I'm having AI anxiety. What is your thoughts on upcoming AI?

[AnNvr]

As already highlighted, I agree that the current AI is just a tool, perhaps in a primitive stage of development. Also, I thing is good to boost generic research and provide massive distraction, enhancing ADHD among the new generations.

No doubt it will get better, and it will wipe a few "rinse and repeat" jobs (those who point coding as one of those they either never code in their life or have no clue what developing is about), but it won't happen any time soon.

This FUD of programming being taken over by the current AI is speculative. I tested both ChatGPT4 and Bard to elaborate code snippets providing parameters of the projects i was working for, just to skip some old fashion Stack overflow research. The code provided by AI was beautifully wrong. It lacks of in-depth analysis of the context to create and debug, and most likely won't fill this gap at any time soon.

The jobs really threatened on the front line I think will be cashier, fast food workers, hotel front desk operators, most of the retail, tourism and hospitality operators. And to be fully honest, I hope for these people it will happen soon. Are detrimental jobs that brings nothing but stress, unhappiness, low income and no prospectives. Better set them free.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[Vntonio]

The jobs really threatened on the front line I think will be cashier, fast food workers, hotel front desk operators, most of the retail, tourism and hospitality operators. And to be fully honest, I hope for these people it will happen soon. Are detrimental jobs that brings nothing but stress, unhappiness, low income and no prospectives. Better set them free.

I'm glad I'm not the only one who thinks the same...

Also, I would say some of the jobs (white collar) where you don't need strong analytical thinking or creativity are going to disappear.
This being said, programming is safe, but for sure it will be changed again (it happened lots of times); I really think it will be impossible as a light programmer, coming from no STEM background, to get a job. But sure, a mid/senior level will still be needed.

 

[tomisin2k]

As already highlighted, I agree that the current AI is just a tool, perhaps in a primitive stage of development. Also, I thing is good to boost generic research and provide massive distraction, enhancing ADHD among the new generations.

No doubt it will get better, and it will wipe a few "rinse and repeat" jobs (those who point coding as one of those they either never code in their life or have no clue what developing is about), but it won't happen any time soon.

This FUD of programming being taken over by the current AI is speculative. I tested both ChatGPT4 and Bard to elaborate code snippets providing parameters of the projects i was working for, just to skip some old fashion Stack overflow research. The code provided by AI was beautifully wrong. It lacks of in-depth analysis of the context to create and debug, and most likely won't fill this gap at any time soon.

The jobs really threatened on the front line I think will be cashier, fast food workers, hotel front desk operators, most of the retail, tourism and hospitality operators. And to be fully honest, I hope for these people it will happen soon. Are detrimental jobs that brings nothing but stress, unhappiness, low income and no prospectives. Better set them free.

Do not be troubled.

History repeats itself. Everyone thought the world was coming to an end when the personal computer became mainstream.

AI is a tool and will always need the creativity of humans to work well.

Someone needs to handle these tool.

 

[Kevin88660]

As already highlighted, I agree that the current AI is just a tool, perhaps in a primitive stage of development. Also, I thing is good to boost generic research and provide massive distraction, enhancing ADHD among the new generations.

No doubt it will get better, and it will wipe a few "rinse and repeat" jobs (those who point coding as one of those they either never code in their life or have no clue what developing is about), but it won't happen any time soon.

This FUD of programming being taken over by the current AI is speculative. I tested both ChatGPT4 and Bard to elaborate code snippets providing parameters of the projects i was working for, just to skip some old fashion Stack overflow research. The code provided by AI was beautifully wrong. It lacks of in-depth analysis of the context to create and debug, and most likely won't fill this gap at any time soon.

The jobs really threatened on the front line I think will be cashier, fast food workers, hotel front desk operators, most of the retail, tourism and hospitality operators. And to be fully honest, I hope for these people it will happen soon. Are detrimental jobs that brings nothing but stress, unhappiness, low income and no prospectives. Better set them free.

A lot of jobs will be removed, the higher management will use ai as an excuse but it has little to do with ai.

