I think the Julius effect is a real phenomenon. There are certain personality types that excel at flattery and bluffing their way through situations. Some of these people are willfully scamming,...
I think the Julius effect is a real phenomenon. There are certain personality types that excel at flattery and bluffing their way through situations. Some of these people are willfully scamming, but I think a larger portion of them are earnestly unaware of their incompetence due to the unhelpful signals they receive from others, and some sort of relentless Dunning-Kruger overestimation of themselves. It's hard enough having a tech career while not being one of these people, because every day you're encountering new systems you've never seen before and you're tasked to figure them out anyway, and you have to be the grown-up in the room and make it work even if your brain just completely refuses to cooperate that day. The longer I work as a software engineer, the more I realize how little I know... impostor syndrome's rampant. I'd be a force to be reckoned with if I never realized that and just kept blustering forward without pausing for reflection.
And failing up is a thing too, I've seen plenty of colleagues bounce from one job to another, each position more lucrative and/or prestigious than the last. My own career has had a similar trajectory too, if I'm being honest, though it's not like this was a strategic decision on my part. I feel like it should be considered a red flag in the hiring process if a candidate's tenure at previous companies is never longer than a year or two. But that rarely seems to come up. Que será, sera.
Anyway, that whole Julius story was the setup for a hamfisted critique of AI and I don't get what the author is trying to say by it. It's true that we're in the middle of a hype frenzy around AI right now, and a lot of the voices driving it don't have the foggiest clue how the tech works or what its limitations or valid uses cases are. But it's a leap to go from that, to writing off all AI tools as nonsense. Don't throw the baby out with the bathwater.
I think the message of the story was clear. It's that AI gets too much credit and it doesn't really have intelligence. The 1979 Peter Sellers movie Being There is about a simpleminded man who is...
I think the message of the story was clear. It's that AI gets too much credit and it doesn't really have intelligence.
The 1979 Peter Sellers movie Being There is about a simpleminded man who is thought to be very intelligent and mysterious by people who should know better. This is where AI is for me.
Oh— it's the idea is that the AI itself is a Julius? I missed that. I thought it was saying that because of other know-nothing Juliuses (in management or positions of influence), the author was...
Oh— it's the idea is that the AI itself is a Julius? I missed that. I thought it was saying that because of other know-nothing Juliuses (in management or positions of influence), the author was getting saddled with useless AI tools. I must have read it too quickly if I missed that.
Yeah. It is confusing but Julius was a real person that had splinter skills, where the highly visible skills masked his other incompetence, and AI reminds the author of this kind of person. Well...
Yeah. It is confusing but Julius was a real person that had splinter skills, where the highly visible skills masked his other incompetence, and AI reminds the author of this kind of person.
Well maybe the whole story was made up but Julius-type people are seen a lot in life.
I started out assuming this was a true story, but disbelieved it after a while. It seems metaphorical? There is a big difference: if an AI tool for programming isn’t working for you, you can...
I started out assuming this was a true story, but disbelieved it after a while. It seems metaphorical?
There is a big difference: if an AI tool for programming isn’t working for you, you can simply stop using it and write code without it.
Or at least that’s true today. I worry that it won’t always be true. Systems will be designed where you can’t avoid using it.
Yeah, the part that made it sound unbelievable was about the “AI” tools – what kind of policy is that? And even if it were in place somewhere, I’m sure you could find ways around it. (I get that...
Yeah, the part that made it sound unbelievable was
I am not allowed to disable them.
about the “AI” tools – what kind of policy is that? And even if it were in place somewhere, I’m sure you could find ways around it.
(I get that the story serves as an allegory/intro to discovering LLMs’ issues for people who might not realize those are a thing; just commenting on whether I hold this for a true story or not.)
As a hypothetical example, let's say for some reason you need to use a web-based development environment, and then you might be forced to use whatever tools they give you. Maybe you have to talk...
As a hypothetical example, let's say for some reason you need to use a web-based development environment, and then you might be forced to use whatever tools they give you. Maybe you have to talk to a chatbot to make changes?
Code Interpreter on ChatGPT works this way. There's no way to edit the code directly. To make changes, you have to convince it to do what you want. It can be fun if you treat it as a game, but I wouldn’t want to have to use it for real.
Programmers have it good because the data is usually stored in plain text files and we can usually pick the text editor we want, but consumer-facing apps often don’t work that way. Sometimes you can’t even cut and paste.
It’s particularly true for dedicated hardware. Being able to update your thermostat’s schedule with a text editor: kind of rare? You use the UI they give you. Maybe some devices will only support voice? “I’m sorry Dave…”
Thanks for the new nightmare. This is an interesting and true perspective. Version control doesn’t really work for Office documents, either, for a similar example… Luckily, at least in the...
As a hypothetical example, let's say for some reason you need to use a web-based development environment, and then you might be forced to use whatever tools they give you. Maybe you have to talk to a chatbot to make changes?
Thanks for the new nightmare.
Programmers have it good because the data is usually stored in plain text files and we can usually pick the text editor we want, but consumer-facing apps often don’t work that way.
This is an interesting and true perspective. Version control doesn’t really work for Office documents, either, for a similar example…
Sometimes you can’t even cut and paste.
Luckily, at least in the browser, there’s still extensions.
