In the rapidly evolving world of AI technology, challenges abound, from hardware dependencies to business model uncertainties and emerging regulatory concerns. The aftermath of the global pandemic has ushered in a chip shortage, amplifying the dominance of certain industry leaders and posing significant challenges for AI development. As we venture deeper into the AI ecosystem, the road to profitability remains murky for many, with the fundamental question being not just about how AI evolves, but how it integrates into our economy, society, and regulatory frameworks. Let’s delve into some of the pressing issues that surround the AI industry today, from chip monopolies to the intricacies of monetization and the overarching umbrella of trust and safety.
Chips: It’s no secret that we are still reeling as an industry from the chip shortage that was caused in part by COVID. In the context of generative AI, this issue is exasperated by the fact that one vendor NVIDIA is the far and away leader in the space; basically, if you want to run large language models at any reasonable level of performance, you’re probably going to be in the market for NVIDIA’s highest end GPUs. As the global AI arms race continues this issue is unlikely to resolve itself anytime soon, however, AMD (NVIDIA’s chief rival) has committed to catching up via a number of interesting AI-focused acquisitions and research projects. Ultimately, it will take time, but most likely NVIDIA will continue to be the market leader in chips that power AI.
Immature Monetization: For startups and businesses venturing into AI, there’s not always a clear path to monetization, making profitability a challenge. The reality is that a lot folks have the idea that AI itself will be the end product; think Iron Man’s Jarvis or the heart-breaking little robot boy in 2001 film AI. I think those cases are going to be extremely rare. My bet is this going to be akin to the gold-rush when the real money was made by the likes of Levis and whoever made pickaxes. That doesn’t mean we won’t have some extremely interesting personified AI-powered bots, but it does mean they will be rare and likely passing fads; anyone else got Beanie Babies still? No, the opportunity is not in building the next Jarvis, but in building AI tools for other developers to bake into their projects. These developers would be focused on a narrow problem set and be looking to fine-tune another firms LLM to hone in our their and their customers’ use-cases. If this business model sounds familiar to you, it should, because it’s the model of OpenAI who (via their ChatGPT LLM) is the current market leader in AI tech.
Trust & Safety Issues: While I generally believe that on the whole AI and frankly most technological advancements are a net good for most people, a number of governments, organizations and individuals have raised concerns about the technology that you could broadly fit under the umbrella of trust and safety issues. These conversations need to happen, but this is not the appropriate venue for it. Nevertheless, there’s a real risk of regulation and other sanctions that could potentially raise compliance costs dramatically, thus eroding profitability; I’d be remiss if I didn’t mention that these costs would be most impactful for small and startup businesses that don’t have the resources and compliance know-how to navigate multiple bureaucracies.
So, yes, there are challenges, but there’s a metric ton of opportunity in AI here to create innovative and profitable new businesses. I’d be remiss if I didn’t mention that we at The Mad Botter INC have are a customer of OpenAI and have an ChatGPT integration with our Alice automation bot; if you’re interested in automating part of your business, drop me a line! Looking forward to your feedback on LinkedIN!
