The weekly download: No 31
The reality of AI and jobs, AI business models in China which confirm that SaaS is not dead plus some wider thoughts on why AI changes what we build, driven by a need for vastly more software...
Feels like a week where interesting stuff has been hitting me from all directions. For the purposes of this download, I have concentrated on a few links that have a direct connection to AI and business change. That means a handful of great articles are not in here - apologies to the authors, I did really enjoy your work.
No, AI will not replace jobs or people
Let’s start with a short but clear “myth” buster. In Finally we have created the silver bullet Dan Davies has set out the reasons why AI will result in wholesale unemployment and the end of work as we know it. In his inimitable style, he makes both the theoretical and practical arguments much more clearly than I could.
I have put myth inverted commas here because I don’t think the alternative is solid enough for even fantasy status. AI will not replace people. If you are thinking about the effect of AI on your business, the idea of “How many jobs can I eliminate?” will get you to the wrong answers very quickly.
AI will create new business models, new products and services and whole new categories of both jobs and organisations. None of this is certain but we can see some signs of where the early leaders are headed if we look in the right places.
A different world, the AI business in China
Grace Shao focuses on China and has a great eye for the way the AI business is developing over there. Part II: How to understand China’s LLMs as a business as the title implies is one of a series. By all means go ahead and read all three but this is the piece I found most interesting.
As Grace points out, the business environment in China is different so what works there will not be exactly a copy and paste into Western markets. Nonetheless, I think her descriptions of emerging business models have something to teach us. Here is one example:
“The same forces that make diffusion fast also make margins thin: competition that races to zero, big platforms that bundle and subsidize, and an ecosystem where distribution matters as much as the model itself.”
She is also clear that this will be a difficult market for US companies (and european - or it would be if there were any significant european companies).
“I don’t think American AI companies can treat China as a straightforward growth market in the mainstream sense, meaning mass consumer distribution and broad enterprise workflow capture. Not because the products are not good enough, but because the ecosystem pushes pricing toward zero and distribution toward local incumbents, on top of regulatory constraints that are obvious..”
I think this really matters because the regulatory/ governmental aspect is an afterthought. It is driven by the characteristics of the market. Here in Europe and the UK, Governments seem to think they can legislate and subsidise an AI industry into existence. Follow the market, not the regulators.
There is lots more in the article. I will pick one final quote which is directly relevant to SaaS:
“The second point that came up repeatedly is that, long term, a subscription model makes the most sense for LLMs. It’s a Spotify-like model. Instead of paying for tokens, most people use only a small amount, just as they listen to only a small subset of songs. A flat fee aligns incentives and creates the potential for durable margins.”
Been true for at least 15 years and fascinating to see that some people believe it still holds.
Software and SaaS: We need more not less
Which leads me to Death of Software. Nah. by Steven Sinofsky. This goes a couple of steps further than my own post on the Fall and rise of SaaS. He captures the key point very neatly early on:
“AI changes what we build and who builds it, but not how much needs to be built. We need vastly more software, not less.”
He is also excellent on the history of technology shifts. Working through three examples which each show two things:
How completely wrong the expert predictions of the demise of certain technologies were.
How new and unexpected models and value emerged from each major shift.
I also like that he acknowledges that both the erroneous forecasters and the prophets of doom have a role to play in shaping how the world reacts to new technologies.
Does not undermine his basic point. Listen to experts by all means but retain the flexibility to learn and evolve. Because committing to some predicted future is certain to be the wrong option.
I especially liked his own “predictions” at the end. A nice way of laying out the foundations of an approach to this wave of uncertainty. Feels like a good way to finish this week:
“1. There will be more software than ever before.
2. AI-enabled or AI-centric software is simply moving up the stack of what a product is.
3. New tools will be created with AI that do new things.
4. Domain experience will be wildly more important than it is today because every domain will become vastly more sophisticated than it is now.
5. Finally, it is absolutely true that some companies will not make it.”
Thanks for reading
Good read as always, Kenny.