I think if your job is to assemble a segment of a car based on a spec using provided tools and pre-trained processes, it makes sense if you worry that giant robot arms might be installed to replace you.
But if your job is to assemble a car in order to explore what modifications to make to the design, experiment with a single prototype, and determine how to program those robot arms, you’re probably not thinking about the risk of being automated.
I know a lot of counter arguments are a form of, “but AI is automating that second class of job!” But I just really haven’t seen that at all. What I have seen is a misclassification of the former as the latter.
A software engineer with an LLM is still infinitely more powerful than a commoner with an LLM. The engineer can debug, guide, change approaches, and give very specific instructions if they know what needs to be done.
The commoner can only hammer the prompt repeatedly with "this doesn't work can you fix it".
So yes, our jobs are changing rapidly, but this doesn't strike me as being obsolete any time soon.
This is actually a great way to foster the learning spirit in the age of AI. Even if the student uses AI to arrive at an answer, they will still need to, at the very least, ask the AI to give it an explanation that will teach them how it arrived to the solution.
No this is not the way we want learning to be - just like how students are banned from using calculators until they have mastered the foundational thinking.
There is research that shows that banning calculators impedes the learning of maths. It is certainly not obvious to me that calculators will have a negative effect - I certainly always allowed my kids to use them.
LLMs are trickier and use needs to be restricted to stop cheating, just as my kids had restrictions on what calculators they could use in some exams. That does not mean they are all bad or even net bad if used correctly.
> There is research that shows that banning calculators impedes the learning of maths.
Please share what you know. My search found a heap of opinions and just one study where use of calculators made children less able to calculate by themselves, not the ability to learn and understand math in general.
That's a fair point, but AI can do much more than just provide you with an answer like a calculator.
AI can explain the underlying process of manual computation and help you learn it. You can ask it questions when you're confused, and it will keep explaining no matter how off the topic you go.
We don't consider tutoring bad for learning - quite the contrary, we tutor slower students to help them catch up, and advanced students to help them fulfill their potential.
If we use AI as if it was an automated, tireless tutor, it may change learning for the better. Not like it was anywhere near great as it was.
Yeah; it gets steps 1-3 right, 4-6 obviously wrong, and then 7-9 subtly wrong such that a student, who needs it step by step while learning, can't tell.
Props to the teacher for putting in the work to thoughtfully grade an AI transcript! As I typed that I wondered if a lazy teacher might then use AI to grade the students AI transcript?
That's roughly what we did as well. Use anything you want, but in the end you have to be able to explain the process and the projects are harder than before.
If we can do more now in a shorter time then let's teach people to get proficient at it, not arbitrarily limit them in ways they won't be when doing their job later.
I think it's a bit like the Dunning-Kruger effect. You need to know what you're even asking for and how to ask for it. And you need to know how to evaluate if you've got it.
This actually reminds me so strongly of the Pakleds from Star Trek TNG. They knew they wanted to be strong and fast, but the best they could do is say, "make us strong." They had no ability to evaluate that their AI (sorry, Geordi) was giving them something that looked strong, but simply wasn't.
Yep, I've seen a couple of folks pretending to be junior PMs, thinking they can replace developers entirely. The problem is, they can't write a spec. They can define a feature at a very high level, on a good day. They resort to asking one AI to write them a spec that they feed to another.
People have tried that with everything from COBOL to low code. Its even succeeded in some problem domains (e.g. thing people code with spreadsheet formula) but there is no general solution that replaces programmers entirely.
My job is to make people who have money think I'm indispensable to achieving their goals. There's a good chance AI can fake this well enough to replace me. Faking it would be good enough in an economy with low levels of competition; everyone can judge for themselves if this is our economy or not.
I don’t think this is the issue “yet”. It’s that no matter what class you are, your CEO does not care. Mediocre AI work is enough to give them immense returns and an exit. He’s not looking out for the unfortunate bag holders. The world has always had tolerance for highly distributed crap. See Windows.
This seems like a purely cynical lacking any substantive analysis.
Despite whatever nasty business practices and shitty UX Windows has foisted on the world, there is no denying the tremendous value that it has brought, including impressive backwards compatibility that rivals some of the best platforms in computing history.
