furilo

Beyond what can be done right now with AI (or LLMs), I like to imagine what will happen in 1, 5 years from now... Some "imaginings":

- It's going to be ubiquitous, it's going to be everywhere and accessing it won't be a concern: just like now we take it for granted that we have electricity and internet connection at any time and almost anywhere at a negligible cost, the same will happen with AI. This is already happening in part because accessing the APIs is very cheap and effective, but it's going to go much further.

- It's going to be invisible: we're going to have models embedded in software and devices (it's already happening: Chrome will embed a model in the browser itself and there will be a native API to use it). If this extends, many software functions will be using AI without us having to deal with it.

- It's going to be private: being local, we won't have to worry about it working on confidential data - the data won't have to go to an external API where we don't know what happens to it.

- It's going to be immediate: the generation speed will continue to increase. There will come a time where generations will just be there.

- Chatbots in vertical applications will be an anecdote: many of us are implementing them because it's easy (to implement and to think, or not to think, because we've seen the success of ChatGPT and we've followed it). But for many use cases it's not the best interface.

- End users won't use AI: in the sense of using prompts, configuring agents, etc. Only power users will do.

- Devs will use/"program" a lot of AI: to make it invisible by integrating it into existing software so that end users don't have to deal with using prompts etc.

- No/low-code will grow by x10/explode: what we've seen so far will be a joke compared to hooking up agents.

And some uncertainties I can't imagine how they'll evolve:

- Generative interfaces: AI will build for us in real time the interface we need.

- Integration of agents: the theory about agents is clear and there is already a lot of software, but I can't see how it will spread.

Open source LLMs are amazing, they will get better and maybe even better that OpenAI's. But: using OpenAI APIs is so easy (they just work, are cheap, they will get cheaper) that added to all the hype, many of us just start using them. We know there are other options, but since this thing we are doing is just a test, we'll have time to change the future since changing is so easy.

But wait: is changing just updating the URL to the API you call?

For many applications, you need to create the embeddings. Each new embedding you create with a given model is digging a little deeper on the vendor lock in...

And if going out of S3 is expensive, imagine having to re-embed all your content with another model/provider...

So, what are your thoughts about the Biggest Vendor Lock-in of the Decade?