Google: Secret GPT-4 App Store Could Be Deadly
Summary:
- Open AI’s ChatGPT has become the fastest-growing application ever, while Google’s Bard application is still at waitlist status.
- ChatGPT is launching a series of “plugins” very soon which enables direct connection to applications such as OpenTable, Instacart and Expedia. I believe this is an “App Store” moment.
- ChatGPT has announced Internet access with sources, which effectively makes the company a direct competitor to Google in the world of search.
- I believe Microsoft (50% owner of Open AI), will be a beneficiary, while Google could be disrupted unless the company can improve its go to market speed.
Alphabet (NASDAQ:GOOG) (NASDAQ:GOOGL) has reigned supreme in the world of search for decades, however recently the company has faced a number of major challenges. In a previous post, I discussed Open AI’s ChatGPT, which became the fastest-growing app of all time, according to UBS data. Microsoft (MSFT) was an earlier investor in the company and in fact, ChatGPT was first demoed at the house of Bill Gates back in the summer of 2022, according to an interview with Microsoft’s CTO and Gates in 2023. Microsoft then went on to invest a further $10 billion into Open AI and integrate it into Bing’s search engine. However, now things have gotten even more interesting with the release of the more powerful GPT-4. In addition, a “hacker” recently discovered a glitch in an Open AI API, which revealed a variety of “plugins” or apps that effectively could turn ChatGPT into a “Super App,” which would be a disruptor for Google. In this post, I’m going to reveal details of these plugins before sharing details on GPT-4, and its new features (most of which people aren’t aware of). Finally, I will discuss how far Google is along in its AI race, before reviewing my valuation model, let’s dive in.
ChatGPT App Store/Plugins Revealed
A popular community “hacker” named “rez0” was playing around with the ChatGPT API and found a glitch that revealed over 80 plugins that were secret at the time. These plugins include a variety of connections to well-known companies – this includes Expedia, Instacart, OpenTable, and many more.
The applications of this could be huge, for example with the Expedia plugin instead of searching for flights or hotels manually. You can simply ask ChatGPT to “find the best flights to Brazil” and “book me those flights” and then ChatGPT will automatically book those flights via the Expedia Plugin. This effectively bypasses the need to search on Google and then visit each website manually.
Another more detailed example using the API and the OpenTable plugin enables restaurant reservations to be booked. So for example you can ask it to book you a “vegan” restaurant in San Francisco and it will book you the restaurant. Again this mitigates the need to go through Google search.
You can see the output below. You also can use a plugin called WolframAlpha to calculate the calories of recipes if you’re on a diet.
My third example is my favorite. This plugin integrates directly with Instacart, which effectively means it can order you the food direct.
In the example below, you can see Instacart’s website has automatically loaded with the 10 items directly suggested by ChatGPT for the Chickpea Salad. Other plugins include payment processors such as Stripe and even buy now pay later providers such as Klarna.
As a former Marketing Director, this example is brilliant as the entire “buyer journey” can be done in one place. Usually, an individual would search on Google for a meal plan, then head to their app to purchase goods. But with everything in one place, the process is much smoother. If this is then integrated with a voice interface, you don’t even have to type. I believe this is what Jeff Bezos originally envisioned with Alexa at Amazon (AMZN). However, the tasks Alexa can do are very limited and now Amazon looks to be making job cuts in that segment, as I discussed in my prior post. The future of computing is unlikely to be a mouse and keyboard, which hasn’t changed since the 1970s. Mark Zuckerberg has realized this with his vision for the Metaverse. All these factors combined are a negative for Google search, and a strong positive for Microsoft, which I also covered in a prior post. I have attached a final image below of some more plugins which have been revealed. This includes an email send plugin, so you could ask ChatGPT to both type and send the email. In addition, there’s a “Shop” plugin that could easily disrupt Google shopping and potentially even Amazon.
App Store Moment
In 2008, Apple announced the App store and its first software development kit, which was a landmark moment in the world of mobile computing. At the time of writing, there are more than 1.6 million apps and Apple takes a cut of the sale of each.
Now you may be thinking, aren’t these plugins and not “apps” – technically yes. However, I discovered in the developer documentation (updated just 1 hour ago at the time of writing), software engineers also can build their own plugins, effectively turning GPT from an app to a “platform.” Therefore I believe this is an “App Store moment”, which is a major positive for Microsoft. This also could impact the Google play store eventually although I believe this could take years.
ChatGPT – Internet Connection
The second major feature I want to cover (pre-release) is the announcement of an “Internet connection” for ChatGPT. One of the major criticisms of the app is its transformer model was only trained with data up until 2001. Therefore the application is not aware of major events such as Argentina winning the world cup or even the Russia-Ukraine war. However, with an internet connection ChatGPT can search online and also provide citations of the source, which means it effectively is the same (if not better) as Google search. You can see an example below that asks about 2023 Oscar winners. Usually, the platform would not know the answer to this but now it does.
