> In reality, they're simply relying on several Playwright automation scripts to do the job for you, which is why they only support four apps: Spotify, Midjourney, Doordash, and UberEats.
I think that part is mostly fine? I'd rather make give a LLM access to https://woob.tech to be my personal assistant while parsing 99% less noise, than have a LLM that parse and understand stupidly complicated web pages, and randomly fail at the task because the name of my doctor is bobby drop tables.
That being said, it can be interesting to use LLMs to assist creating woob plugins.
you can't automate playwright without a decision making component in front of it, they are definitely using a transformer there. one could train a llama and make it perform triggers to playwright automations. you can even get deep into transformer tokenization and create action tokens and a formal grammar for your generation, build a parser on top of your predict function and have a "lam" working. the fact that they use playwright does not imply it is not generative ai. i'd say it is really hard to do those actions without a transformer involved
Midjourney does not have a public API and I'm pretty sure that automating a Midjourney account is against the TOS, so I wouldn't expect that functionality to last long.
I think that part is mostly fine? I'd rather make give a LLM access to https://woob.tech to be my personal assistant while parsing 99% less noise, than have a LLM that parse and understand stupidly complicated web pages, and randomly fail at the task because the name of my doctor is bobby drop tables.
That being said, it can be interesting to use LLMs to assist creating woob plugins.