Ignite 2023 - Recap and Reflection

It’s been a couple weeks since Ignite wrapped up, I missed the opportunity to post that evening and between travel, life, and thanksgiving time has flown by. I have had plenty of time to process the conference so far and a number of conversations with people about how the event went. I figured I would just do a final summary of the event and my thoughts.

Microsoft’s Big Bet #

There have been a number of breakthroughs in my life that really changed the way the entire world interacted with technology. Things like smart phones, high speed internet, virtual machines, and cloud computing just to name a few off the top of my head. There are always those people or companies that recognize the potential impact in the beginning and jump on board, and usually it takes a few years for the rest of the world to catch up.

Microsoft is not hiding the fact that they think of AI as the next big step forward in technology, and they have pivoted 100% to becoming one of the leaders in this space. Attending Ignite this year left me with the strong impression that they think this is an advancement that is going to be something used by every consumer on each personal device, and every business in every critical application or interface. They have thrown way more resources than I would have imagined towards being the leaders in this space, and looking back on what I saw at the conference, it is really incredible what they have accomplished already in such a short time.

In conversations with other individuals about generative AI like ChatGPT, Google Bard, or Bing Chat (ahem, I mean Copilot), there are a lot of mixed reactions. Most people have some degree of uncertainty, anger or fear about it. There are concerns about it replacing jobs, copyright infringement, becoming a crutch for students, or even rising up in a terminator-esque revolt. But on the reverse, a lot of smart people are finding ways to use it really efficiently, cutting out a large portion of grunt work and allowing them instead to focus on refinement or clarifying an idea or message.

On the business side of things, I know there are a lot of companies with similar concerns, with the added anxiety of using AI and inadvertently putting their sensitive data at risk. There’s also a general sentiment I’ve heard amongst people working in the technology space that it is somewhat useful for basic scripting, but not much beyond that. All in all, AI has felt like this tech that people are fascinated with but very unsure of.

Microsoft is betting on the fact that they are seeing essentially limitless potential in this fascinating technology, and within a couple year everyone else in the planet is going to be going all in on the foundations that they are building today with OpenAI. The crazy thing is, they are already starting to prove it and everyone else is probably going to catch up a lot quicker than most realize.

My Thoughts #

Up until, during, and even for a few days after Ignite, I had a low opinion of AI in it’s current state. I also did not share many of the concerns expressed by other people, I was mostly apathetic towards it. The little bit I experimented with ChatGPT left me underwhelmed and I had seen enough direct instances of AI “hallucinations” (answers or responses that were just wrong) to shelf my interest with the intention of exploring it more later after it had advanced.

However, I enjoy asking other people how they were using AI products out of curiosity, and in the hopes that I could get some inspiration for how to use it myself. In most cases it seems to be a tool that was great for drafting the body of an email, letter, speech, or some other language-oriented product. From there people said they would tweak what was given to them with a few additional prompts and it would be able to assist them with delivering a final product while saving a good portion of time. My favorite example of this was a friend of mine who said he mostly used it to write letters of recommendations for his staff.

After thinking about what Microsoft was presenting, they really seem to have taken this use case and then exploded it to apply to as many scenarios as they can. They have Copilot for Github that assists with developing applications or scripts, which was already a pretty common use case. They also have Azure Copilot which can interact with, analyze, or manage any resource tied to the Azure Resource Manager data plane, which allows it to be a really powerful user interface for a complicated Azure platform.

But then they also have built Copilot for 365, which can take a static pdf report and generate an entire powerpoint presentation for you, or it can scan your email responses and then generate emails for you using your tone of voice or communication style. It can build excel forumlas, pivot tables, or integrations for you. Or it can generate a list of important tasks for you based on meeting notes or recent emails. All of these things kind of feel like a natural progression for AI but what really makes them brilliant is the fact that the AI itself appears to be able to interact with any office application or data in almost any form, and when you think about how that works it’s really kind of mind-boggling.

I went to a presentation about the 365 Copilot where the development team was talking about how they made the product work. They talked about how an AI bot can’t reliably use something like Visual Basic to do everything they need it to do, so they had to essentially build a lightweight, AI-oriented language on top of office to allow it to manage those applications to do everything they need to do. And then they had to tailor it to accurately interpret these sort of 365 requests and perform actions unique to that technology.

The more I thought about what I saw, the more interested I am in trying these products for myself, and teaching myself how to use them. And in every keynote, in every statement about the trials they have done so far at other companies, the statistics seem to support the idea that people really like these products. In almost every instance, they talked about a 60-70% decrease in mundane tasks, a ~50% increase in the ability to focus on more meaningful work, and a large percentage of users saying they wouldn’t want to lose the ability to use the product. These stats were consistent across every Copilot product they mentioned.

Consider me a convert, as I am looking for ways to take better advantage of these products in the future. I think Microsoft has made a good bet, and I don’t think it will take long for the rest of the world to catch up.