What happens when people lose their jobs?

Someone made the point: “But today, about 50% of the total wealth is owned by just 1% of the population, which means that a huge chunk of the economy is already diverted from traditional ’employment and consumer spending’ and redirected towards catering to the rich. So it seems that in the future, the rich can continue to concentrate even more wealth in their hands without any repercussions for them.”

Wealth distribution has always been ruled by a kind of Pareto principle, with the top 1% controlling 20 – 40%.

And that works when we still have a functioning economy – relatively low unemployment, with people spending money.

But that changes if we have unemployment of 10, 20, 50% of the population due to AI taking jobs. The IMF is predicting 40-60% of jobs in developed economies will be effected.

“Effected” doesn’t necessarily mean “lost”, but we don’t know what it means. In the past, technology that has replaced jobs but we’ve always been able to re-skill people, find them other things to do for an income. But in a world where AI is taking knowledge-worker jobs, and robots are taking manual labour jobs, I don’t see what kind of work is left. Maybe new things we can’t even imagine will be invented. But what kinds of work can exist that are safe from AI and robots? Not many that I can think of. Chrissy’s job as a violin teacher will probably be safe… if people can still afford lessons for their kids. But the list of jobs that are safe seem pretty limited.

Most wealth is held in assets – shares and/or property, bonds and cash, some gold and crypto. But their value is always relative to the broader economic health of the market.

So let’s say we have massive unemployment. That means people don’t have income. Which means they can’t spend money (unless their income is replaced by something else, eg a UBI, or some other kind of welfare). Which means downward pressure on prices. Which means downward pressure of profits. Which means businesses fail (unless they compensate by replacing their own employees with AI, which may or may not make the problem worse). Which means more unemployment. Real estate prices fall. The share market falls. The price of bonds, gold and crypto falls. Capitalism fails. And if the unemployment persists, it can’t recover.

You can’t have rich people if nobody is spending money in the economy. Wealth has no meaning in an economic collapse.

So let’s assume AI does replace lots of jobs. We will need to make major structural readjustments to the economy – either replacing incomes with some other kind of financial assistance to people who have lost their jobs (and can’t find replacement jobs), or totally restructuring capitalism into some kind of post-scarcity economy, eg the Trekonomics (https://en.wikipedia.org/wiki/Trekonomics).

Mindmapping The Future

What happens in the next five years?

– AI gets massively smarter and more capable

– Probability – 90%

– Based on the statements of pretty much everyone working in the field, Altman, Ilya, Musk, Gates, Kurzweil, Hassabis, LeCun, and including those like Hinton, Wolfram, etc who have no direct skin in the game

– Even if it isn’t 100% LLM (which I doubt it will be) and includes interconnections with specialist systems using symbolic logic and/or other approaches

– Quite a few people who are involved in AI are predicting AGI by 2027.

– What does the world look like when we have machines that are smarter than every expert human in every domain and are available to everyone for $20 a month?

The answer is WE DON’T KNOW. We cannot predict. And when we have arrived at a place where we honestly can’t predict what life will look like in five years, that, by definition, is a technological singularity.

– We cannot predict, but we can make some educated guesses.

– Businesses will try to use AI to increase profits

– When?

– When AI becomes more reliable, which Altman and others are saying with confidence will happen with GPT5, due out this year

– It’ll start slowly, then go very quickly

The first layer will be tasks that are low risk / high cost / benefit

– 2025-26

– Coding (with humans overseeing / checking the code)

– Customer service (web only, then voice, then retail)

– Analysts (with humans overseeing)

– Writers (PR, journalists, marketing)

– Graphic design

– Industrial design

– The biggest short term impact will be just the explosion of intelligence. Imagine a world where PhD level intelligences are available for $20 a month. What will businesses do with all of that intelligence? A million new scientists to solve our biggest problems, reading and analysing all of the research, developing new trials, running those trials in virtual environments, presenting the best vectors to humans for lab experiments. Imagine a million new scientists hitting the world overnight. (https://cameronreilly.com/a-million-new-everythings/)