I find it remarkably ironic that this essay about how AI is like an awful coworker uses an AI-generated image as its hero. EDIT ... ok so maybe the stock image site isn't ai generated? I maintain...
I find it remarkably ironic that this essay about how AI is like an awful coworker uses an AI-generated image as its hero.
EDIT ... ok so maybe the stock image site isn't ai generated? I maintain that it's confusing and that they should include a statement making the fact explicit. but in this case I might have been being over cynical, sorry everyone
The title of the image is Clickworker 3d-printed and, although there are a couple of weird spots, it does seem to me like it's an actual photo of something someone 3D printed.
The author also uploaded another image of the same 3D printed figures:
I think the Julius effect is a real phenomenon. There are certain personality types that excel at flattery and bluffing their way through situations. Some of these people are willfully scamming, but I think a larger portion of them are earnestly unaware of their incompetence due to the unhelpful signals they receive from others, and some sort of relentless Dunning-Kruger overestimation of themselves. It's hard enough having a tech career while not being one of these people, because every day you're encountering new systems you've never seen before and you're tasked to figure them out anyway, and you have to be the grown-up in the room and make it work even if your brain just completely refuses to cooperate that day. The longer I work as a software engineer, the more I realize how little I know... impostor syndrome's rampant. I'd be a force to be reckoned with if I never realized that and just kept blustering forward without pausing for reflection.
And failing up is a thing too, I've seen plenty of colleagues bounce from one job to another, each position more lucrative and/or prestigious than the last. My own career has had a similar trajectory too, if I'm being honest, though it's not like this was a strategic decision on my part. I feel like it should be considered a red flag in the hiring process if a candidate's tenure at previous companies is never longer than a year or two. But that rarely seems to come up. Que será, sera.
Anyway, that whole Julius story was the setup for a hamfisted critique of AI and I don't get what the author is trying to say by it. It's true that we're in the middle of a hype frenzy around AI right now, and a lot of the voices driving it don't have the foggiest clue how the tech works or what its limitations or valid uses cases are. But it's a leap to go from that, to writing off all AI tools as nonsense. Don't throw the baby out with the bathwater.
I think the message of the story was clear. It's that AI gets too much credit and it doesn't really have intelligence.
The 1979 Peter Sellers movie Being There is about a simpleminded man who is thought to be very intelligent and mysterious by people who should know better. This is where AI is for me.
Oh— it's the idea is that the AI itself is a Julius? I missed that. I thought it was saying that because of other know-nothing Juliuses (in management or positions of influence), the author was getting saddled with useless AI tools. I must have read it too quickly if I missed that.
Yeah. It is confusing but Julius was a real person that had splinter skills, where the highly visible skills masked his other incompetence, and AI reminds the author of this kind of person.
Well maybe the whole story was made up but Julius-type people are seen a lot in life.
I started out assuming this was a true story, but disbelieved it after a while. It seems metaphorical?
There is a big difference: if an AI tool for programming isn’t working for you, you can simply stop using it and write code without it.
Or at least that’s true today. I worry that it won’t always be true. Systems will be designed where you can’t avoid using it.
Yeah, the part that made it sound unbelievable was
about the “AI” tools – what kind of policy is that? And even if it were in place somewhere, I’m sure you could find ways around it.
(I get that the story serves as an allegory/intro to discovering LLMs’ issues for people who might not realize those are a thing; just commenting on whether I hold this for a true story or not.)
As a hypothetical example, let's say for some reason you need to use a web-based development environment, and then you might be forced to use whatever tools they give you. Maybe you have to talk to a chatbot to make changes?
Code Interpreter on ChatGPT works this way. There's no way to edit the code directly. To make changes, you have to convince it to do what you want. It can be fun if you treat it as a game, but I wouldn’t want to have to use it for real.
Programmers have it good because the data is usually stored in plain text files and we can usually pick the text editor we want, but consumer-facing apps often don’t work that way. Sometimes you can’t even cut and paste.
It’s particularly true for dedicated hardware. Being able to update your thermostat’s schedule with a text editor: kind of rare? You use the UI they give you. Maybe some devices will only support voice? “I’m sorry Dave…”
Thanks for the new nightmare.
This is an interesting and true perspective. Version control doesn’t really work for Office documents, either, for a similar example…
Luckily, at least in the browser, there’s still extensions.
I find it remarkably ironic that this essay about how AI is like an awful coworker uses an AI-generated image as its hero.
EDIT ... ok so maybe the stock image site isn't ai generated? I maintain that it's confusing and that they should include a statement making the fact explicit. but in this case I might have been being over cynical, sorry everyone
If you mean the image at the start of the article, I don't think it's AI-generated, is it?
Here's a much higher resolution version:
https://betterimagesofai.org/images/downloads?url=%2FImagesAI%2FMax_Gruber_Clickworker_6000x4000.JPG&size=extraLarge
The title of the image is Clickworker 3d-printed and, although there are a couple of weird spots, it does seem to me like it's an actual photo of something someone 3D printed.
The author also uploaded another image of the same 3D printed figures:
https://betterimagesofai.org/images?artist=MaxGruber&title=ClickworkerAbyss
from what I could tell of the better images of ai website I thought they were all ai generated on there?