Join Gen AI Enterprise Leaders in Boston on March 27th for an exclusive evening of conversations about networking, insights, and data integrity.request an invitation here.
“This is venture money, not adventure money.” This was the loving response I once got from a VC when a dear friend of mine pitched an idea. But when we are in it, stages of the hype cycle With the advent of new technology, that vigilance completely disappears. After all, VCs have to put all the money they raise, especially when everyone else is making the same swing, and the cost of missing out on something big is the minus of swinging and missing out. higher than the surface.
Similar dynamics play out within most companies, and the current technology is AI and everything remotely associated with it. large language model (LLM): They are AI. Machine Learning (ML): It's AI. Call that project that you are told every year unfunded, call it AI, and try again.
Billions of dollars will be wasted on AI over the next decade. If that sounds like a backwards interpretation, it shouldn't be. Big waves of technology come with excitement, even before we know how real and transformative it will be. Search, social, and mobile have all had far-reaching and lasting impacts, but virtual reality (VR) and cryptocurrencies are even more restricted.
But you wouldn't know that just by reading headlines from five years ago. Everyone is now eager to show how much they are spending on AI and how it will change everything. This shotgun investment approach will inevitably result in some big hits and many failures. The same dynamic is at play in venture capital, where corporate leadership often greenlights investments that lead to optimistic, or at best misplaced, hopes and adventures in the name of AI.
It does not deviate from the fact that LLM is a revolutionary technology. See how quickly ChatGPT reached 100 million users compared to other innovative companies.
![](https://venturebeat.com/wp-content/uploads/2024/03/image1_46edd2.png?resize=680%2C382&strip=all)
Almost every company uses some kind of initiative LLMs and AI. So how do you decide where to bet and where you are entitled to win?
If you understand these three things clearly, you can cut down on 80% of your wasteful spending.
- Understand the total cost over time.
- Ask why others can't do it.
- Make some bets that you're willing to make.
1: Understand the total cost over time
When you decide to say „yes“ to it, Next AI project, look at the cost of resources needed now and in the long term to maintain that project. His 10 hours of work on the data science team often buries his five times as many hours in Engineering, DevOps, QA, Product, and SysOps. Companies are littered with fragments of projects that were once good ideas but lacked the sustained investment to sustain them. It's hard to say no to AI initiatives today, but saying yes too often often means you don't put enough money into some of the things that are worth supporting tomorrow.
Another aspect of cost is the marginal cost that AI increases. Training, running, and maintaining these large models is expensive. Overusing AI without a corresponding increase in downstream value will squeeze profits. Even worse, withdrawing released or promised features can lead to customer dissatisfaction and negative market perceptions, especially during hype cycles. Just look at how several failures, not to mention the early days of IBM's Watson, quickly tarnished Google's reputation as an AI leader.
2: Ask why others can't do this.
We quickly forget what we learn from textbooks. Everyone has read about commoditization. The same lessons I learned from being pushed around in real life stick with me. When I worked as a chip designer at Micron, our core product was a near-perfect product: memory chips. No one cares what brand of memory chips are in their laptop or how much they cost. In that world, the only long-term sustainable advantages are scale and cost.
The technology industry can be bimodal. There are exclusives and merchandise. If you say „yes“ to the following question, AI initiatives, ask yourself, “Why us?” Working on something that becomes commoditized over time is no fun, especially when there are no scale or cost advantages. Please take it from me. The only one that will definitely benefit is Nvidia and his AWS/Azure. The only way to avoid this is to focus on areas with defensive moats. Get a head start with priority access to data, unique insights into use cases, or applications with powerful network effects.
3: Make some bets with the intention of following through.
The simplest bets are those that will make the business you're already in better. BASF old commercial “We don't make what you buy, we make what you buy better.” The best bet is if applying AI can give momentum to the products you already make. Easy to run and extend. The second easiest bet is to move up or down the value chain or expand laterally into other sectors.
The most challenging but important bet requires cannibalizing your current business with new technology. Otherwise, someone else will be cannibalized. Be prepared to place more bets on these few bets that passed his two tests and follow through on those bets. Leave the rest to VCs and startups.
The hype around AI is real and justified, but if there's one lesson we've learned over the years, it's that these cycles involve not only healthy investments, but also a lot of waste. is. By following some of the tips outlined above, you can maximize your investment's chances of delivering algorithmic results.
Mehul Nagrani is Managing Director, North America. in moment.
data decision maker
Welcome to the VentureBeat community!
DataDecisionMakers is a place where experts, including technologists who work with data, can share data-related insights and innovations.
If you want to read about cutting-edge ideas, updates, best practices, and the future of data and data technology, join DataDecisionMakers.
You may also consider Submit an article It's your own!