This might be somewhat off topic, but do you really think the next five years will proceed as you predict if 90% of the workforce is rendered redundant? Are you not worried the pitchfork-wielding neo-Luddite mobs will be burning down data centres and attacking corporate HQs once the unemployment rate hits a mere 20-30 per cent? Might that not be a (predictable – as you flag but don't really explore) fly in the Operating Model Compression Playbook?
I worked on some of these as part of setting things up at a startup 18-21 months ago (when I still had the leeway to attempt it at early stages as a technical founder). It's great to see how far things have come, at least by way of the discourse and (hopefully) an increasing understanding of the value of these things.
"Operating model compression flips the “10x developer” idea on its head - if you increase productivity 10x you only need 10% of the original resource footprint. That’s a 90% compression ratio."
That holds true, as long as we assume that code=product.
Which is false.
And I'd bet in 5 years from now it will stay so.
In fact, if all those laid-off developers jump on the "I'll just build my own product" bandwagon, we should expect tons and tons of new products out there. But we won't see a similar change in the buying power. What follows is:
* There would be even fiercer competition between aspiring product developers
* There will be a lot of noise for customers (a lot of vibe-coded apps, probably nor responding to actual need)
If we take hints from other industries on what a similar "democratization" of means did, it wasn't all rosy. Thanks to Spotify and others, anyone can reach the whole world with their music. Yet only very few top artists can make a living out of that. For everyone else, it's a hobby. *Despite* the access to hundreds of millions of potential customers.
This is more about more than just code though - it’s about all context to run a business: procedures, SOPs, policies, frameworks, client records, archives, network drives - literally everything. Live recording of all meetings and pushing that context to the repo as a transcript. Interviews taking multiple hours with key people in every function. Anything to get an edge.
I get it. We can be more efficient with a lot of stuff we do now.
We can store and extract information in way more convenient way.
The question remains: what is the goal?
Because all these things are but means to an end. Another wave of improvements, like digital information (vs printed), or internet (vs physically shipped/faxed), etc.
I commented on the article, since you mentioned ends which I don't think can have a chance to materialize en masse.
Assuming vibe coding will become a viable path to build software in the long run (it is not even close now), what will the whole business ecosystem will look like? Because that aspiring entrepreneur will be working in this future-state ecosystem, not the current one.
In other words, if you're the only person who has a Gutenberg press, then you're bound to leapfrog your competition. If everyone has one, then it's just a commodity. Same with access to the internet. Same with AI capabilities.
Speculatively (and I reserve the right to be wrong), there will be less advantage coming from using AI to get "literally everything" more efficient than from making "literally everything" meaningful and aligned with goals. Making a meeting more efficient won't do much if that meeting shouldn't be happening at all or has little purpose whatsoever.
Drucker famously said: "There is nothing so useless as doing efficiently that which should not be done at all."
Now, if your vision comes true, then:
a) We can see the improvements because we will have smaller organizations, and smaller organizations are, by design, more effective (per person)
b) We will face a social crisis as a huge part of knowledge work will not require knowledge *workers*
I would bet on another future, though. One that rhymes with what happened in the past. The change will be more gradual (yet it might still be considered punctuated equilibrium). Instead of scaling down, companies will be inventing new things to work on, so we won't see that rapid change in the job market (again, at a large scale; I don't deny many individuals will be on the sharp end of the stick). Finally, the size will remain an advantage, and industries will be dominated by relatively few big corporations rather than a dense network of small players.
Such 'democratization' of means is never all rosy - in the beginning, alongside huge amounts of hype, like you said, there will be a lot more entrants into the field, most of them being 'noise', but eventually only a few of them survive and there will be consolidation?
Doesn’t the same apply to operational compression of the entire economy? Plausibly, 90% of current companies (and business value created) only exist to address the “inefficiencies” that will be replaced by AI-driven processes. It’s not clear to me what is left of the economy after a couple of rounds of operational compression, and what is the business value actually being created at the end?
Yep - that’s one of the next issues to investigate - where does demand come from? How do societies function?
Accumulation of capital / physical resources protected by state or quasi-state power might even be a short run thing in this scenario - like grabbing all the chips before the timer runs out.
Insightful thanks. I’m curious if consumers are ready to make the leap to supporting AI first companies to the extent outlined by the latter years here. Or will it feel strange and uncomfortable for many people. Will companies suffer from not having a full suite of staff? Because users won’t be able to relate… Or will they not care? This is an unknown right now… Yes operating costs can go much much lower but there’s a human side which will be missed
I'm writing on this topic. As AI-driven production costs drop, prices will also tend to drop, sometimes significantly. But that will be, as you say, at the expense of the human side. So there likely will in parallel be a human-based offering which will cost much more, but will mainly be preferred for reasons of that personal touch or prestige.
