The far end of the pro-AI movement believe that we are “on the foothills of the Singularity” – that is, we’re about to reach the invention of AGI, or ‘Artificial General Intelligence’. These we can split into the AI Pessimists and AI Optimists. The Optimists believe AI is going to change everything – it will replace all human jobs and human labour while creating unbelievable economic growth and prosperity. The Pessimists believe AGI is the most dangerous invention humanity has ever strived for, and unleashing it could end in nothing less than an extinction event.
The other end of the spectrum is the AI skeptics, who believe that AI, and especially LLMs and Generative AI, is the biggest scam in the history of scams. That LLMs are just a big word prediction machine (which, on a very reductive level, they kind of are).
I’m somewhere in the middle.
There is a LOT of hype around AI. The amount of money that is being spent on GPUs and data centres to train and run inference across these models is staggering. The circular financing between the big AI companies feels to me at best extremely dangerous.
But the value is real.
I’ve seen it in software development and coding. I’ve seen teams shipping code to production that was written by AI (and reviewed by humans).
I’ve seen how useful LLMs can be to help creative thinking or act as a strategy sparring partner. I often use it now to brain dump a bunch of unstructured thoughts out of my head in stream-of-consciousness style, and let the LLM organise it and play it back to me in a structured way.
I think the biggest problem isn’t that there’s no value here… it’s that the big AI companies are promising a bit too much of the future. The models are becoming ever more capable, but they can’t do everything. Not yet at least. There appears to be very little evidence that they are really displacing human jobs yet.
Right now it feels to me that giving everyone in white collar work LLMs today is a bit like when we gave everyone in the finance department spreadsheet software in the 80’s. While it is true that the advent of Spreadsheets drastically reduced the need for clerks and manual data entry specialists, it also exponentially increased the demand for financial analysts. It moved the value up the stack, increased quality and reduced cost… which didn’t result in fewer jobs, it resulted in increased demand, which led to more jobs… It’s an imperfect analogy, but illustrative of the dynamic.
This I guess is the bull case against massive job displacement due to LLMs… they will create efficiency, reducing cost (and therefore jobs), but then lower cost will generate more demand, leading again to more jobs…
We’ll see how it plays out. For now, what we know for sure is that there’s definitely value being delivered by LLMs… whether it lives up to the insane promise to justify the even more insane investments is the trillion dollar question!
Leave a Reply