>Generative AI is great, but it's not like you could just assign an arbitrary job to a present-day LLM, give it access to an expense account, and check in quarterly with reasonable expectations of useful progress.
I'd be interested in seeing the results if so. Like someone with $500k to burn just handing it to an AI and telling it to found a startup, including incorporation, vibe coding and deploying a product, publishing a website, bringing on cofounders and/or advisors and/or employees as needed, fundraising, etc.
My guess is finding agentic tooling that could even push the model hard enough to attempt that in earnest would be a problem, and the compute bills would add up quickly, but it would be a very interesting experiment. Probably with hilarious results, but seeing what the actual failure points are would be useful information. Maybe it would actually end up hiring and paying someone to do something insanely illegal.
There have been a few attempts at “building a business by following ChatGPT’s instructions.” All the ones I’m aware of failed, and were really just a scheme to grab eyeballs.
There’s also a website showing a company’s attempt to use multiple LLMs to close its (real-world, but historical) accounting month-end books. They fail miserably with compounding errors.
Has anyone tried this yet?