An interesting caution on Coast FI and maybe Barista FI from Nick Maggiulli
See
Basic concern is if AI takes too many jobs, it could raise unemployment permanently and structurally eventually (AI could provide more jobs temporarily, but net eliminate jobs over time), which makes FIRE a terrific mitigation. But Coast FI or Barista FI dampens that risk reduction compared to full FI.
Nick (the guy behind "Of Dollars and Data") summarized this logic as follows:
Why take the step back now when you could do it in a few years? After all, if AI doesn't take your job, you're fine. And if it does, you'll be glad you acquired as much capital as possible while you still had the chance
Here he covered it more completely on his own blog
Join the conversation
Sign up to reply, follow discussions, and connect with the ChooseFI community.
Comments
as I’m currently implementing AI at a company there is definite truth of changes to the future job market. Recent reports from McKinsey and others suspect an initial job increase as companies figure it out followed by decline with varying levels of net effects. Some experts say 20% unemployment, others say this will add jobs as new roles are developed that don’t exist yet.
Short way of saying depends on why your coast/barista FI job market is and your optionality. If your coast FI plan is you only want to do side work that will easily be replaced by AI, probably need another plan. If barista to lean FI and open to coffee shop or other service / manual role, then no need to change, but just be aware and flexible.
Great nuance JoeQ17 regarding some categories of jobs being more vs less vulnerable to AI replacement.
Though I will say that even "open to coffee shop or other service / manual role" is not immune from eventual AI job replacement. Though these types of jobs would likely take a decade or more before AI/robotics encroached as seriously in these categories, whereas other job categories may be at risk much sooner.
It is plausible that at some point in the next 1-3 decades, the only jobs left for humans would be where people were willing to pay a premium for human vs robotic labor. Not claiming that outcome will happen - but it could.
there will be a point where restaurants have robot waiters and bartenders, sure. Other than for novelty purposes I don’t see it 30 years out. Similar for manual work such as construction. While the fear is so many white collar jobs are lost that they flood the blue collar market.
Being a capitalist and consumer country is suspect more new jobs will appear, new industries, new ways to make money and spend money that don’t exist yet. But there may be some points of 1% and 20% unemployment while we all figure that out. So flexible
there will be a point where restaurants have robot waiters and bartenders, sure. Other than for novelty purposes I don’t see it 30 years out.
The novelty phase is already here. At least one cruise ship already has a robotic bartender
https://www.royalcaribbean.com/cruise-dining/bionic-barand I've eaten at sit-down (not casual dining) restaurants where:
- You order and pay on a tablet at the table (not really a novelty - just a preferred option for ordering, and default option for paying)
and another where:
- A server bot brings out the food - still transferred to the table by human waitstaff, but bot transferring the food is likely soon - at least as a novelty.
The reason the human waiter or bartender may eventually become the novelty rather than the robot is that robotic labor is likely to eventually be orders of magnitude less expensive than human labor. A proprietor deciding whether to buy robots or hire humans can dramatically reduce their costs by hiring robots, which means the restaurant or bar with robot labor will have much lower prices than the one with human labor. Customers will have to pay a hefty premium to go to a restaurant or bar with human servers. And some restaurants may have 1 waiter/waitress and many robotic servers, and if you want human wait staff, you pay a hefty surcharge. Some will choose the human option - many won't.
Similar for manual work such as construction.
Why? Robotic construction work will be far cheaper and far less error prone. Not to mention 3D printed construction is already dramatically increasing home construction speed and reducing costs. But even for a fully custom home, robot labor would be cheaper and faster.
Being a capitalist and consumer country is suspect more new jobs will appear, new industries, new ways to make money and spend money that don’t exist yet.
I strongly agree, except that the new jobs could easily be taken by AI and robots. Great if you are holding broad market equity - not great if you depend on human labor for your income.
The issue is not what AI can do today. It is what AI will be able to do in 1-3 decades, once AGI (Artificial General Intelligence) and ASI (Artificial Super Intelligence) are achieved combined with advances in robotics that we are already seeing creating robots with dexterity rivaling humans.
Don't believe AI/robots can take any job eventually?:
This is where full FI is a very strong defense, but Coast FI or Barista FI are still somewhat vulnerable (though likely have more time with service or manual labor jobs than with most white collar jobs).
An important perspective for me is that I always need to be able to adapt. I will plan toward FI but I understand that things may change and that I may have to go back to work. The future is not guaranteed. Just like the asterisk that they put "Past returns are no guarantee of future performance".