The second layer will start to happen when there is confidence in AI capabilities from the experience with the first layer

– 2027-2030

– AGI might arrive around this time, too

– Higher level jobs will start to be replaced

– middle management (because there will be less people to manage)

– legal

– accounting

– HR (again, less people to hire / manage)

– recruitment

– psychologists – everyone has a free AI therapist who knows you more intimately than your own family

– medical – everyone has a free GP

– business strategy

– animation

– acting / film and tv production (as more and more is done with AI)

The third layer will come with cost-effective humaniod robots

– 2030-2040

– manual labour in mines, factories, workshops, maintenance

– When?

– When the TCO of a robot is cheaper than a human

– A labourer costs how much? $50 – 150K a year, depending on the industry?

– A robot will last how many years? 10?

– When you can buy a robot for <$500K, it becomes economically viable.

– Goldman Sachs:

– [“The total addressable market for humanoid robots is projected to reach $38 billion by 2035, up more than sixfold from a previous projection of $6 billion”](https://www.goldmansachs.com/intelligence/pages/the-global-market-for-robots-could-reach-38-billion-by-2035.html)

– [“The manufacturing cost of humanoid robots has dropped — from a range that ran between an estimated $50,000 (for lower-end models) and $250,000 (for state-of-the art versions) per unit last year, to a range of between $30,000 and $150,000 now. Where our analysts had expected a decline of 15-20% per annum, the cost declined 40%.”](https://www.goldmansachs.com/intelligence/pages/the-global-market-for-robots-could-reach-38-billion-by-2035.html)

– “The team’s base case is for more than 250,000 humanoid robot shipments in 2030, almost all of which would be for industrial use. Our analysts’ base case is for consumer robot sales to ramp up quickly over the next decade, exceeding a million units annually in just over a decade.”

– [“In December 2023, billionaire venture capitalist Vinod Khosla made this prediction.](https://www.linkedin.com/pulse/humanoid-robots-here-soon-millions-billions-them-peter-h-diamandis-wdgne/)

“By 2040 there could be a billion bipedal robots doing a wide range of tasks including fine manipulation. We could free humans from the slavery of the bottom 50% of really undesirable jobs like assembly line and farm workers. This could be a larger industry than the auto industry.””

“Robohub, a nonprofit robotics organization, provides a perspective on this. They argue that the “holy grail” for humanoid robots would be crafting sophisticated tech under a $50,000 price tag.

This figure isn’t arbitrary—it aligns with the annual wage of a single shift of labor at just over $18/hour, resonating with the ongoing labor shortages in low-wage industries.

On the other hand, Macquarie dives deep into the cost breakdown for early-stage humanoid robots. Their estimate? A slightly more optimistic $40,000. With allocations like $10,000 for sensors and chips, $5,000 for torque sensors, and $8,000 for precision reducers, they’ve dissected the cost matrix intricately.”`

– Of course, by 2035, many people will be out of work, so who will be able to afford a robot?

– unless… we have not-for-profit robot factories, staffed by robots, making other robots, the costs fall dramatically

– and AI helps us develop nanotechnology, so we can have tiny robots breaking down waste products into their molecular components (oxygen, carbon, hydrogen, nitrogen, silicon, copper, iron) and then building new components out of them

– every home has their own nanofabricator and the first thing you do when you get one is make one for your neighbour

How will businesses use AI?

– Improved customer service experiences

– AI agents know more about the company, more about the customer, are cheaper to run, faster, better at customer relations, can have whatever accent / speak whatever language the customer prefers (no more complains about Indians, etc)

– Reduce employee headcount

– Less managers required

– Less administration required

– Less coders required

– Less people-facing people required

– AI agents can have conversations / take orders / make sales calls with far high efficiency than humans (either voice only, email, or realistic human avatars in a video call)

– What do people do that machines (AI + robots) can’t do?

– Less teachers required

– When / How will unions get involved?

– What about white collar workers who aren’t unionised?