AI shovelware pump-n-dump is an entirely different short term game that will never get anywhere near Microsoft levels of success. It's more like the fly-by-nights in the dotcom bubble that crashed and burned without having achieved anything except a large investment.
You misunderstand me. While I left Windows over a decade ago, I recognize it was a great OS in some aspects. I was referring to the recent AI fueled Windows developments and Ad riddled experiences. Someone decided that is fine, and you won't see orgs or regular users drop it...tolerance.
You are describing tradition (deterministic?) automation before AI. With AI systems as general as today's SOTA LLMs, they'll happily take on the job regardless of the task falling into class I or class II.
Ask a robot arm "how should we improve our car design this year", it'll certainly get stuck. Ask an AI, it'll give you a real opinion that's at least on par with a human's opinion. If a company builds enough tooling to complete the "AI comes up with idea -> AI designs prototype -> AI robot physically builds the car -> AI robot test drives the car -> AI evaluates all prototypes and confirms next year's design" feedback loop, then theoretically this definitely can work.
This is why AI is seen as such a big deal - it's fundamentally different from all previous technologies. To an AI, there is no line that would distinguish class I from II.
This is actually a really good description of the situation. But I will say, as someone that prided myself on being the second one you described, I am becoming very concerned about how much of my work was misclassified. It does feel like a lot of work I did in the second class is being automated where maybe previously it overinflated my ego.
SWE is more like formula 1 where each race presents a unique combination of track, car, driver, conditions. You may have tools to build the thing, but designing the thing is the main issue. Code editor, linter, test runner, build tools are for building the thing. Understanding the requirements and the technical challenges is designing the thing.
The other day I said something along the lines of, "be interested in the class, not the instance" and I meant to try to articulate a sense of metaprogramming and metaanalysis of a problem.
Y is causing Z and we should fix that. But if we stop and study the problem, we might discover that X causes the class of Y problem so we can fix the entire class, not just the instance. And perhaps W causes the class of X issue. I find my job more and more being about how far up this causality tree can I reason, how confident am I about my findings, and how far up does it make business sense to address right now, later, or ever?
is it? I really fail to see the metaphor as an F1 fan. The cars do not change that much; only the setup does, based on track and conditions. The drivers are fairly consistent through the season. Once a car is built and a pecking order is established in the season, it is pretty unrealistic to expect a team with a slower car to outcompete a team with a faster car, no matter what track it is (since the conditions affect everyone equally).
Over the last 16 years, Red Bull has won 8 times, Mercedes 7 times and Mclaren 1. Which means, regardless of the change in tracks and conditions, the winners are usually the same.
So either every other team sucks at "understanding the requirements and the technical challenges" on a clinical basis or the metaphor doesn't make a lot of sense.
I wonder about how true this was historically. I imagine race car driving had periods of rapid, exciting innovation. But I can see how a lot of it has probably reached levels of optimization where the rules, safety, and technology change well within the realm of diminishing returns. I'm sure there's still a ridiculous about of R&D though? (I don't really know race car driving)
Sure there is crazy levels of R&D but that mostly happens off season or if there is a change in regulations which happen every 4-5 years usually. Interestingly, this year the entire grid starts with new regs and we don't really know the pecking order yet.
But my whole point was that race to race, it really isn't that much different for the teams as the comment implied and I am still kind of lost how it fits to SWE unless you're really stretching things.
Even then, most teams dont even make their own engines etc.
Do you really think that rainy Canada is the same as Jedddah, or Singapore? And what is the purpose of the free practice sessions?
You’ve got the big bet to design the car between the season (which is kinda the big architectural decisions you make at the beginning of the project). Then you got the refinement over the season, which are like bug fixings and performance tweaks. There’s the parts upgrade, which are like small features added on top of the initial software.
For the next season, you either improve on the design or start from scratch depending on what you’ve learned. In the first case, It is the new version of the software. In the second, that’s the big refactor.
I remember that the reserve drivers may do a lot of simulations to provide data to the engineers.
