Preparing for when the machine stops
idiallo.com
idiallo.com
Both are great and horrifying at the same time.
In this story, everyone relies on automation so much that they forget how to do mathematics. A lone, nerdy operator re-learns the skill of 2 digit multiplication in the process of working through automated weapons targeting issues, astounding his superiors.
The concept of forgotten arts seems like a common thread here.
This is a PDF of the text. https://ia600806.us.archive.org/20/items/TheFeelingOfPower/T...
This is the Wikipedia overview. https://en.wikipedia.org/wiki/The_Feeling_of_Power
First heard this one at school.
This thought exercise also gets interesting with software development. There are few prospects available for new grads and junior devs now. Relentless offshoring, headcount reductions, and AI promises from CEOs have hollowed out the tech landscape. There are very few opportunities for young people to professionally develop the knowledge to keep the systems running.
And now most startups are focused on eliminating staff via AI. The people who would keep the systems running. I'm not sure where all of this leads to in a few years.
What's lacking is incentive and drive: there are just too many other shiny tools to distract us, including jobs that sound more glamorous and/or pay just as well.
A young person with the drive to learn just about anything has the means to do so, if they really want.
W.R.T. any large industry crash, no one's going to care if you can do your Angular (or specific tech skill) stuff in 1 days vs. 5 days, so why the emphasis on speed in this scenario? Both System 1 or 2 thinking is fine here.
We've enjoyed some pretty great technical advances for the past 40 years, even with a 10 year "dark age" we're still net ahead. We can rebuild and relearn a lot of stuff in 10 years.
If a manager doesn't know anything about what their employees are working on, they are basically fucked. That much holds up with LLMs. The simple stuff mostly works, but the complex stuff isn't going to pan out for you, and it will take a while to figure out that's the direction you went in.
The best answer is the one that fits the aesthetics of my approach--one that didn't exist before (there was only the problem before), but the answer is simple, straightforward, or adaptable.
Having multiple answers is good because different minds evaluated the question. It is a buffet of alternatives, starting from others' first principles, mistakes, and experience. Some are rejected outright from some tacit taste organ. Others become long-lived browser tabs, a promise to read carefully someday (never).
All this is void if it turns out using SO is similarly degenerative in the same way, though.
That way we can quickly assimilate knowledge from the AI and theoretically always have at least as much knowledge as the AI.
I suppose it also means that we can verify that the AI is not lying to us.
I used to think about this in math class. I could figure out what to do with my calculator most of the time, but I didn’t have any intuitive sense of how things worked. The sine, cosine, and tangent functions are still just black boxes to me, I have no idea what they actually do or how I would calculate their values. I often daydreamed about finding myself on a desert island, needing to make use of trigonometry to rebuild civilization, but not being able to find the angles that I needed.
Lots of other skills are lost this way. I don’t know how to join wood or sew a stitch, but I do know how to operate a nail gun and work a sewing machine. I couldn’t fix either device, but if I couldn’t find anyone to fix them and couldn’t obtain new ones, I would likely have bigger problems to worry about. Most people will only ever need to view these devices as black boxes; the benefits of specialization generally offset the costs introduced by abstraction, absent major market disruptions (e.g. supply chain breakdowns, changing regulatory frameworks, etc.). Most people in human history have spent their lives as generalists on a farm. This hedges the individual against a lot of risk (the generalist can likely always find some work to do), but the real strides in risk management are made by specialists living in urban settlements.
It's why "preppers" buy an armory of weapons, but don't make friends with their neighbors, become contributors to their local town or advocate for infrastructure improvements or sustainable farming policies.
Interestingly, when I've seen interviews with deeply committed preppers, they almost always seem to come around to the conclusion that community is the most significant factor in their planning.
I find it fascinating how that juxtaposes with the (possibly well earned) cliche of preppers as intensely individualist libertarians who reject society. I suspect there's some relationship here to the idea that if you go far enough to the left or right, you find that the spectrum is circular and not linear.
They are functions which tell you how to relate different parameters of a triangle. Concepts in mathematics are often interconnected, and trig functions appear in a lot of other (interesting) contexts as well, but fundamentally I don't see them as anything "more" than the SOHCAHTOA you were taught at school.
> or how I would calculate their values.
Without a lookup table? You would need some kind of a way to express the functions in terms of other mathematical functions which you know how to do, like multiplications and additions. Sometimes you can do this with a series expansion. Computers sometimes use a variant of the CORDIC algorithm. Both of those things are clever ideas in their own right, but you don't need to understand them to know what trig functions do.
