This stuff, chatbots and generative AI, is not the artificial intelligence that actually matters. What matters is the AI that's behind the scenes.
From the Disconnect substack, July 27:
Generative AI will enrich investors and be deployed against everyone else
Since late last year, generative AI has been all the rage — and no product has received more attention than ChatGPT. After its release, it quickly racked up more than 100 million monthly users as the media obsessed over transcripts of journalists’ conversations with the chatbot as though it was really engaging with them and the industry promised this was just the beginning: generative AI was going to upend society — and potentially even threaten the future of humanity.
But ever since the beginning of the hype cycle, there’s been ample reason for skepticism of the narratives being deployed by OpenAI and the other companies and executives incentivized to make chatbots and image generators out to be the next big thing. Was it really any surprise that a free tool got a lot of users to give it a shot when virtually every news program couldn’t shut up about it? And were the expansive claims the industry was making ever an accurate reflection of the reality of the technology?
Earlier this month, data from SimilarWeb — the same company whose figures were used to tout ChatGPT’s rapid user growth earlier this year — showed that just over six months since its launch, ChatGPT saw its worldwide traffic fall 9.7% in June, while unique users dropped 5.7%. Some defenders of generative AI explained the decline by pointing to students getting out of school and not using it for their essays any longer, but SimilarWeb’s data shows the service’s rapid growth began to stall out back in March, well before its June decline....*****... Even if we can explain the drop in traffic by changes in student usage — something I’m not convinced by — it still doesn’t bode well for such a new product that’s supposedly only beginning to transform how we live, work, and engage with digital services. I would argue that statistic points to the fragile foundation the generative AI hype has been built on, and how its reputation is much more hot air than real promise.The next big thingWhen generative AI emerged onto the scene at the end of 2022, Silicon Valley was in a rough place. After building an operating model on the assumption of low interest rates, central banks were rapidly hiking them for the first time in over a decade, squeezing off easy access to capital and demolishing the entire idea of how a tech company should lose boatloads of money to achieve rapid growth. Meanwhile, the industry’s last big bets had also collapsed.The crypto hype peaked in November 2021, then experienced a series of crashes and bankruptcies through 2022, and the metaverse didn’t fare much better. After a ton of companies jumped on Meta’s bandwagon, they just as quickly abandoned their experiments as it became clear the public saw Mark Zuckerberg’s virtual fantasy as more of a joke than a bright future. But those challenges left the industry in a difficult position. Growth was slowing, the pandemic boom experienced by major tech companies was over, large layoffs were being made across the industry, and there was no next big thing.
Silicon Valley relies on a boom and bust cycle to survive. Malcolm Harris, author of Palo Alto: A History of California, Capitalism, and the World, describes how Silicon Valley’s takeaway from the dot-com crash at the turn of the millennium wasn’t to avoid another bust at all costs, but to keep creating new bubbles they could profit from until their collapse, then begin the cycle anew. After crypto and the metaverse, it wasn’t clear what would fill the void — but something was needed to shore up an industry in the sort of rut it hadn’t seen in years. Investors were awaiting a new idea to drive another wave of investment, then ChatGPT exploded onto the scene.
Repeating the playbook
Generative AI followed the typical model of a tech hype cycle. One minute, few people had ever heard of the product, let alone used it. Then, all of a sudden it was everywhere and we were bombarded with stories about its magical qualities and unparalleled capabilities alongside warnings about what it could mean for jobs, society, and the future of our entire species. It was no wonder people were curious and even worried, especially when they didn’t have the skills to assess the tech’s potential for themselves. But it was also very similar to how other tech products have exploded onto the scene in the past, only to be largely forgotten a few years later.For me, the generative AI cycle contains eery parallels with the last time AI and automation were supposed to upend society. Maybe you remember, because it was about a decade ago. At that time, the industry and its media mouthpieces were breathlessly telling us about how the robots were on the cusp of taking many of our jobs. Truck and taxi drivers were about to be eradicated by self-driving cars, baristas and food service workers were going to be replaced by robotic arms, AI bots would take the place of artists and journalists, and robots would proliferate into many other areas of the economy....
....MUCH MORE
Here, with additional commentary, The Richter Scales:
