@Daniel Tomberg not sure what made you think it’s a solo founder solution. Even the footer is rich in company information, not to mention terms page, GDPR, privacy, etc. everything connected to the limited company

@Shane Terry thanks, Shane. That’s the feedback I was expecting. To be honest, the whole website ( the marketing one) is mainly for Google and AI bots. The average trader Jack is not even reading anything past the 3 buttons on the front page, not to mention reading something on the other 17 pages of the website. Hence why that first section there, before any scroll, has a create account, watch video and live test the system. Pretty sure my traders will not go further whatsoever. That’s the part where they need to be convinced

P.S. my answer is specific to the way you phrased the question. If you only mention Claude, the assumption is you are referring to the Claude web interface, the chat. Cause if you are referring to Code, Cowork, Design, API usage, etc, discussion is completely different 😀

@James Day definitely using the API, mate. You can track usage there as well as tracking your usage natively using https://platform.claude.com/dashboard

Think of the use of AI in business terms first and foremost. RL requires compute power, best I’ve seen online cost a fortune, you most likely need tens of thousands of budget just for that. I doubt this is the route you want to pursue considering the other questions you asked. Use Karpathy method and create a framework of how to work with the AI prior to starting to think what project you want to build. If you are also asking how to prompt Claude, that would suggest the lack of a basic understanding of prompt engineering, in which case I would say better start with an organised way of learning the basics of that, in which case I strongly suggest to start here: https://www.anthropic.com/learn

@Pat Bishop i see no reason for an AI model to suggest a cheaper model. it's our responsibility to never start working without having a clear breakdown of costs, optimise usage based on task difficulty, and make sure we have a proper budget to complete any project. also, i've seen so many complex solutions built by people which failed miserably during the monetization stage. when they realized they cannot charge people less than the usage of the SaaS solutions.

@Jack Piro all good.

@Sgt Poulson I don’t think it’s just fear, mate. It’s also costs. Any individual with some ideas can start using the likes of OpenClaw, Hermes, or any similar. Cause they are outstanding problem solvers. But most of the people using them will enter an infinite loop in which they try to achieve something in impossible niches and they will just end up with huge costs and zero outcomes. On the other hand, using AI in established businesses and agents like the ones mentioned above, is a 10x productivity booster. Only problem is that it might come with less need for additional humans in the loop.

Here something I have created for my university foundation year, which has been noticed by a few companies in just a couple of days, and they started implementing it as a software. Because it’s model agnostic and it can be easily fit in different LLMs and not just that, but also helps both parties involved in demand and supply of workforce. And if you think of it, it’s nothing but a basic agent skill, it’s not rocket science: https://github.com/akeronforces2026/construction-job-hunter