Most projects don’t change that much either. Head over to a big open source project, and more often you will only see tweaks. To be able to do the tweaks require a very good understanding of the whole project (Naur’s theory of programming).
Also in software, we can do big refactors. F1 teams are restricted to the version they’ve put in the first race. But we do have a lot of projects that were designed well enough that they’ve never changed the initial version, just build on top of it.
> I know a lot of counter arguments are a form of, “but AI is automating that second class of job!”
Uh, it's not the issue. The issue is that there isn't that much demand for the second class of job. At least not yet. The first class of job is what feeds billions of families.
Discussing what we should do about the automation of labour is nothing new and is certainly a pretty big deal here. But I think you're reframing/redirecting the intended topic of conversation by suggesting that "X isn't the issue, Y is."
It wanders off the path like if I responded with, "that's also not the issue. The issue is that people need jobs to eat."
Kind of tangentially related to weird ways tech works: a few weeks ago I finally disassembled my original DMG-01 Game Boy to fix it. There was decades of battery acid corrosion that took a ton of cleaning and resoldering and reflowing the screen connections.
After hours and hours of iterations I could get it to work perfectly, just once, for each cartridge. I would clean it a bit more, try a game, things would work great. I’d try another game and it the copyright logo would fail. So I’d clean it up a bit more. Swab the port and try it again. It worked! Then another game… nope.
I eventually realized that the isopropanol was making a weak connection work fine, and then I guess it just kept working once power was flowing.
No matter what I tried I couldn’t get it to stay fixed, so I keep a handful of cotton swabs and a small dispenser of isopropanol in the carrying case. I’ll swab a cartridge before inserting it and it works every time.
So now I have a Game Boy that requires alcohol to work.
I think I’m old enough to have experienced this cycle so many times with so many businesses that I just feel kind of silly to hate on Apple or Microsoft or whoever. They’re all just maximizing profits as designed.
I think people find it easier to scowl at the villain du jour than to dig into the deep complex issue of when capitalism doesn’t work, when the government isn’t doing enough, and what we could do about it… or the feeling that we really can’t do much.
That's why people don't dig into the deep complex issues. Because it's uncomfortable, and forces one to confront the potential reality that their worldview, and everything they've known about how our society works is wrong, broken, and collapsing in front of them.
It can be a very distressing and depressing state of mind. There's a reason "ignorance is bliss" is a common trope, because there's some real truth to it. For some, it's better for emotional and mental wellbeing to ignore the problems of reality and remain ignorant.
> For some, it's better for emotional and mental wellbeing to ignore the problems of reality and remain ignorant.
I think it isn't just some, it's effectively everyone, the nature of being human. Instead, there's a group of people who are willing to sacrifice their emotional and wellbeing to face these problems of reality, and try to use the limited power they do have to improve them, for the greater good.
We can do a lot if we pressure the company or the regulations around it. Maybe not right now in this current regime, but tides will shift.
The issue is that people's attention spans on this are much too short. The fervor around this may not even last to the end of this month, let alone until a change in power allows a new administration to properly go after the company.
You don’t need to solve the problems of capitalism to call bullshit bullshit. Saying “companies maximize profits” doesn’t magically make the behavior acceptable and when Apple does this, it’s not just “the market at work,” it’s the use of market power.
Complaining about it is part of the system operating the way it operates. It’s factored in already. I just think that it’s not really interesting. It’s reasoning about the instance, not the class.
Sounds like Americans are in general fine with all of it. Voting patterns hold. General sentiment still remains aligned with the status quo. There does not seem like there are any consequences for the representatives to not represent the people.
This sounds like my experiences in Toronto. It’s less adversarial than the experiences I've had in the U.S.
My experiences were basically a form of, “Hey we saw something that caught our attention and might be an issue. Let's work through addressing this."
One case it was a handful of 3.5" galvanized nails. "Whoops. Okay, so, this bag used to be my makeshift toolbag. My other one ripped and I had to get one last minute--" "No problem. Can you remove them? You can either surrender them to us or we can get them mailed back to you, but I'm guessing it's not worth it..." I was so defensive because to me it looked bad but they weren't actually after me in the way I thought they'd be.