> I often daydreamed about finding myself on a desert island, needing to make use of trigonometry to rebuild civilization, but not being able to find the angles that I needed.
If you had a circle, a ruler, a pen and some paper then you'd get the idea to make yourself a lookup table pretty quickly!
I think the point I'm making is that we get used to assuming that knowledge is so deep and complex a thing that we can never really know anything. But often, knowledge isn't as intricate as that. If everyone forgot about trig functions today, they'd be rediscovered tomorrow.
Most of what I've done is in software, I could not build computer from electronic parts, not even full adder from memory. Maybe I could read some schematics, but most measurement devices use chips as well, it would very difficult.
I think preparation should be skills, useful in any environment.
> Learning to learn is a noble idea. But more important is learning to unlearn, and knowing when to resist the comfort of automation.
I feel strongly about “teaching people to teach themselves” which seems a direct analog to “learning to learn”, but I am at a complete loss for what “learning to unlearn” means, especially as it relates to resisting automation.
Is the idea that you need to “unlearn” wanting to “learn” automation so you can keep “learning” more deeply about things you have already “learned”?
The example is unrealistic because that's a very easily anticipated single point of failure.
We've been building systems without single points of failure for millennia.
There will never be only one machine. Already in AI we have dozens of models. They come from multiple diverse - unaffiliated! - cultures, countries, and ideologies.
It is not and never will be a monoculture. Not at the level of which machine is which and not at the level of which machine can help understand the workings of which other machine. It’ll be many machines augmenting our capabilities in many ways.
A pencil is a powerful tool that is horrendously complex thing to make - so many steps to create one. Yet they’re plentiful and we take them for granted.
They're gonna pay the price for their foolishness, and paying the price is the universe's mechanism for keeping itself in balance.
Take what's going on the US for example...
The price will be born by all of us.
You probably meant borne, but I do agree we're all somewhat responsible too...
No doubt that is true, but...
Suppose an extortionist comes to my business and says "hey, that's a nice business you got there, it would be a shame if someone was to ruin it with a bunch of horseshit tariffs"
I'd say 'yes sir no problem" and then do my goddamn best to murder those fucks in their sleep. Making them pay the price for extorting me will make sure that its not gonna happen again.
So now suppose the business next door is being extorted. And I tell the owner that I'd be happy to murder the extortionists for them. But they say "nah dog, they're sending the bad guys to El Salvador, it'll be fine".
How do expect me to care about that guy? Fuck that guy.
"Love your neighbor as yourself" exists in some form in every major pre-modern religion and philosophy because it's a cornerstone of society.
Not over-compassionate, truthless love or yet another government program, but caring about those we run into enough to--at minimum--personally help meet their basic needs and defend or rescue them from immediate threats to their safety.
If we're going to let anything go, it shouldn't be that.
Seriously, though… How will you “carve out a little island”? Computers are the kingdom of AI.
I went to Toronto Metropolitan University (ranked 800-1000 globally) and actually designing that basic CPU was too hard for most students, so they just gave us VHDL code to copy & paste into Quartus II.
Their VHDL code had bugs in it which I discovered after doing the project myself from scratch and comparing my answer with the "correct" one.
This has been the policy for many years now. Shockingly our co-op placement rates have dropped from 70% to 30%.
[0] https://cwur.org/2024/rensselaer-polytechnic-institute.php
Most people don’t know how to grow a potato. So?
Yes, horses were replaced, but go outside of city and there are many horse farms, and some of them offer horse riding classes to general public. Should something happen to the cars, the knowledge will come back. Same way, even if an average JS programmer has no idea what "differential pair" stands for, there are plenty of people who _know_ what this is, and can at least read the eye diagram.
Personally, I'm pretty concerned about both. The fact that many people don't have basic survival skills like sourcing their own food, safe drinking water, and heat. And the fact that that many people lack basic thinking skills: ability to detect misinformation, or deal with the challenges and inaccuracies of flaky AI.
In an ideal world society everyone who is capable has a higher level of training in both. But under modern oligarchic capitalism there are advantages to ensuring that people have neither skill: survival or thinking.
I feel that's been an issue for a long time, with the heavy reliance on dependencies. AI tools are really just a furtherance of the model already in use.
I have the luxury of developing software as a craft; not as a vocation. I deliberately do stuff "the hard way," because I feel more fulfilled, doing so.