People have a bad intuition about economic mechanisms, but that does not mean that AI won't cause huge regressive effects.
The transition from independent artisans to factory work didn't cause massive unemployment. But it cause the former artisans to lose out, because they no longer had the bargaining power they had before - their work was now a commodity. This resulted in "Engel's pause", a period in which wages stagnated, and so conditions got worse because prices also increased.
But AI also isn't just automation. It's the automation of automation. You may think that you will just level up and learn the new skill set of managing AIs, but the AI companies will be collecting the data to automate that too. Often by watching you.
From a recent article in The Economist. It turns out that once again many of the jobs that people predicted would be eliminated by AI have instead received a boost, at least for now. This closely follows another recent report on how Indian BPO firms are surprisingly thriving due to AI.
"Moreover, AI is helping offshore workers in the Philippines do new, more complex jobs. They are picking up work training AI models or supervising AI agents. Hospitals in America are increasingly outsourcing the checking of insurance eligibility and filing of medical records to Filipinos packing AI tools. Mr Gallimore says more “high-value” work, such as accountancy or engineering, is going offshore in part because “AI is a leveller. You can now teach someone complex stuff quickly” and still get it done more cheaply than in America or Europe. Some of this is done at so-called global capability centres, captive back offices that are booming in India and also now growing in the Philippines."
It seems like AI has incidentally commoditized the complement (procedural knowledge), and in turn created greater demand for inputs like human judgement and "empathy".
> It seems like AI has incidentally commoditized the complement (procedural knowledge), and in turn created greater demand for inputs like human judgement and "empathy".
I think as well there is currently the following: AI still needs an operator, in the same way than a bulldozer needs a bulldozer operator. And like with bulldozer, there are always something that business stakeholder wanna dig, now that digging is faster.
I visited India recently and when I saw just how many companies have a GCC there I was mindblown. Lots of cool offices, engineering and technical roles, great work life balance and generally a different vibe. Literally every company that I could recognise had a back office there.
> Hospitals in America are increasingly outsourcing the checking of insurance eligibility and filing of medical records to Filipinos packing AI tools
So, insurance claims are handled by people who operate in different legal system and by unaccountable undererministic system. And people are pressured to perform fast.
This is not about "automating complex task". This is about "create accountability sink for fraud we want them to do and create system that will perform said fraud".
Yeah, paradoxically Indian software outsourcing is growing. I guess the way the companies are thinking about this is we need some sort of human oversight for all the work the LLMs are doing, might as well pay someone the least amount possible to do that.
Don't understand why this should not have been obvious. If you have technology that allows expensive workers to do things cheaper why wouldn't same technology allow cheaper workers to do stuff even cheaper... And cheap is good...
I know a freelance translator and they have more business than ever, AI being "mostly" right is not good enough in many businesses. You need oversight and most importantly - accountability in the form of a human that can put their stamp of approval.
I’m not sure if this is true. I really want the hiring story to be true, but the responses I am getting from people on the ground vs industry generated reports isn’t meeting up.
The only report that I know of that is available is the ICRIER report.
I haven’t read the report itself, but it’s supposed to show modest hiring growth.
Conversely I know that most of the outsourcing vendors are seeing reductions in headcount when it comes to outsourced tech operations.
"Computer says no" is about unhelpful bureaucracy. Being knowingly locked to a specific procedure and unwilling to divert from it is a different thing from mistakes being made that you are blissfully unaware of. The problem is now invisible and the patient is dying for no reason anyone can discern.
The "unhelpful bureaucracy" arises precisely because the process is structured so there's no point in the chain that can actually consider and solve your problem directly. The process is the point, outsourcing work to the opposite end of the globe has long been a perfect way to guarantee no it's robust against errors like "compassion" and "helpfulness", and that's as true of LLMs performing those functions tomorrow as it was of people doing it yesterday.
Well, possibly. To paraphrase the great American philosopher Donald Rumsfeld - there are known knowns, known unknowns, unknown knowns and unknown unknowns. And making the actual process a black box for the people managing it will create a lot more Unknown Unknowns.
