I heard Wayne Holmes make this argument at a recent conference. It’s not the first time I’ve heard someone suggest recently that a crash is coming, with Gary Marcus probably being the most influential voice. It’s certainly true we’ve been in a hype cycle and we might currently be passing the peak of inflated expectations, it doesn’t follow that another winter is coming. I suspect Wayne’s framing of an autumn is more accurate and this is probably a good thing on balance, because it enables a more careful grappling with the implications of these technologies within higher education beyond the accelerated pace of development.
It’s not obvious to me this is guaranteed though. I suspect if Claude 3 Opus had been released by ChatGPT and called GPT 5 it would have reignited the hype wave. Its new capacities simply aren’t getting the wider attention they would if OpenAI had released the model. Likewise there’s still the possibility of an actual GPT 5 or AppleGPT coming with significant implications, even if the evidence of this admittedly doesn’t look great.
The problem I think with hype wave explanations is that they tend to be deflationary, pointing to past hype waves in order to explain away the longer term implications. This obscures the timespans over which sociotechnical systems tend to influence organisations and wider society e.g. the societal impacts of social media being felt now almost two decades after the major companies launched. Furthermore, I just don’t think we can meaningfully point to something like the MOOC in order to dismiss GAI as hype, as Martin Weller persuasively summarises here:
The obvious comparison with MOOC hype is AI hype. So one could draw the conclusion that in 10 years we’ll be going “remember AI?”. I think that’s unlikely, it looks set to be a technology that will integrate into existing tech and is causing higher ed to ask fundamental questions of itself and practice. If AI does nothing else but get rid of the essay as the default assessment mode, then it’s impact is profound. But the MOOC lesson should at least give us caution over some of the more revolutionary, rapture type claims. I’m sure everyone will make sure they don’t make any over the top claims this time, eh?
https://blog.edtechie.net/mooc/dont-look-back-in-anger-or-anything-else/
Furthermore, is it so obvious that investment in GAI will dry up in the absence of profitability? If interest rates fall over the next few years, the economic viability of the future stories that Silicon Valley fundamentally sell (Dave Karpf has been very good on this recently) will increase correspondingly. We might not be going back to the low interest rate environment which facilitated the initial consolidation of big tech, but the conditions in which they struggled post-pandemic and which led to the desperate latching onto GAI aren’t going to last either. This is increasingly being folded into their cloud businesses which is the core dynamic of platform capitalism, rather than some frothy extra which can be dispensed with at will. The geopolitical dimension of AI being used to fuel the cold war leaves tech firms bound up in the national security apparatus as well, to a much greater degree than was previously the case. The economic and political interests at work here suggest the GAI investment could continue even as we move through the hype cycle, suggesting perhaps that movement ‘through it’ will be much slower as a result.
Mark Carrigan 