Very simple data automation tool can replace cashier and retail. You can always scan the bar code and pay for your grocery in a super market with simple diy machine. We aren’t pushing automation 100 percent and the decision is political. At least this is the case in my country. A lot of the big brand names supermarkets are partially owned by the government through sovereign wealth fund. You just cannot make all minimum wage workers go jobless. That is a recipe for social backslash. Bullshit jobs keep people paid. That is a political decision.

Elon musk tried firing 80 percent of twitter workforce and it works fine.

Its common knowledge that a lot of big organisation are not lean, even the big techs. It is always 20 percent of the people doing 80 percent of the work. The layoffs in big tech we see today are just firing people who aren’t productive anyway. AI is just a good excuse to retrench without sending massive termination letters saying “You have to go because you suck”. That political pushback will be more.

Someone said big Techs keeps the workers not because they need them, but because they can afford to.

AI will replace lazy workers and salary scammers.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[akshayazariah]

This FUD of programming being taken over by the current AI is speculative. I tested both ChatGPT4 and Bard to elaborate code snippets providing parameters of the projects i was working for, just to skip some old fashion Stack overflow research. The code provided by AI was beautifully wrong. It lacks of in-depth analysis of the context to create and debug, and most likely won't fill this gap at any time soon.

This is operating on the assumption that such tools will either progress at an inconveniently slow rate, or effectively remain the same in terms of abilities. That's just not happening.

I do think a lot of the repetitive, mundane tasks related to programming will be neatly tackled by NLP models in the (extremely) near future. With that said, though, the creative aspect of it will always remain. Really, AI tools will just weed out the innovative from the masses.

 

[WilliamSherman]

Honestly, the worry that AI might take over programming jobs seems a bit much for now. AI's good at helping out with small coding tasks or fixing minor bugs, but it just doesn't get the big picture like humans do. Programming isn't just about typing code; it's about really understanding what's needed, building big projects, and making important choices on how everything works together.

 

[MJ DeMarco]

Honestly, the worry that AI might take over programming jobs seems a bit much for now. AI's good at helping out with small coding tasks or fixing minor bugs, but it just doesn't get the big picture like humans do. Programming isn't just about typing code; it's about really understanding what's needed, building big projects, and making important choices on how everything works together.

Oh, did you write this using AI? Because multiple users are reporting your posts as AI, and using AI for interaction here is grounds for removal.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[SamMitchell1]

Hello, I am interested in hearing your thoughts on upcoming AI tools. Initially, they were incredible, but now they are causing anxiety. Do you share this feeling, or am I alone in this? The ongoing bettle between Altman and Musk regarding AI, as well as the six-month pause on all AI Labs, have added to my concerns. Furthermore, it appears that some jobs, such as content writing, coding, and graphic design, are already being taken over by AI to some extent. All of this is quite alarming.

It's not even the begining. From what I see in just few years this shit will be developed even more. ChatGPT is working on 20% of what it can be in nearest future. Time to diversify

 

[TempusFugit]

I'm not sure about AI Anxiety. I for one am thankful to our future robot overlords for helping us entrepreneurs.
All joking aside, AI actually is extremely helpful for web dev if you are just starting out (like me). I would say that the barrier of entry is slightly lower because of it, but you still need to learn basic HTML/CSS/JS skills in order to ask something like chatGPT the right questions. You can't just be like "code this for me". You need to go back and forth between your code and asking the AI to perform the correct action for you.
Anyhow, that's just my 2¢.

 

[srodrigo]

I just installed GitHub Copilot again today, to give it a fair try (paying for 1 month should make me take it more serious than during the free trial last year). So far, so good; programmers' job security I mean, because if this is the state-of-art AI for coding, jobs are not going anywhere.

The good part is I only paid $10, so if it continues being close to useless, at least I won't feel like I've been ripped off.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[Bounce Back]

I just installed GitHub Copilot again today, to give it a fair try (paying for 1 month should make me take it more serious than during the free trial last year). So far, so good; programmers' job security I mean, because if this is the state-of-art AI for coding, jobs are not going anywhere.

