As you may have heard, Microsoft is reportedly investing a whopping $10 Billion in OpenAI, the firm behind ChatGPT and DALL-E. The eight-year-old firm with no mainstream product or revenue model yet has dominated the news headlines ever since it released an AI Chatbot to market in Nov 2022. So, we decided to check in on what the hoopla is all about.
So, what is ChatGPT and what can it do?
ChatGPT—the GPT stands for generative pre-trained transformer, is a specialized GPT-3 based natural language model, that can do many things from formulating answers and generating content to having a real conversation with you. This AI tool applies ML algorithms (reinforcement learning) optimized for human-like dialog scenarios. Think of it like a free chatbot that can assist you with tasks such as composing essays/blogs, writing code, generating art while answering any question you pose, and mixing and matching ideas for creative solutions.
ChatGPT is based off a generative AI algorithm called transformers (in this case a large language modeling Encoder-Decoder neural network). Until recently, machine learning was largely limited to classification and predictive modeling based on supervised and unsupervised learning. Generative AI was a breakthrough of sorts as these algorithms can create new content including text, audio, video, images, and even code.
So why is Microsoft so gung-ho about this?
To dig into this, I decided to look top-down from the story that the 2022 financial numbers point to.
The table below represents the Quarterly topline revenue numbers as reported by Microsoft in its earnings call across its 3 major segments.
|FY’ 23 Q2*
|FY’ 23 Q1*
|FY’ 22 Q4*
|Productivity and Business Processes
* In Billions
As you may have guessed, a few trends seem to emerge.
- The personal computers segment has seen a steady decline in growth
- The revenue from its cloud segment (Azure) forms the largest share and is the one showing massive signs of growth potential.
So now while Microsoft has a need to arrest its decline in personal computing and capitalize on growing its cloud and Productivity suite, as anyone doing a competitive study of the major cloud service providers such as AWS, Azure, GCP etc. may tell you, the competition in this aggressively growing market is so fierce to gain share that these providers achieve feature parity in a matter of weeks, if not months.
If Microsoft can manage to successfully integrate ChatGPT use cases with some of the huge Corporate/Consumer datasets that it houses as part of its productivity (such as LinkedIn, O365, Teams, etc.) and/or Azure suite (as it is hoping to), then this could be the silver bullet that can help revive Bing and catapult its cloud business growth multifold. This could provide the key differentiator for Microsoft to stay ahead of the game, by adding the missing link of feeding contextual data intelligence to enable hyper-personalized result sets from ChatGPT.
All that is fine and dandy, But what is it that this ChatGPT can do?
As it turns out quite a bit, actually. It has widespread applications ranging from mission critical workloads to generating jokes on the fly, from disrupting supply chains to providing recipes for butter chicken, from designing e-commerce websites and doing technical reviews, and storytelling to acting as a social media influencer or advertiser.
Like most AI models and tools out there, ChatGPT’s result set sees an exponential increase in relevancy when additional context is provided through contextual prompts. And this is where the wealth of Corporate/Consumer data that Microsoft holds can start yielding quick and rich dividends.
Sample use cases include rich UI-based interfacing services targeted towards:
- Personalized LinkedIn recruiter using your Corporate and LinkedIn profile
- Azure Services Cost optimizer virtual guide based on your subscription and usage history
- Azure troubleshooter to help RCA and suggest potential fixes based on log data
- Richer Employee services and assistance using the Active Directory personnel info for better employee engagement
- Active Email Responder learning your personal communication and preference style
- Enhanced Cognitive Search that can learn from your immediate past actions on what your exact need is and yield hyper-contextualized and personalized result sets
- Personalized Sales assistant who can enable SAs to have better prospect conversations by surfacing multi-hashed incentives/feeds etc. using personnel/Company/Industry/Macros economic indicators from CRM and other sources
- Developer assistant for providing advanced boilerplate code to truly enable kickstarter packages
This list is by no means even an attempt to look at how many ways this can disrupt each of the product/service offerings from Microsoft. As you may begin to notice, literally the sky is the limit and we have just begun scratching the surface. Even as we speak, this tool has already started disrupting the educational industry with colleges and schools now waking up to the fact that the students can just get their complete homework done from ChatGPT with more human-like responses than ever before and are now rushing to put compliance guidelines around this.
While the world is waiting to understand what the pros and cons of this new power entails and still others wait with bated breath on what this means for their career, we could all do well by better understanding and tackling risk and regulation before we start seeing further manifestations of this amazing technology going mainstream and us playing catchup like what happened with the Crypto world.
More to come from Bottle Rocket on this topic for sure.