This might be somewhat off topic, but do you really think the next five years will proceed as you predict if 90% of the workforce is rendered redundant? Are you not worried the pitchfork-wielding neo-Luddite mobs will be burning down data centres and attacking corporate HQs once the unemployment rate hits a mere 20-30 per cent? Might that not be a (predictable – as you flag but don't really explore) fly in the Operating Model Compression Playbook?
A few guild-like professions will make a stand but those making decisions on how to allocate capital will win.
call me cynical but I don’t see any feasible path as at today’s date for slowing this trend down.
I like the new phrase. Thanks. I will use it.
Thanks Jurgen
I worked on some of these as part of setting things up at a startup 18-21 months ago (when I still had the leeway to attempt it at early stages as a technical founder). It's great to see how far things have come, at least by way of the discourse and (hopefully) an increasing understanding of the value of these things.
"Operating model compression flips the “10x developer” idea on its head - if you increase productivity 10x you only need 10% of the original resource footprint. That’s a 90% compression ratio."
That holds true, as long as we assume that code=product.
Which is false.
And I'd bet in 5 years from now it will stay so.
In fact, if all those laid-off developers jump on the "I'll just build my own product" bandwagon, we should expect tons and tons of new products out there. But we won't see a similar change in the buying power. What follows is:
* There would be even fiercer competition between aspiring product developers
* There will be a lot of noise for customers (a lot of vibe-coded apps, probably nor responding to actual need)
If we take hints from other industries on what a similar "democratization" of means did, it wasn't all rosy. Thanks to Spotify and others, anyone can reach the whole world with their music. Yet only very few top artists can make a living out of that. For everyone else, it's a hobby. *Despite* the access to hundreds of millions of potential customers.
This is more about more than just code though - it’s about all context to run a business: procedures, SOPs, policies, frameworks, client records, archives, network drives - literally everything. Live recording of all meetings and pushing that context to the repo as a transcript. Interviews taking multiple hours with key people in every function. Anything to get an edge.
I get it. We can be more efficient with a lot of stuff we do now.
We can store and extract information in way more convenient way.
The question remains: what is the goal?
Because all these things are but means to an end. Another wave of improvements, like digital information (vs printed), or internet (vs physically shipped/faxed), etc.
I commented on the article, since you mentioned ends which I don't think can have a chance to materialize en masse.
Assuming vibe coding will become a viable path to build software in the long run (it is not even close now), what will the whole business ecosystem will look like? Because that aspiring entrepreneur will be working in this future-state ecosystem, not the current one.
In other words, if you're the only person who has a Gutenberg press, then you're bound to leapfrog your competition. If everyone has one, then it's just a commodity. Same with access to the internet. Same with AI capabilities.
Speculatively (and I reserve the right to be wrong), there will be less advantage coming from using AI to get "literally everything" more efficient than from making "literally everything" meaningful and aligned with goals. Making a meeting more efficient won't do much if that meeting shouldn't be happening at all or has little purpose whatsoever.
Drucker famously said: "There is nothing so useless as doing efficiently that which should not be done at all."
Now, if your vision comes true, then:
a) We can see the improvements because we will have smaller organizations, and smaller organizations are, by design, more effective (per person)
b) We will face a social crisis as a huge part of knowledge work will not require knowledge *workers*
I would bet on another future, though. One that rhymes with what happened in the past. The change will be more gradual (yet it might still be considered punctuated equilibrium). Instead of scaling down, companies will be inventing new things to work on, so we won't see that rapid change in the job market (again, at a large scale; I don't deny many individuals will be on the sharp end of the stick). Finally, the size will remain an advantage, and industries will be dominated by relatively few big corporations rather than a dense network of small players.
Such 'democratization' of means is never all rosy - in the beginning, alongside huge amounts of hype, like you said, there will be a lot more entrants into the field, most of them being 'noise', but eventually only a few of them survive and there will be consolidation?
Doesn’t the same apply to operational compression of the entire economy? Plausibly, 90% of current companies (and business value created) only exist to address the “inefficiencies” that will be replaced by AI-driven processes. It’s not clear to me what is left of the economy after a couple of rounds of operational compression, and what is the business value actually being created at the end?
Yep - that’s one of the next issues to investigate - where does demand come from? How do societies function?
Accumulation of capital / physical resources protected by state or quasi-state power might even be a short run thing in this scenario - like grabbing all the chips before the timer runs out.
Insightful thanks. I’m curious if consumers are ready to make the leap to supporting AI first companies to the extent outlined by the latter years here. Or will it feel strange and uncomfortable for many people. Will companies suffer from not having a full suite of staff? Because users won’t be able to relate… Or will they not care? This is an unknown right now… Yes operating costs can go much much lower but there’s a human side which will be missed
Thanks Matt - what will be lost through this sort of change is the big unknown…
I'm writing on this topic. As AI-driven production costs drop, prices will also tend to drop, sometimes significantly. But that will be, as you say, at the expense of the human side. So there likely will in parallel be a human-based offering which will cost much more, but will mainly be preferred for reasons of that personal touch or prestige.
https://ashstuart.substack.com/t/econ