The second time was that I had an "Arduino Starter Kit" full of bundled up wires and random chips and such. Once they saw the box they didn't even ask me to un-shrinkwrap it, and unlike the nails, didn't re-x-ray the bag.
Both times they rotated their screen and pointed to the box framing the item in question on the colourized x-ray.
Meanwhile, the TSA looks at me like I'm, at best an annoyance, and at worst a criminal, when I ask them to inspect my camera kit manually (film, not digital). And that inspection consists of swabbing 35mm film canisters - like, the shell of a 35mm roll is going to tell them anything useful?!?! It's a complete sham.
I guess they're probably operating on the assumption that at worst a few short nails stuffed in a small film canister are no worse than the metal handle from a rolling suitcase.
The swab is for common explosives. The canisters are a bit on the small side but I guess could still pose a threat if packed with high explosive and a bit of shrapnel.
The apparent annoyance (or worse) is the part that gets me. The entire process just feels needlessly adversarial. At least they didn't insist on patting you down or emptying out your bag!
I think for film specifically it might be for drugs? Seems like a very convenient way to smuggle contraband. You can’t open it to inspect it, you can’t xray it either otherwise it will ruin the film.
Worst and most aggressive pat down I have ever experienced was in Toronto for no reason that I can think of, so I have learned to be stoic about all interactions with gate keepers, regardless of country. You never know when someone had a bad cup of tea just before the met you.
New York is the worst security I've ever come through for just being needlessly horrible. Like screaming at people because they didn't literally put their feet on the "footprints" on the floor.
Toronto was fine. Just a slightly incredulous conversation about how we could take 3 weeks off to travel Canada.
Especially if you've been in New York for a few days, being yelled at shouldn't be taken so personally. Especially when you consider how many people badly need instructions yelled at them because they're so very confused, I can see why they do it!
While there's U.S. Customs agents in Pearson, the entirety of security is done by CATSA. I cannot imagine U.S. Customs doing any sort of pat down. I'm not sure they'd even be allowed to do anything like that in Toronto. I think they're pretty much only allowed to screen and admit or reject.
That's been exactly my experience recently in the US. Most recently it was some Hot Hands hand warmers. They just had me go to the end of the line where you get your bags ouf of the scanner and the agent brought my bag down there on the other side of the rollers. They set it on the table in front of me, and there was a monitor above the table where they pointed to the hand warmers on the screen. They said something along the lines of, "Looks like you might have some hand warmers in the main pocket, would you mind taking them out?" I pulled them out, showed them to them, they thanked me and I put it back in the bag and went on my way. This was in Juneau, AK.
Hard to know what accounts to bother responding to these days. This is likely one of them as it fails to offer any worthy substance beyond a barely whined grievance. But I have first hand experience that the things described in this post are absolutely not gone.
I suspect there are agents of lesser minds at work hoping to stir instability. We aren’t swindled as easily as other peoples.
I imagine it’s far more economical to have one foundry that can make a general purpose chip that’s overpowered for 95% of uses than to try to make a ton of different chips. It speaks to how a lot of the actual cost is the manufacturing and R&D.
The only real problem I could see is if the general purpose microcontroller is significantly more power-hungry than a specialized chip, impacting the battery life of the earbuds.
On every other axis, though, it's likely a very clear win: reusable chips means cheaper units, which often translates into real resource savings (in the extreme case, it may save an entire additional factory for the custom chips, saving untold energy and effort.)
I have a 2020 Forester and I've come to describing it as "I no-longer drive on the highway, I manage the car." Sometimes I'll get nervous and take over. But even in stop-and-go traffic, it has behaved perfectly.
My only complaint is that there's an over-eager PID loop with lane keeping. If I want to pass a transport truck and want to kind of edge to the left of my lane when doing so, it will keep trying to compensate, which I can feel in the wheel, so I compensate for as well. And if I let go of the wheel and let it win, it suddenly flings me towards the right side of the lane.
I suspect this is because it isn't programmed to think that I'm making adjustments, it probably just thinks there's some weirdness in the vehicle dynamics/road characteristics that requires extra compensation.
