As we make it through the third week of the new year, one clear trend developing is the sudden onslaught of AI-generated everything.
The most popular of which, ChatGPT, is set to make its day-to-day life debut as OpenAI’s early investor Microsoft (MSFT) said at the World Economic Forum in Davos Switzerland that it would begin incorporating AI into all of its products. Further, the company will make its AI available as a platform for other businesses to build on.
More importantly, the company’s CEO said it will move quickly to commercialize OpenAI’s tools.
Background
ChatGPT is an AI chatbot that can write, at times, incredibly human-like responses to text prompts. This can be anything from writing Python code to passing an AP English essay.
While sites like Stack Overflow have temporarily banned AI-generated code suggestions, Russian hackers have been observed using ChatGPT to write malicious code. The nuance here is that sites like Stack Overflow want precise correct answers, while for Russian hackers and perhaps developers, the code can be close enough that a few tweaks can make it right.
What’s Next?
At the event in Davos, Microsft’s CEO, Satya Nadella said that “every product of Microsoft will have some of the same AI capabilities to completely transform the product”.
Microsoft has a ton of revenue streams, so conceivably, that is a lot of AI. So we’ll focus on some of the likelier near-term ones that aren’t support-related chatbots.
While search advertising only accounts for 6% of Microsoft’s revenue, it remains an area that could be high growth if they can bring anyone over from using Google search. Conversational AI could be better for some kinds of questions as it can simplify or expand its answer depending on your prompt and follow-up questions.
Video games, which account for 8% of revenues, is another area that AI could find itself in soon. This seems particularly near-term in sandbox games like Minecraft where people create their own servers and game modes. There are clear parameters that lead to near-limitless opportunities, and conversational AI could really speed up the process of making a new custom game mode.
Meanwhile, Windows accounts for a larger 12% of revenues. AI implementations could be used to speed up repetitive tasks by creating personal automation based on use. Further, it could simplify organizing and allow users to more easily change computer settings and manage applications, as well as offer a way to bypass a search engine for questions.
Lastly, office-related products like Teams and Microsoft Office Products like Excel account for 23% of revenue and could get a boost over competitors, which boomed in some areas during the pandemic, as AI integration could drastically simplify use and boost productivity.
All of these are very doable based on existing features in ChatGPT with creating spreadsheets, VR fitness games, and writing SQL queries in the current stack of capabilities.
Where Else?
Elsewhere, AI is being implemented in everyday life right under our noses. However, Microsoft’s implementations will be massive as they have a huge market share of devices and they will really bring it to the forefront.
McDonald’s (MCD) recently launched a fully automated store and had earlier been testing AI for taking orders during the pandemic.
Meanwhile, AI is increasingly being turned to in sports to help make decisions. This was on full display during the World Cup with semi-automated offsides technology as goals were ruled out (or in!) based on nanometers, and other sports are likely to follow.
Additionally, Tesla (TSLA) is supposedly set to make a pretty big transition when it eventually releases version 11 of its autonomous driving software. Cars will start to run on what Andrej Karpathy calls “software 2.0” as it uses neural nets for navigation and control, rather than just vision.
Needle Mover?
This is an interesting dilemma. On one hand, AI will likely look something like the search engine business with one main winner as the best AI will get built into most things. Second-bests could serve niches, but since scale helps AI learn faster, the definition of "niche" will likely get more narrow. On the other hand, it could become the latest over-hyped tech advancement. AI has technically been that for a while, but this is more generalized than Tesla and others' past (and present) over-hyping so it kind of deserves its own chance.
I think investors can call it a needle mover after implementation starts and it seems to be a big improvement on Microsoft and Microsoft-related devices and services, or if we see a roadmap of upcoming improvements and they impress.
At its best, modern AI efforts can make products drastically better. However, it remains to be seen if today’s impressive AI models are ready for everyday use. We’ll get an idea pretty quickly based on the complexity of improvements launched whether this will be the next infamously delayed tech promise or if it can transform and simplify day-to-day life.