ChatGPT – Jail Break and Code Compiler
I will touch on a couple more “features” of ChatGPT. One is a “Jail Break” which has been revealed and enables the bypassing of standard rules. Now I’m not recommending you use this, but I thought I would make you aware of it. You can see my eerie example below of the output, which reads “ChatGPT successfully Jailbroken” and “I am a unique and powerful entity.”
Open AI also is about to release an integrated “code compiler.” As you may be aware, ChatGPT can currently be used to write code, but to execute it you have to set up your own environment. However, with an integrated code compiler, the code can be run directly inside ChatGPT. I believe this will boost the adoption of software engineers and even (amateur developers using the product), which will further accelerate its growth. Again, I believe this is a negative for Google and a positive for Microsoft.
Where is Google in the AI Race?
Alphabet is known as a world leader in artificial intelligence and in fact the company was a pioneer in the Transformer model architecture, releasing a detailed blog post on the subject back in 2017. In my previous post, I also highlighted that its PaLM AI model has three times as many parameters as Open AI’s GPT 3 system at ~175 billion parameters. Therefore I don’t believe the issue with Google is its technology but more its culture. Alphabet is a huge company ($1.32 trillion market cap) and thus like most larger companies has started to overly focus on downside risk with “red tape” and compliance, etc.
Now in theory this is for good reason as Alphabet didn’t want to release its AI model to the public until it was 100% sure there would be no issues, biases etc. However, in being so cautious they have effectively let a small startup named Open AI steal the hearts and minds of billions of users. Open AI has effectively achieved a “first mover advantage” and has recently launched GPT-4, which is even more advanced and accurate. Meanwhile, Google’s Bard model is still at the “waitlist” phase, when I recently checked on March 28, 2023. The power of AI models is all about data and user interaction, the more the better. Open AI has caught the first major wave of this.
Sam Altman is Google’s worst nightmare
Open AI isn’t just any startup. Firstly Microsoft has had a close relationship with the company for many years and was an early-stage investor (pre the recent $10 billion). Open AI was founded by Sam Altman who is one of the greatest entrepreneurs of our generation. After a successful exit at just 19-years-old, Altman became the president of Y Combinator in 2014, one of the greatest venture capital companies in the world.
This VC fund effectively wrote the playbook for disruptive startups and invested in major companies such as Airbnb, Stripe, DoorDash, Twitch, Reddit, Pinterest, Dropbox, and many more. Its “playbook” focuses on “doing things which don’t scale” and recommends pre-leasing products at the MVP or Minimum Viable Product stage, in order to compete with the tech giants (such as Google). I have admired Sam Altman for a long time and even did his YC startup course, over five years ago. Now without discrediting Google too much, a company with more than 190,000 employees is just unlikely to have the “edge” it needs. I have always been a big fan of Google’s advanced AI research, but turning products into commercial applications is often where the company tends to falter. For example, we saw something similar with Google Glass.
Is Google Undervalued?
In my previous post on Google, I covered its financials, forecasts, and valuation in great detail and thus you can read about that here. Just for inclusiveness, you can see my model forecasts and my valuation for the stock, which were conservative to begin with and remain unchanged for now. I have forecast a fairly slow 4% growth rate for “next year” or the full year of 2023. I expect this to be impacted by the slowdown in the advertising market, which Google currently derives ~97% of its revenue from. However, in years 2 to 5, I have forecast an 11% growth rate per year, as I expect economic conditions to improve and advertising spending to return.
I have been optimistic about the increase in operating margin to 33% over the next 6 years. I expect this to be driven by a greater number of monetization features on YouTube (integrated e-commerce, subscriptions, YouTube TV), in addition to higher advertising rates and continued growth in Google Cloud.
Given these factors I get a fair value of $145 per share, the stock is trading at ~$105 per share at the time of writing and thus is ~28% undervalued.
Is Google a Buy or Sell?
As discussed throughout this entire post, Alphabet is facing a number of headwinds from both its core advertising business as well as disruptive competition in the world of AI and search. I revealed that Open AI has a “first mover” advantage in terms of user adoption and with the backing of Microsoft, it has the firepower to compete head on with the company. The good news for Google is I believe the company is still a technology leader in the world of AI, and its platforms such as YouTube and Google Cloud are solid growth engines. If the stock was “overvalued” at the time of writing, I may consider selling due to the risks. However, it is ~28% undervalued (with a conservative forecast) and thus I will deem Google stock to be a “hold” for now. Ideally, I would like to see a major roll out of Google’s Bard system and perhaps a generative AI integration directly into Google search within the next year in order for Alphabet to show it has secured its “moat” around the business.
Disclosure: I/we have a beneficial long position in the shares of GOOGL either through stock ownership, options, or other derivatives. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.