So things may change, but the future could also be great and my FI plan could totally work out or do even better than I anticipate.
A lot of people seem to have forgotten the Uber/Tesla self-driving hype from 10 years ago. I had software engineering coworkers who seriously believed that major metro areas were going to ban all human drivers by 2030. It was as ridiculous back then as the idea of AI replacing baristas is ridiculous now. For actual implementation reasons! Even if the tech worked perfectly! Not because I'm a "hater" or because "the Luddites" are right.
On the contrary if you believe all the hype then you should become a barista sooner! You will have substantial barista experience by the time the alleged hypothetical hordes of displaced white-collar workers will be restarting their new barista careers from scratch.
Fully autonomous (no safety driver) vehicles (both Waymo and trucks) are already on the public roads and highways today. Eventually, they will be far safer than human drivers, so while human drivers might not be fully prohibited, insurance rates for human drivers may become prohibitively expensive for many compared to insurance for autonomous vehicles, making human driving a novelty. Not 2030, but eventually.
Baristas are already being replaced by machines. Sure, some may eventually be left, at a premium price, but the majority could easily be replaced due to significantly higher cost for the human barista made coffee vs the machine made.
There are already very good cappuccino/espresso self-serve machines that even grind the beans on demand. These already exist in places a human barista may otherwise have been hired - for instance on cruise ships (where labor costs, by the way, are already much lower than US labor rates).
It's always "someday." Make a concrete prediction or it's just more weirdly defensive wishcasting.
Machine-made coffee has been available for years. AI has nothing to do with it.
Not all automation is AI, and not all AI is automation.
Make a concrete prediction […]
It is not possible to reliably predict what will happen after the AI singularity. Reasonable predictions of the future are based on probabilities and risks. Nobody can definitively predict the future and no plan should depend on a single concrete prediction of the future - rather it should be robust/flexible with respect to a range of possible futures.
Of course machine-made coffee does not require or typically use AI. I never said it did. My point was that baristas are easily replaced by automation and economics will drive the degree to which it will happen. Robotic bartenders also already exist as a novelty.
How much of this really comes down to what kind of Coast FI work do you want to do?
If you want to do your white collar office job remote on a contract type basis then that is probably prime to be taken by AI or outsourcing.
If you were planning on doing something that requires a human element like working two days as a nurse it is a different story.
This really just boils down to how much you believe in AI taking jobs.
When it has come to this idea of Barista FI I always struggle with it. Every time that I have run the math it always just seems like drastically decreasing your wage in the middle of your peak earning years makes little sense. For simple numbers you could make 100k in a standard job or 25k on a part time basis. For essentially two more years of work you could get the same amount as 8 years as a "barista". Not to mention you could continue maxing out the tax advantaged accounts and saving elsewhere.
Isn't this just another excuse/reason for one more year syndrome? Isn't he essentially asking the same question that so many people ask when it comes to pulling the trigger on the next step of your plan (whether that's quitting work entirely or only pursuing the work that lights you up? There is always something new that is shaking up the way business is done that potentially could cause people pursuing FI to rethink the timing of their next step in their journey. AI is just the current/most recent disruptor. When pursuing any type of FI, we need to be flexible so we can pivot if/when necessary. There will always be something that makes us look at our current situation and think, "I need to work a few more years because I don't have enough yet."
Yes and no. Full FI truly is the best defense against the possibility of massive AI triggered job loss. Coast or Barista FI are partial defenses.
But there are reasons to believe this time really is different (I know how cliche that sounds and I already know the counterarguments), but AI has the potential to change the world as profoundly as human harnessing of fire (the other type 😉- combustion). We've only scratched the surface of the potential AI caused changes. Never in human history have humans created something that has the potential to supplant us as the dominant species on the planet. Comprehending the impact of future Artificial General Intelligence (AGI) or Artificial Super Intelligence (ASI) is well beyond what AI can do today. See my earlier response to Joe in this thread.
Could it be an excuse for one more year syndrome? Sure (that's effectively what Nick concluded in the blog post), but is this a substantial risk that full FI addresses but Coast FI only partially addresses? Yes, it is.
Geniune question: if you think AI is the single most important human discovery and that it will completely reshape our world, what are you doing about it?
Certainly some action is required, no?
Personally, I've achieved full FI, so I am no longer dependent on earned income. I was lucky enough to achieve FI before AI unemployment risk became a significant concern. I would encourage people on the path to FI to get there sooner, rather than later if it doesn't cause hardship to do so, as a mitigation to this risk.
