competition could lead to a rapid expansion of the AI-powered virtual assistant ecosystem, pushing the technology further and creating new opportunities for businesses and developers alike.
Nvidia unveils foundations
Nvidia Foundations is a new cloud services platform that enables developers to build on top of generative AI tools. Known for their graphics-based products, Nvidia has become heavily invested in AI. They now offer tools such as BioNeMo, Nvidia Picasso, custom large language models, and visual models tailored for domain-specific tasks. Accessible via browser and simple API plugins, Nvidia Foundations is inviting developers to build their products on their platform. As Engadget puts it, "NVIDIA's newest cloud services offering, AI Foundations, will allow businesses lacking the time and money to develop their own models from scratch 'to build, refine and operate custom large language models and generative AI models that are trained with their own proprietary data and created for their unique domain-specific tasks."'
This development is particularly interesting for marketers, as it introduces more diversity into the AI marketplace and emphasises the importance of AI adoption. With a focus on enabling AI-assisted content production and streamlining creative workflows (among other things), Nvidia Foundations has the potential to revolutionise the creative industry. As AI continues to evolve, it becomes increasingly important for businesses to integrate AI technology into their digital products to stay competitive and avoid becoming laggards in their industries. The launch of Nvidia Foundations highlights the growing number of tools and resources available for developers to leverage AI capabilities.
Of particular interest to life science businesses is BioNeMo, one of Nvidia Foundations' tools. As Nvidia describes it, "BioNeMo is a cloud service for generative AI in drug discovery. With NVIDIA cloud APIs, researchers can quickly customise and deploy domain-specific, state-of-the-art generative and predictive biomolecular AI models at scale. BioNeMo enables researchers and developers to use generative AI models to rapidly generate the structure and function of proteins and biomolecules, accelerating the creation of new drug candidates". Exciting stuff!
Google gets into the chat assistant game
This week, Google announced Bard, its chatbot offering which is equivalent to ChatGPT. For us, the launch of Bard demonstrates that Google is trying to play catch-up in the AI chatbot market. Available through a waiting list, Bard is designed for content generation, brainstorming ideas, and other use cases similar to what ChatGPT offers. Some of its advantages include the ability to retrieve information from the web, making it more versatile than relying solely on its training data.
Although Google is making an effort to enter the AI chatbot space, it still seems to lag behind its competitors. However, we suggest that it may still be worth signing up for the waitlist to explore how Bard can be used for content generation and to potentially achieve different results. Additionally, the introduction of Bard could give users a glimpse into the kind of chat assistant experience they might have across Google Apps in the future, including Gmail, Docs, and Sheets.
In our tests, comparisons between Bard, Bing Chat, and ChatGPT reveal that Bard and Bing Chat offer similar experiences, while ChatGPT seems to stand out as a major improvement thanks to the deployment of the GPT-4 model. Some users have observed that Bard feels like "GPT-3.5 with a few links," suggesting that it may be somewhat outdated or restricted compared to ChatGPT.
Overall, while Google is attempting to establish itself in the AI chatbot market with Bard, it appears to be struggling to keep pace with the rapid innovations and capabilities of competitors like ChatGPT and Bing Chat. Despite this, it is important for businesses and marketers to keep an eye on Bard and other emerging chatbot technologies, as they may offer unique features or potential benefits that could enhance their content generation, brainstorming, and overall productivity.
AI gets creative
The recent launch of new AI tools for creative work has made waves in the marketing industry this week. Canva, a popular graphic design platform, announced ten new features at their annual conference, which included Magic Design, Magic Draw, Translate, Magic Eraser, Magic Edit, Beat Sync, Presentation Generator, Magic Write, Text to Image, and Create Animation. These features leverage AI to improve content generation, design, and overall productivity for marketers, enabling them to create visually stunning work quickly and efficiently.
Adobe Firefly, another AI-powered creative tool, offers image generation from prompts, AI-drive vector re-colouration, context-aware brushes, and a range of impressive smart-text effects. Our impression is that the inputs have the potential to be "publication ready" with minor or even zero tweaking, something that hasn't necessarily been true of AI-driven image tools to date (although this is changing all the time with the release of software updates like Midjourney v5 and the models available using applications like DreamStudio).
