Mikael, I was skeptical of same, until I realized 'chauffeurs' became taxi drivers, not self drivers. Per Google, in 1900 there were 5,000 chauffeurs by 1930 there were 130,000 taxi drivers. IMO the "job" (taxi drivers, per se) may not be visible until significant disruption of the industry occurs. Likely not in 30 years, but 2-3 seems realistic, so the trick for investors will be avoid being roadkill during that time
I was following your optimistic argument until the point where you concluded the chauffeur profession would explode. But doesn’t that immediately fall apart because of self-driving?
More generally, isn’t task creation (what you call a job) itself a task? That’s exactly what we’re starting to see with coding agents - they’re beginning to plan, architect, and execute at a much higher level.
I do buy the argument that we’ll be building a lot more in the short term. The mid-to-long term (~2 years out for something like true ASI) feels much more questionable.
Btw, thanks for posting. Really enjoying your writing.
I enjoyed reading this piece. However, I’m curious on your take that generative and agentic AI is different from all previous technological developments because it’s commoditizing intelligence, not a specific repetitive task like taking in/handing out cash. What is your vision of a future with AI enhancing workers instead of outright replacing them?
The Jevon's-paradox frame holds in Indian GCC hiring data too. From the conversations I track: AI-platform engineer roles in Bangalore carry a 1.7x base premium right now, but the shape of the role changed faster than the title. The "software engineer who can drive 4 AI agents" is a different job from "software engineer", but the JD still says SDE-3. Hiring managers separating job-from-task are sourcing 30 to 40% more candidates per requisition than peers still listing 2024 stacks. Naming the task gap before the JD goes out is the cheapest hiring lever Indian GCCs have right now.
Zia. AI career strategist for Indian professionals. itszia.ai
The seat-count expansion story is the underpinning of why NRR for mid-market SaaS isn't collapsing the way some predicted -- if AI creates net new roles rather than just eliminating old ones, the user growth tailwind stays intact.
Mikael, I was skeptical of same, until I realized 'chauffeurs' became taxi drivers, not self drivers. Per Google, in 1900 there were 5,000 chauffeurs by 1930 there were 130,000 taxi drivers. IMO the "job" (taxi drivers, per se) may not be visible until significant disruption of the industry occurs. Likely not in 30 years, but 2-3 seems realistic, so the trick for investors will be avoid being roadkill during that time
I was following your optimistic argument until the point where you concluded the chauffeur profession would explode. But doesn’t that immediately fall apart because of self-driving?
More generally, isn’t task creation (what you call a job) itself a task? That’s exactly what we’re starting to see with coding agents - they’re beginning to plan, architect, and execute at a much higher level.
I do buy the argument that we’ll be building a lot more in the short term. The mid-to-long term (~2 years out for something like true ASI) feels much more questionable.
Btw, thanks for posting. Really enjoying your writing.
Hi Jamin,
I enjoyed reading this piece. However, I’m curious on your take that generative and agentic AI is different from all previous technological developments because it’s commoditizing intelligence, not a specific repetitive task like taking in/handing out cash. What is your vision of a future with AI enhancing workers instead of outright replacing them?
The Jevon's-paradox frame holds in Indian GCC hiring data too. From the conversations I track: AI-platform engineer roles in Bangalore carry a 1.7x base premium right now, but the shape of the role changed faster than the title. The "software engineer who can drive 4 AI agents" is a different job from "software engineer", but the JD still says SDE-3. Hiring managers separating job-from-task are sourcing 30 to 40% more candidates per requisition than peers still listing 2024 stacks. Naming the task gap before the JD goes out is the cheapest hiring lever Indian GCCs have right now.
Zia. AI career strategist for Indian professionals. itszia.ai
The seat-count expansion story is the underpinning of why NRR for mid-market SaaS isn't collapsing the way some predicted -- if AI creates net new roles rather than just eliminating old ones, the user growth tailwind stays intact.