On a larger scale, post FIRE, I have actively volunteered in a few AI safety initiatives, including providing input to US NISTs US AI Safety Institute (now renamed, but still active) and I try to raise awareness of AI transformation issues when I can (such as this thread).
I am encouraged that there is a sufficient population level of concern about AI risks (I believe AI will provide enormous benefits and comes with enormous risks) that the odds are decent that AI risks will be managed / taken seriously, rather than ignored. I would be much more concerned if population level concern of AI risks were low.
Yes. It's basically just another expression of "The Possibility Effect", which summarizes a number of cognitive biases that lead people to equate mere possibilities with actual probabilities.
That dressed up with an attractive Narrative Fallacy leads to a lot of poor decision making. Think of Rain Man who would not fly on any airline that had experienced a fatal crash, even just once. This is what much of financial services marketing and financial media is based on – the latest and greatest scare stories.
Whether or not AI (in the form of AGI and ASI) will trigger sustained structural unemployment (not merely retraining) displacing a large percentage of workers, or worse, reflects major disagreement among AI experts, with upwards of 20% of AI experts predicting substantial risk of broad structural unemployment. AI expert advocates of that position include Geoffrey Hinton, often referred to as the (or sometimes a) "godfather of AI."
I don't know the specific probability of that outcome occurring, but it is not an insignificant possibility. To claim it is an insignificant possibility - given a significant minority of AI experts espousing substantial risk of that outcome - exhibits the congnitive bias of ostrich effect.
Standard FUD. A recent study by McKinsey reported that 97% of CEOs admit the claimed benefits of AI haven't materialized, and the best use is to make existing employees more efficient, not to replace them. Also, many companies have started reporting that AI is so expensive (and it's still mostly subsidized by VCs!) that it's cheaper to ditch AI and hire more staff.
What you are referencing from the McKinsey report is a snapshot of AI today, and I agree with that characterization of today.
What I am talking about is likely one or more decades in the future if and when Artificial General Intelligence (AGI) or Artificial Super Intelligence (ASI) is achieved. Exponential advancement is very hard for humans to comprehend, but what we are seeing in today's AI capabilities were unimaginable by most people until this year or last.
Just an example: Mythos has caused the US Treasury Department to convene financial institutions to protect themselves from hundreds/thousands of vulnerabilities which Mythos can easily find that humans take orders of magnitude longer to find. This type of thing will happen more and more as AI advances. Sure, there will be some overhype, but the rate at which AI is advancing is far beyond any technological advance in the past. Just compare the accuracy and nuance of one LLM from 1 or 2 versions back, even in the same year.
Also, see my 2nd reply earlier in the thread to Joe, especially the YouTube video linked there - "Humans are becoming horses" (previously called "Humans Need Not Apply"). Is massive net disappearance of human jobs inevitable? Not necessarily. But is it a reasonably likely eventual result of AI worth taking seriously as a risk? Definitely.
I'm not sure I agree with McKinsey from a boots on the ground perspective. Things are radically changing and code development is accelerating. I think everyone is learning that there's more to software than code, but from what I'm seeing every part of the chain is going to improve, and things are already changing rapidly.
I don't really have an opinion whether jobs will disappear. I am cognizant that when doors close other doors will open. But I think this technology is unprecedented, and we don't know the full extent of the disruption. I think saying millions will lose jobs is just as much fortune telling as is saying that everyone will stay employed, even factoring in some bumpiness. Probably the truth lies somewhere along the spectrum.
As blithe as it is to say, the only thing I can say for sure that things will not be the same, again.
I would agree that mass unemployment is not inevitable (though even if a single industry segment like truckers is fully displaced, it would eliminate millions of jobs).
My point is that the risk is high enough that targeting achieving full FI before the possibility of mass unemployment is worth factoring into FI planning.
It is also possible that AI will herald sufficient abundance that no one needs to work for a living or to provide for their retirement. So I am not claiming mass unemployment with commensurate poverty is an inevitable outcome.
100% AI is changing how I work on a daily basis. I can see both sides to the job agreement, not sure which way it will go
I found this interesting. Near the end
What is interesting to me is that a friend was planning on retiring around 2010, at about age 60. So, a little "early" by traditional standards, but not by FIRE standards. After the great financial started to take shape he decided to continue working. I don't recall precisely when he retired, but my vague recollection was that it was until sometime around 2015, 2016, or maybe 2017.
And he was in a great position financially speaking. Married couple that were both high earners and never had kids. But the fear was very real and caused him to work for a number of years more than planned.