AI-driven tools like Canva and Adobe Firefly certainly have the power to revolutionise the creative landscape, making it easier for marketers and non-designers to create high-quality content quickly and efficiently. They also offer benefits like streamlined workflows and the ability to easily update content with a click of the button (e.g. add a new logo, change the colour palette etc.).
However, we have some concerns that the widespread use of these AI tools might lead to a "templated world", where designs become too similar and lack uniqueness. Therefore, it's crucial to strike a balance between using these AI tools and maintaining a human touch to ensure that your brand remains unique, memorable, and engaging (and there will always be a need for human creativity and expertise to deliver innovative ideas and refined execution).
Recent papers released on our route to AGI and the impact of AI on jobs
Towards the end of the podcast, Paul and Martin's attention turned to recent research papers on AI: One paper from Microsoft researchers explores the capabilities of GPT-4 (and whether it is a major step towards AGI), while another from OpenAI focuses on how GPTs may affect the workforce.
The OpenAI research paper, on the other hand, explores the potential impact of generative pre-trained transformers (GPTs) on the labour market, specifically in the US. According to the study, 80% of the workforce could have 10% of their tasks impacted by GPTs, while 19% of workers may see 50% of their tasks affected. Interestingly, higher-wage , "white collar" jobs, appear to be most affected.
The rapid advancement of AI models like GPT-4 has the potential to change the way we work, increase productivity, and displace certain jobs (although, this is just theoretical, at this point). While AI could bring about improvements and efficiencies across various industries, it's essential to consider the potential social and economic repercussions. Policymakers, businesses, and individuals will need to adapt to this new landscape, ensuring that the benefits of AI advancements are distributed equitably and that workers have the opportunity to acquire new skills and transition to emerging job roles.
Tool of the week: AskYourPDF and ChatShape
Next, Paul and Martin discuss two tools of the week: AskYourPDF and ChatShape. Both are ChatGPT-enabled chatbots that allow users to query information from web pages and PDFs. AskYourPDF allows users to upload a PDF, such as an eBook or a peer-reviewed paper, and receive a brief summary of the content. Users can also ask questions about the content, and the tool provides answers based on the uploaded PDF.
ChatShape is a Chrome extension that reads data on a web page and allows users to query it with natural language through a chat interface. This means users can ask questions about content on a webpage, such as a Wikipedia article, without having to read the entire text. We're excited about the future development of ChatShape, which will eventually allow users to access all content on a given website (e.g. an entire blog series or resource section of a website.
Both AskYourPDF and ChatShape are available to try for free, although users should be cautious about uploading sensitive information to AskYourPDF and understand the access permissions you are granting to the ChatShape Chrome extension.
Showcase segment: Ninja marketing and sales application of the week
At the end of the podcast, Paul and Martin describe a story they saw on Twitter explaining how one enterprising developer built a GPT-4-driven app to enable him to send highly customised prospecting emails that automatically generated new web app ideas to pitch to potential clients with a very high success (sock-cess?!) rate. The only way to do this story justice is to embed the Twitter thread below, so we have - please be aware, this is content from outside of BioStrata and there is explicit language below.
My buddy @umar482 runs software development shop in Pakistan— Spencer Scott (@AKASpencerScott) March 23, 2023
(He is the guy I hired for @Median)
Like every business owner, he wants more customers
Historically, some of his best customers are Shopify owners that he built custom apps for
This is where it gets crazy
How to subscribe to the Artificially Intelligent Marketing Podcast
Hopefully you found this blog post useful... but it only touches the surface of what Paul and Martin cover in the podcast. You can listen to the full episode for free here (or search for it using your favourite podcast app). You can also access the back catalogue of episodes here.
The creation of this blog post was supported using Whisper API to transcribe the original podcast audio, and GPT-4 (ChatGPT) to summarise the transcript into blog sections. These sections were then manually curated and edited by one of the human members of the BioStrata team. The blog image was created with the help of AI using DreamStudio.