Andrew O'Harney

Very poorly. Neural networks in AI require specific, often opaque adjustments to perform tasks that the human brain handles more efficiently. AI models require significant computational resources and often fail to generalize beyond their training conditions, unlike the human brain's adaptable and energy-efficient system. Thus, while neural net based models are powerful tools for certain applications, they do not replicate the underlying principles or efficiencies of biological neural processes.

As prompt-based technologies become more prevalent, the organizations controlling the AI have significant influence over the range of acceptable outputs. This control can potentially restrict artists and other creatives, confining their expression to what is deemed suitable by those in charge of the AI tools. This isn't limited to visual arts; it affects writers, programmers, and others who rely on AI-enhanced tools, shifting the landscape towards more restricted creative possibilities. The ethical concern here is significant: it touches on censorship and the potential dilution of originality and innovation.

Future proof your business by considering that tomorrow's customers won't be people, but rather will be sentient AIs themselves. Begin with your revenue model: you might consider on accepting payment in cryptocurrency or AI-generated art, predicting these will be the gold standards of future economies. On market: bypass traditional industries and focus on niches yet to be created by the AI revolution.

If you're still pre product market fit, identify who benefits the most and double down there. Not everyone's going to be over the moon with what you've built, and that's fine. Focus on those who see the magic in your AI. Pour your energy into making the experience unforgettable for them. Tailor your AI, make it indispensable.

When using generative AI use specific prompts to counter bias. For instance, when generating images, include in the prompt that people from a diverse range of cultural backgrounds should be included. Be careful when doing this though - also ask that subject matter is taken into account and adhere to actual historical, cultural, or differential biases that truly exist.

AI or not, focus on traction first - not funding. In fact, never raise money unless you absolutely must. Even then, explore non-dilutive financing options before VC/angel. Get that MVP out there, get your first few customers or snag letters of intent. These days there are plenty of open source and accessible models that won't require a tonne of foundational research first. Real, in-the-wild use beats speculation any day. If that plan is not possible and you're building something rock solid at the core of AI— you may as well think super big, be prepared to show off your credentials and go all-in on the bet. Bottom line: Deliver value, not just ideas, and you're golden.