The AI Jobpocalypse 180
The truth lies in the middle
First, AI was to automate every job if you believed the Technology press and the founders of the AI labs. Now, both the AI lab founders (Dario and Sam) as well as Technology media say there will be no Jobpocalypse.
To me, this latest turn smacks of IPO positioning - neither company wants massive public backlash ahead of IPO pricing. It is positioning much the same way Anthropic positioned its GPU shortage into a marketing campaign where its forthcoming Mythos model was “too dangerous and effective to make public until further review”, when in fact Anthropic lacked the GPU capacity to make Mythos affordable. Funny how Opus 4.8 was released as Anthropic capacity came online and Mythos will arrive any day now.
My experience with the models is that they are effective, even if at times very frustrating (limited context, lack of common sense, lack of persistent memory). I use the models 10-12 hours per day for coding and they are 100% capable of automating tasks and orchestrating larger workflow automations. I see it with Kilby, which took a long time to build, required a lot of research, and requires continuous investment to stay ahead of Anthropic and OpenAI. Yet, for all of Kilby’s automation capability, broad knowledge about everything, and deep knowledge around the Capital Markets and Leadership, it hasn’t been terrible to write code for primarily as a result of how good the coding models are (Kilby does write code as well).
My view is that first companies will automate tasks, then they will rethink their entire workflows as functional silos break down (I’ve seen this on the marketing side of TEK2day). Publishing these short articles and long reports require a fraction of the time they used to require, primarily because the models Kilby/Anthropic/OpenAI/Gemini automate the header artwork and can surface items from my disparate notes that may be useful to include in a given article.
More productivity means more revenue. I think this is what people miss about AI - they only think of the Operating Expense savings side, when in fact the Revenue opportunity side of the equation is probably the larger opportunity over the long-term as has always been the case with automation.
This podcast episode with Op Leaders at ServiceNow (NOW) speaks to what I see with my workflows and what I have seen with other companies. NOW has skin in the AI game, but what is said here is true:
Try Kilby HERE