Remember Netscape Navigator? Yeah, me neither. It always appeared that standalone chatbots would meet the same fate, although for wholly different reasons and not because of just competition. It has been a needless new form factor that nobody had ordered.
ChatGPT and its likes, in its current form, might be in line for the vintage software museum and quicker than Netscape, PalmPilot, or Blackberry. The Large Language Models (LLMs) can be approached for information from anywhere: let’s term it “Query”ing. We are moving from simply "finding" and “searching” as we now ask questions, have conversations, and expect fully worded answers.
Chatboxes were the way to go, at first. But the velvet rope is now at the entrance of apps with integrated LLM capabilities. Microsoft added the Query through Copilot early on and now it is not only in Google’s Search but also in WhatsApp and Instagram. The need to open a new tab to ask is reducing rapidly.
There are two interesting, issues as a result. Few were able to price for find or search since their inception decades ago, and it will be interesting to see if the same holds for Query. Adobe announced additional pricing to query the PDFs this week. Microsoft has been attempting the same for a few months, and from the lack of noise around Copilot, one wonders how successful it has been. Google is contemplating pricing for enhanced search, as well. Given the free products on offer and far more featured Query about to enhance Querying of multiple documents through operating systems or mail applications, one wonders whether any of these attempts will lead to meaningful revenues.
The bigger issue is for the LLM makers, as most of them are being consigned to look for the best revenue opportunities as picks and shovels or a backend of someone else’s products. Effectively, they have to become B2B and not B2C. For some LLM providers at least. Trading the starring roles for a backstage pass is not just a valuation issue but also enormously difficult, given the pace at which new LLMs keep coming up at the most competitive end. So far, being around for a while in LLM space provides little competitive gains: in the last month, three new LLMs claim to be as best as anything around.
The disruption risk exists for most monoliner software businesses beyond “Query” as we have discussed multiple times because of the software commoditization. Having an existing business, and one with un-disrupt-able moat is becoming critical for a host of players.