I have a 2023 Crosstrek, my wife has a '21 Ascent. I have the same habit you do - edging away from large trucks slightly - and both of them do the same thing you described to me.
It's essentially that Subaru's lane system actually has two levels: it has lane keeping where it's just trying to keep you inside the lines, and then on top of that it also has lane centering which is pretty much what it says.
Just a note for you or anyone reading who has a recent Subaru and doesn't know already: if you find the centering really bothersome, you should be able to be able to go into the settings on the instrument cluster display (up/down arrows at the lower left behind the wheel, toggle it until you get to the "hold for settings" option), find the Eyesight settings, and turn off lane centering. It will still try to keep you inside the lane markers but won't try to park you right in the center of the lane. In that mode, it's more like the Honda Sensing system I had on my 2016 Civic.
I go back and forth a bit on it but mostly keep it in lane centering mode now - I've gotten used to how it positions the car in the lane, and it lets me focus more on what's going on around me than micromanaging lane position and such.
> It's essentially that Subaru's lane system actually has two levels: it has lane keeping where it's just trying to keep you inside the lines, and then on top of that it also has lane centering which is pretty much what it says.
Same with Hyundai except they call them "Lane Keeping Assist" (LKA) and "Lane Following Assist" (LFA) and I have trouble remembering which one centers you and which one just keeps you from leaving the lane.
To me just based on the names I'd have expected keeping to be the one that actively positions you (it keeps you centered) and following to the one that just reacts when you are going to depart the lane (it keeps you following the lane).
Mostly now I just remember that the one that comes on automatically any time I'm going 40+ mph is the reactive one, and the one that I have to explicitly turn on is the centering one (although both come on automatically on certain highways based on data from the navigation system).
idk whether subaru is exact the same as hyundai but i basically turned lane centering off on my hyundai. when possible i only use radar cruise control, and lane follow. if i want to overtake, turn on my signal and it'll automatically safely increase speed to set cc speed and let the lane follow off. it's pretty seamless.
lane centering is a bit too annoying for me, i need to keep my hands on the wheels anyway.
> I have a 2020 Forester and I've come to describing it as "I no-longer drive on the highway, I manage the car." Sometimes I'll get nervous and take over. But even in stop-and-go traffic, it has behaved perfectly.
I drive an old beater from 2001, but... I really don't think I understand why people want these in-between not-quite-autopilot features? To me it's like, it would be one thing if you could completely turn your brain off, or look at your phone, or rest. But since you can't, it seems like this stuff makes it more difficult to pay the appropriate amount of attention? For me, if I'm already driving somewhere, and have to pay enough attention to know if an emergency is about to happen, I might as well just do the driving.
As a technologist, I like lane-keep assist because it feels fundamentally more right that my car by default follows the road than keeps going with the turn radius I had previously input.
Cruise control with minimum distance helps me keep a sound distance even as other cars keep packing up and reducing distances on a busy highway. My previous car (Mercedes) was great at detecting if a new car coming in front of me was accelerating, if so it didn't adjust the distance as aggressively. Much better behavior than my current Kia.
Auto-break features are sweet as they react really fast. If that can avoid deploying an airbag in my face, I'm all for it.
I agree it's a lot like managing, with six buttons just to do the above, but from a bottom-up approach, each feature has value in its own right.
> For me, if I'm already driving somewhere, and have to pay enough attention to know if an emergency is about to happen, I might as well just do the driving.
Where do you draw the line? Would you prefer not having a steering and brake servo? Would you prefer sticking out your arms instead of having flashing lights? Would you prefer feeling every bump in the road to having suspension?
To me these systems just feel like natural evolution of the car concept, something that's been going on for 120 years. What Tesla failed at was putting their heads in the clouds and hoping something awesome would eventually pop out the other end. While the established car makers did incremental improvements.
GP said s/he didn't understand why anyone would want these in-betweens. I gave an explanation as to why.
Based on what you're saying, it seems the divide arises from some drivers classifying these features as physical comfort, and some as mentally disengaging.