– Governments will try to use AI to reduce costs, improve services

– People will use AI to keep their jobs, in their personal lives until their jobs have been replaced with AI

– What happens when people start losing their jobs?

– High unemployment

– Less cash in the economy

– Business will suffer as there is a cash squeeze

Governments have to intervene to stop economies from collapsing

– Using the usual economic tools – interest rates, printing money, handouts

– AI tax on corporations, goes to UBI / welfare

– Does that create a disincentive on corporations to replace workers in the first place?

– But it will happen in stages, first the jobs will be replaced, the governments will be slow to react

– As people lose their jobs they will become angry – at corporations, at governments, at technology companies

– The property market will implode as people are forced to sell their houses

– Share market will tumble as people pull money out and businesses struggle with less cash in the economy

What holds back adoption of most technological revolutions?

– business / consumer apathy – the market just doesn’t care as much as the tech companies think they will or has a downright negative reaction, eg Google Glass

– doesn’t apply to AI, everyone wants it

– Cost – products are just too expensive to gain enough traction, eg Segway

– consumer level will be low cost / free services (eg iPhone)

– business will justify it by reducing headcount

– Requires people to change behaviours / habits, learning curve too difficult, cost to benefit ratio too high

– AI will have a very low learning curve, it speaks natural language, and will teach you how to interact with it by suggesting prompts

– Will economies with greater government controls, eg China, fare better?

Some people will learn how to use AI to be smarter

– Analyse legal documents to avoid falling into traps with insurance, finance, employment, etc

– Analyse politicians speeches, bills in front of Parliament, news stories

– Organisations (business, political, etc) will use AI to create more propaganda / lies

– But some people will be able to use AI to see through it

– However many won’t use it that way

– AI tools will be built into our every day devices

– from our phones and computers to our Roomba, car, etc.

– Low cost, just a chip which connects to a cloud LLM or runs locally, small footprint model

– Mega Corporations will try to offer their own LLM, eg General Motors, to control the user experience, and will burn billions of dollars, and ultimately fail

Your personal computing device will be your dominant AI agent

– It will be your intermediary with the world

– It will read your emails, text messages, watch / record / save / analyse what you’re doing during the day (both on your devices as well as IRL), listen to your calls, your conversations, etc, and it will all be stored online, backed up, indexed for retrieval.

– People might scream about privacy issues – but we’ve been here before (CCTV, cookies, mobile phones tracking us, credit cards online) and what we’ve learned is that people will trade privacy for convenience, services and benefits if the trade-off seems beneficial

– “I’m not committing any crimes, so what do I care?”

What are the implications?

– Legal – people will have access to video / audio recordings of every interaction, every conversation

– Employers might have access to your recordings made during work hours – but you also have access to all of your conversations with them and colleagues and customers

– The End of Deception?

– Nobody will be able to lie and get away with it

– “Oh my AI was turned off” will be as suspicious as cops turning off their body cam

– Infidelity

– Crime

– What if the cops get an alert when someone turns off their AI monitoring device?

– “Sir, why was your AI turned off at 11pm on the night of January 2?”

– Will wearing an AI device become mandatory?

– We could do that already with some kind of audio/video recorder.

– Courts / police will have access to most recordings for trials

– Marital – spouses will demand access to each other’s recordings

– Arguments about who said / did what will disappear

– But will be replaced by “it’s what I *meant*”, even if it’s not what I *said*”

– It will suggest ideas / products / services / music / shows to improve your life

– It will filter out all advertising / marketing unless you opt-in

– So that’s the end of advertising, marketing, influencers

– It will be your therapist, dietician, coach, friend, advisor, consigliere, teacher and mentor

– Hackers will develop open source models that will be good enough for many daily activities, run locally, or in a trusted environment (eg Wikipedia)

– They won’t be as capable as the massive models, eg OpenAI / Microsoft, but will be good enough for many tasks

What if AI doesn’t get massively smarter?

– AI only gets marginally smarter and more capable

– Probability – 5%

– AI does not get smarter or more capable

– Probability – 5%