OpenAI's ChatGPT Plugins and the emergence of an Apex Aggregator
I'd be really interested to learn about the implications for finance and investing? ChatGPT plug in seems very similar/analogous to "smart order routing", where Wall St. traders are able to leverage AI to find out where the best trading venue and supply is for their particular trade. In other words, it matches buyers and sellers, cutting out the middle man (broker).
There seems to be a lot of use cases for ChatGPT in finance..."Find me the best performing large cap growth mutual fund over the last 15 years"..."Put me in contact with three mortgage lenders offering the lowest rate currently"..."How many treasury futures do I need to buy to close my duration gap" (SVB could have used this).
Would love to hear your ideas on potential implications for finance/investing/wealth management. Thanks Packy!
Great read. But the idea of an Apex Aggregator is frightening. Agree that inserting themselves into money flow is a path but centralized control soon follows when someone inserts themselves in the money flow (ex, Apple's 30%?). I think OpenAI is on the path of running the Web2 growth playbook...shit, YC practically wrote it.
The real way Facebook and Google make money: surveillance capitalism.
So AI is just the latest? Good to know.
Wow, this was an excellent piece! I'd be interested in hearing your thoughts on the potential impact this may have on the content creation market, if any – such as writing newsletters, producing video content, and so on.
I was doing research on Ben Goertzel's SingularityNET project, and reading this piece sparked a thought. He envisions a system where AIs connect through a blockchain network and use APIs to collaborate with each other, assisting in solving problems users are addressing within the platform's ecosystem.
This plugin/API relationship involving apps connecting to ChatGPT seems somewhat similar, albeit with a different infrastructure.
Fantastic read, thank you!
I wonder if one day we could use ChatGPT to intermediate fights with our partners. You vent all your frustrations to ChatGPT, and it tells your partner, "John doesn't feel heard by you, and he knows you're tired when you get home from work, but he wishes you would answer, 'How was your day?' with more than just, 'Fine.'"
This is a great read, I'm as excited as you are about the implications of Plugins ChatGPT. Hopefully this whole thing will take a better turn than what happened with web2 companies and their siloed monopolies...
"If it comes down to a choice between plugin suppliers and users, I think OpenAI will side with the users"
This would be interesting to see if OpenAI will side with the users.
Wow, this was such an interesting read and really set my mind running in many different ways, I've enjoyed several of your posts but this one was in another league so I wanted to do the pledge thing... I know it says future rather than the money being taken now, but I would have liked to see a one off payment as an option if you would consider that? I've been trying to do the right thing for online content creators I love, but given the cost of living I worry about signing up to anything monthly/yearly at the moment but often really happy (especially after payday hehe) to fire over a one off thank you as and when something particularly resonates (a pay as you go option if you like) :)
This is great. I have three things in mind after reading this:
1. From a game-theory perspective, all it takes is one single company in a certain vertical to build their ChatGPT plugin, and the rest in the vertical will eventually have to follow suit just for the sake of keeping up with the competition, as you mentioned. Adoption seems inevitable.
2. With ChatGPT plugin, everyone will now have their own PA. I believe this is something unprecedented in history. There are many mundane decisions in our day-to-day lives that can be outsourced. Usually, I don’t really care what I should have for lunch, and I don’t want to spend 10 minutes opening the app and considering a bunch of options, which sometimes creates decision fatigue. In Japan, they have a type of dining called Omakase, in which you outsource the decision to the chef. There is no menu; the chef decides what you eat. Same thing here, I believe. These kinds of plugins can help us save a lot of time by outsourcing our judgment on low ROI decisions. I guess many will be willing to pay $10-20/month for this.
3. The human-focused user interface as we know it today in most consumer apps will become less relevant over time if more and more decisions are made via ChatGPT without the user browsing the app. I wonder about the long-term implications of this on designers and front-end engineers.
When we talk about differentiation in the context of increasing market competition, I do agree with you that it could lead to lower prices and a better world. However, when information itself is the product (such as with small websites and blogs) things could go wrong. Essentially, OpenAIs Browsing plugin grab all the data from these sites and give it away for free without leaving any value for the creators. If they cannot make money anymore, it may led to a less diverse and more concentrated system of information and thoughts.
Love this. Thanks for putting it together. Great read.
Great post! I'm just as thrilled as you about this unexpected development and can't wait to see how this will shake up various sectors. Regarding the view that in this new realm, the branding will matter less and becoming "the best" will be what businesses have to do to get picked up, I'm not so sure about that.
I agree about the branding part but plugins would have to lean on external information sources, like customer reviews or restaurant descriptions, to reach its conclusions at least in the beginning. Consequently, businesses will end up gaming the system with some kind of PEO (Plugin Engine Optimization) to gain ChatGPT's favor.
A potential avenue to a more objective assessment might involve ChatGPT gathering firsthand feedback from users post-experiences, employing the same simple thumbs up or down system. But even that could be gamed on the long run. Anyhow, fascinating times!
First time reader and what a piece to get introduced to Not Boring. OpenAI could take personalization to the extreme after aggregating plug-ins. It knows your food preferences, it'll know how you prefer to talk, it'll know everything because it's all stored on chat logs. Google knows your search results and clicks but ChatGPT will know what actions you take with the plug-ins. That said, no way Google will just continue to slow roll things - they'll release something in several weeks. (Though DeepMind scientists did sign a petition to stop further AI development in the indsutry).
Great read and agree with your general premise. One point I wanted to discuss:
“Brand matters less than before. Actual differentiation matters more…. there’s incentive for a restaurant to focus on making really fucking great French Onion Soup”
Whenever a platform has emerged that can supposedly distinguish the best French Onion Soup in a playing field of French Onion Soups, players have emerged to game that system. There are many factors involved in distinguishing good from mediocre; data on these factors will need to be delivered and that’s a vulnerability. The platform will only be able to distinguish as well as its data sources allow. There will be AI search optimization, experts will spring up, and some crappy products will be successfully promoted. As much as there will be a focus on "making really fucking great French Onion Soup" there will also be a focus on making what appears to be really fucking great French Onion Soup to an AI.
On the supply side, if AI enables a proliferation of subpar content or products, brand (or something like brand) will matter even more. When I buy things on Amazon, I often have to buy the same thing multiple times because the first thing will break or be defective in some way that was not obvious from the reviews or PDP. It’s not Amazon’s fault, but players will game their system.
It is true that AI will enable individuals to access more data than is easily accessible today, and that will generally lead to better decisions. But I think in that world, brand will continue to be a moat. It may even be more of a moat.
(It might be helpful to define brand, which is your logo and advertising but also your reputation as maintained through expensive signaling and over a long period of time.)
Platforms that lead to a proliferation of hard-to-differentiate entities can create more of a moat for brand. I don’t think that’s necessarily a bad thing. If you invest in brand, then you are likely investing in other aspects of your product as well. It’s not perfectly correlated but it is correlated.
Most consumers are susceptible to deception, and for businesses it will be well worth it (even necessary) to optimize and game these systems. As a consumer that just wants an actually good French Onion Soup, I will probably still leverage a variety of sources (including offline sources), or direct my AI at a subset of trusted sources (brand!), to make decisions.