The cognitive load is greatly reduced when using these features. Honestly, adaptive cruise control in the city is a godsend. Not having to deal with watching speed . start and stop traffic is also automated for me. Driving on a highway is also great .. You can drive much further without needing a break.
Same. Even cruise control is kind of useless because people in front of you don't necessarily use it and are very inconsistent in their speed.
So you end up constantly having to engage/disengage, rendering the whole thing moot.
I think something like autopilot could be implemented at the infrastructure level (sensors and emitters along the road), but people wouldn't like that because it would mean being unable to set your speed or overtake.
The car exists for "freedom," but it is really an inefficient mode of transportation from both a time-use and energy-use perspective.
What we really need is a mix between rail/train and car/road.
ACC generally has a 3-4 second time interval that it permits between you and the car in front of you. I live in SoCal, so a lot of my driving is on very aggressive routes. The 4-second gap is mechanically safe but it's practically unusable because it creates a void large enough to invite other cars to lane change in front of me. So when that car merges in, the ACC detects a violation of the safe braking distance and decelerates to reestablish the gap. I call it the "cut me off" loop when we're on trips.
And before anyone suggests that I start tinkering around with the settings, I have adjusted it and the damned thing just resets itself constantly.
The beauty of ACC is it lets your disengage mentally. You can be aggro if you want to with it on, but I found it's just not emotionally worth it to get mad at being cut off anymore in a car with ACC. ACC just handles going forwards and I'm not having to touch gas nor brake. If I'm not touching either, I don't have to panic react to getting cut-off, just make sure the ACC is handling it, and if that's all I need to check, vs slam on the brakes, then eh.
Ah yes, I never used that. My car isn't very recent (about 10 years old now), and I drive very little (about 2-3k per year; I take the train to go anywhere far) because I hate it.
But the adaptive part would make it much more useful indeed.
However, something that is extremely annoying in France is that speed limits tend to change very often and abruptly.
I just think that trying to solve the problem solely at the car level is always going to have too many limitations...
I prefer to steer, but radar cruise control takes a lot of the frustration out of minor speed fluctuations in front of me on the highway. I don’t feel as much need to pass all the time.
Have a EX90 I got on a really great deal, we drove it cross-country and it was mind-blowing how little I had to do. If it didn't make you touch the wheel / pay attention we could have basically done the entire trip without incident minus off-ramps.
But there was one thing that was quite bad, similar to yours. While passing a semi I pulled it to the left side and it actually yanked us right so hard and then over-corrected once again. Super scary moment, the only issue of the whole trip, but basically never passed with it on again.
How you like that car? I test drove an early model that was really a pre-release dealer demo. It was a great ride but I also didnt get to do a whole lot with the sales guy next to me and a tight deadline to get back home.
Made me want to invest in Volvo. They fixed most of the software issues, what's left is a shockingly nice experience. Touch screen can be a bit annoying is all, but Google integration and design of the UX in the touchscreen is really good.
My volvo also has a "not perfectly tuned" PID loop. With "autopilot" engaged it keeps weaving constantly left and right inside the lane im in. Have gotten used to keeping a firm-ish pressure on the steering wheel at all times to compensate. But drivers behind me must have thought me drunk before i got the hang of it.
This is lane keep assist not lane centering and dangerous to use as a lane centering feature as it’s not designed to do that hence the ping ponging behavior
Lane keep assist is always enabled, this "copilot" or whatever it's called is an extra feature i can manually enable over the default lane assist. And it will steer and follow the road quite well in most conditions. But sometimes it starts the ping pong behaviour.
Should also be noted that i never take my hands of the wheel. And the volvo is quite fast at beeping at me if it detects that i dont hang onto the wheel.
Don’t most of these systems release the lane-keeping when you turn on the turn signal? Does yours not, or do you not signal until you are trying to exit the lane? (Genuinely asking.)
But if your job is to assemble a car in order to explore what modifications to make to the design, experiment with a single prototype, and determine how to program those robot arms, you’re probably not thinking about the risk of being automated.
I know a lot of counter arguments are a form of, “but AI is automating that second class of job!” But I just really haven’t seen that at all. What I have seen is a misclassification of the former as the latter.
reply