Big Plays or Little Plays: The Key to AI Revenue
July 11, 2024
I keep thinking about the billions and trillions of dollars required to create a big AI win. A couple of snappy investment banks have edged toward the idea that AI might not pay off with tsunamis of money right away. The fix is to become brokers for GPU cycles or “humble brags” about how more money is needed to fund the next big thing in what venture people want to be the next big thing. Yep, AI: A couple of winners and the rest are losers at least in terms of the pay off scale whacked around like a hapless squash ball at the New York Athletic Club.
However, a radical idea struck me as I read a report from the news service that oozes “trust.” The Reuters’ story is “China Leads the World in Adoption of Generative AI Survey Shows.” Do I trust surveys? Not really. Do I trust trusted “real” news outfits? Nope, not really. But the write up includes an interesting statement, and the report sparked what is for me a new idea.
First, here’s the passage I circled:
“Enterprise adoption of generative AI in China is expected to accelerate as a price war is likely to further reduce the cost of large language model services for businesses. The SAS report also said China led the world in continuous automated monitoring (CAM), which it described as “a controversial but widely-deployed use case for generative AI tools”.”
I interpreted this to mean:
- Small and big uses of AI in somewhat mundane tasks
- Lots of small uses with more big outfits getting with the AI program
- AI allows nifty monitoring which is going to catch the attention of some Chinese government officials who may be able to repurpose these focused applications of smart software
With models available as open source like the nifty Meta Facebook Zuck concoction, big technology is available. Furthermore the idea of applying smart software to small problems makes sense. The approach avoids the Godzilla lumbering associated with some outfits and, second, fast iteration with fast failures provides useful factoids for other developers.
The “real” news report does not provide numbers or much in the way of analysis. I think the idea of small-scale applications does not make sense when one is eating fancy food at a smart software briefing in mid town Manhattan. Small is not going to generate that. big wave of money from AI. The money is needed to raise more money.
My thought is that the Chinese approach has value because it is surfing on open source and some proprietary information known to Chinese companies solving or trying to solve a narrow problem. Also, the crazy pace of try-fail, try-fail enables acceleration of what works. Failures translate to lessons about what lousy path to follow.
Therefore, my reaction to the “real” news about the survey is that China may be in a position to do better, faster, and cheaper AI applications that the Godzilla outfits. The chase for big money exists, but in the US without big money, who cares? In China, big money may not be as large as the pile of cash some VCs and entrepreneurs argue is absolutely necessary.
So what? The “let many flowers bloom” idea applies to AI. That’s a strength possibly neither appreciated or desired by the US AI crowd. Combined with China’s patent surge, my new thought translates to “oh, oh.”
Stephen E Arnold, July 11, 2024
Comments
Got something to say?
-
- Subscribe to Beyond Search
Feature archive
News archive
-
