Innovation, leverage, commoditize gameplay on LLMs providers

Innovation, leverage, commoditize is a gameplay from #wardleymaps

I’m not the first one thinking that #LLM owners are using it as gameplay (OpenAI, Google, Microsoft…).

But why?

Right now, in May 2023 we don’t know what are the user needs in that space. I mean the user needs with details and strong volume of usage (not one user but a great volume of users).

In addition to that, the more interesting things to be built on LLMs are not the user needs of the user of Today, but the user needs of the user of tomorrow (let’s say 2030).

By that reason to hear what is the real use requested to ChatGPT, bard and others is important.

How?

The ILC gameplay contains three main basic parts:

  • Innovate: Let others to build specific solutions (innovate) and delegating the risk of identifying mainstream uses to others.
  • Leverage: Identify patterns with metadata, so in this way we can identify what the end user is really requesting.
  • Commoditize LLMs (focus on specific use cases, increase efficiency of operations, increase the performance of right answers…).

Where is the main focus areas?

The main flows of data are marked in blue (map below), which are very imprecise It’s relevant to understand the uses of LLMs to make the right bets (right now too soon to know strong uses on vertical applications)

That information is gold!

Other gameplays

Right now I see other gameplays being used:

  • Education: they all are trying to increase the base of users by educating them. They are also educating governments (For instance this week Sam Altman is traveling through Europe explaining many prime ministers what they are doing, and some weeks ago main CEOs were in the white house).
  • Market enablement: mainly offering their solutions through an API with tokens (which enable to generate some revenue).
  • Open approaches: some companies are opening part of their code to enable the acceleration of the proper code and increase the number of specialists being able to build solutions on that space (right now there’s a lot of talent scarcity in that space).
  • Talent Raid: Removing core talent from a competitor is a classical and I imagine that behind the scenes companies are shielding their key resources, while others are truing to hunt them.
  • All gameplays related to ecosystem, where Sensing engines is key part of it,
  • create their own Applications? maybe this is something they are going to do once they recognize strong patterns.
  • Embrace and extend: If new applications are build, to remove the competition is easy: they just increase the cost of the tokens and just remove them from the ecosystem.

Any thing I missed?

Update October / 2023

Update November / 2023

OpenAI announced on November 6th a set of news in the Dev Conference about many aspects of the company. One of them is the ability of adding “GPTs” or applications that can be consumed by users from a store. There are other relevant news, but this particular one has much effect on the ILC gameplay:

Big LLMs providers (November / 2023)

OpenAI is creating this new market space at solution level and they are looking for a massive and quick adoption of it in the next months, looking for higher volumes of consumption:

Big LLMs providers (November / 2023) showing capital flows

Some questions that I do not have answer right now

  • Is this a new type of solutions store?
  • How much volume will be required to have positive ROI?
  • How much time will take Microsoft and Google to enable a similar approach?
  • What about AWS?
  • Has Grok something to say? or it’s still too early

Update December/2023

Leave a Comment