The core limitation with current GenAI systems is they mandate human curation of complex formatted data production…
It’s a lived technical reality for a few, but not necessarily an obvious realization for many GenAI observers, that the current state of the technology limits the model’s interaction with data to complete interactions only. In simple terms, the models cannot currently interact outside of themselves with separate applications to affect pieces or parts of large data files or applications. A basic example would be the models could perform a grammar check on a document, but not automatically interact with MS Word to fix only the errors it identifies. You would either need to have the model attempt to rewrite the entire (or some whole section of) text to fix the errors (hoping to note induce more in the process), or you would have to manually fix the errors separately yourself. We saw this same thing with diffusion models that generate images until recently, when notable major diffusion model providers started adding additional tools to their user interfaces to allow image subsection alterations. Soon with agents we will see similar capabilities come to LLMs.
Agents will be the major capability to unlock a wide number of use cases, as it is the key to automating large workflows of complex actions. Currently there are several third-party software platforms that do a version of this automation. Make.com is perfect example of this type of customizable AI workflow automation with LLMs. Like a Zapier for AI, you can use these workflow tools with LLMs to build out pipelines of automation to construct very complex and parallelized outputs. The key piece though that is still limiting all these 3rd party platforms is price. All of them use the API interfaces of the major LLM providers, and those all charge on a per token basis. So, it’s pay as you go, and it’s easy to go very expensive very quick. With agent integration into primary model provider systems (like ChatGPT and others), will we see the continuance of flat rate monthly subscriptions? When you can do so much more how will the companies scope that cost to the user?
The next level of ground breaking capability might also just bring with it the next level of significant cost to the average user.