At a former job for a regional bank in South East Asia my boss used to joke that outsourcing is really great for the company, because now they can fail to deliver their software projects for cheaper.
POGOs are criminalized and often operated by illegal immigrant Chinese people, it's not really representative. At least that's my understanding having lived in Manila for a few years.
Outside the English-speaking world, there's generally been a knowledge gap when it comes to programming. Better resources weren't being shared, and AI helps bridge that gap. You could learn the basics, but higher-quality materials were hard to come by. AI actually solves that—so people's reactions depend a lot on their environment
Even learning through open source participation requires understanding the English speaking culture, and that's a kind of knowledge that only people who have lived in those countries can really grasp.
In my case, most of my knowledge came from Microsoft's learning center materials. To actually learn deeply, I had to read papers. And I only knew about them because I went to graduate school. But there were already many libraries that made it easier to apply that knowledge without reading the papers directly.
So I think how useful AI is depends a lot on where you live and what kind of environment you're in
Overall it seems we don't have good intuitions for what kinda work AI 'displaces'. Years of hand-wringing about programmers being put out of a job and instead programmers are the ones paying $200/mo for LLMs
And on the flipside a lot of AI vendor ideas like helping doctors turn voice transcripts into notes has just been annoying for the prospective user--they feel they may as well type it up themselves if they then have to review and correct the AI notes
The transition from independent artisans to factory work didn't cause massive unemployment. But it cause the former artisans to lose out, because they no longer had the bargaining power they had before - their work was now a commodity. This resulted in "Engel's pause", a period in which wages stagnated, and so conditions got worse because prices also increased.
But AI also isn't just automation. It's the automation of automation. You may think that you will just level up and learn the new skill set of managing AIs, but the AI companies will be collecting the data to automate that too. Often by watching you.
"Moreover, AI is helping offshore workers in the Philippines do new, more complex jobs. They are picking up work training AI models or supervising AI agents. Hospitals in America are increasingly outsourcing the checking of insurance eligibility and filing of medical records to Filipinos packing AI tools. Mr Gallimore says more “high-value” work, such as accountancy or engineering, is going offshore in part because “AI is a leveller. You can now teach someone complex stuff quickly” and still get it done more cheaply than in America or Europe. Some of this is done at so-called global capability centres, captive back offices that are booming in India and also now growing in the Philippines."
It seems like AI has incidentally commoditized the complement (procedural knowledge), and in turn created greater demand for inputs like human judgement and "empathy".
I think as well there is currently the following: AI still needs an operator, in the same way than a bulldozer needs a bulldozer operator. And like with bulldozer, there are always something that business stakeholder wanna dig, now that digging is faster.
Oh dear
So, insurance claims are handled by people who operate in different legal system and by unaccountable undererministic system. And people are pressured to perform fast.
This is not about "automating complex task". This is about "create accountability sink for fraud we want them to do and create system that will perform said fraud".
> https://researchprofiles.ku.dk/en/publications/winners-and-l...
Surely, AI helps them do stuff even cheaper, and cheap is good, so their business should be booming?
The only report that I know of that is available is the ICRIER report.
I haven’t read the report itself, but it’s supposed to show modest hiring growth.
Conversely I know that most of the outsourcing vendors are seeing reductions in headcount when it comes to outsourced tech operations.
I think that two arent as different as you believe.
Even learning through open source participation requires understanding the English speaking culture, and that's a kind of knowledge that only people who have lived in those countries can really grasp.
In my case, most of my knowledge came from Microsoft's learning center materials. To actually learn deeply, I had to read papers. And I only knew about them because I went to graduate school. But there were already many libraries that made it easier to apply that knowledge without reading the papers directly.
So I think how useful AI is depends a lot on where you live and what kind of environment you're in
And on the flipside a lot of AI vendor ideas like helping doctors turn voice transcripts into notes has just been annoying for the prospective user--they feel they may as well type it up themselves if they then have to review and correct the AI notes