The good part is I only paid $10, so if it continues being close to useless, at least I won't feel like I've been ripped off.

I'm not allowed to use it at work due to security reasons (it has to send your code currently to gpt servers is my understanding).

But I got it for my own development and I won't go back. I don't think it replaces devs but makes me easily twice as productive when I am in the zone but I have to wade through a lot of junk it suggests and it fails on some contextual things only humans can really weigh in (bigger picture goals you are trying to do with the code, design decisions, code style stuff, etc.).

I actually think its not really going to help junior/mid levels that much - it will hurt them long run if they use it wrong (spoon feeding vs learning why). Seniors + Principal engineers definitely can crank out what they already know how to do quicker though and also gives you an unembarassing way to bounce ideas off yourself where as before you'd have to bug your coworker.

 

[StrikingViper69]

I just installed GitHub Copilot again today, to give it a fair try (paying for 1 month should make me take it more serious than during the free trial last year). So far, so good; programmers' job security I mean, because if this is the state-of-art AI for coding, jobs are not going anywhere.

The good part is I only paid $10, so if it continues being close to useless, at least I won't feel like I've been ripped off.

From what I’ve heard, it’s ok writing bits of code, for example the odd function, but you still need a programmer to check it and put it all together.

Unless you’re being paid to write “hello world”, programmers are going to be around for a while

 

[srodrigo]

I'm not allowed to use it at work due to security reasons (it has to send your code currently to gpt servers is my understanding).

But I got it for my own development and I won't go back. I don't think it replaces devs but makes me easily twice as productive when I am in the zone but I have to wade through a lot of junk it suggests and it fails on some contextual things only humans can really weigh in (bigger picture goals you are trying to do with the code, design decisions, code style stuff, etc.).

I actually think its not really going to help junior/mid levels that much - it will hurt them long run if they use it wrong (spoon feeding vs learning why). Seniors + Principal engineers definitely can crank out what they already know how to do quicker though and also gives you an unembarassing way to bounce ideas off yourself where as before you'd have to bug your coworker.

I guess everyone's brain works in a different way. It's cool that you found Copilot useful.

I find it useful to type less, but for any mildly complex thing, it just gets in my way. For example, say you want to build a feature that has frontend and backend. You need to build the frontend in baby steps until you make some API call. Then same with the backend, starting with the API and ending in the DB. But Copilot sucks at this, you need to guide it... at which point it just breaks my flow and I'd rather just do it as usual because it's faster for me to do it than to explain and nitpick lines of code. The amount of explanation needed for it to give you an example that doesn't fit at all is just too much work and distraction. And it's okay at creating small pieces of code, but at that point it's become more of a better auto-completion tool than something mind-blowing. I'll keep trying to use it though, maybe I'll report back that it's great in a few weeks.

I think Juniors are going to have a hard time in general. The market is horrible even for seniors, let alone people with little experience. And I don't think it's going to get much better in the long-term.

I do think that there'll be an "AI fixer" role soon after people who have no clue about what they are doing start cranking out crap made by AI that barely holds together.

From what I’ve heard, it’s ok writing bits of code, for example the odd function, but you still need a programmer to check it and put it all together.

Unless you’re being paid to write “hello world”, programmers are going to be around for a while

That's my same thought. I see some people on Twitter over-exaggerating about Copilot, and I can only thing most of them are writing trivial apps. Even a CRUD feature is a struggle to build with this tool.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[Bounce Back]

I guess everyone's brain works in a different way. It's cool that you found Copilot useful.

I find it useful to type less, but for any mildly complex thing, it just gets in my way. For example, say you want to build a feature that has frontend and backend. You need to build the frontend in baby steps until you make some API call. Then same with the backend, starting with the API and ending in the DB. But Copilot sucks at this, you need to guide it... at which point it just breaks my flow and I'd rather just do it as usual because it's faster for me to do it than to explain and nitpick lines of code. The amount of explanation needed for it to give you an example that doesn't fit at all is just too much work and distraction. And it's okay at creating small pieces of code, but at that point it's become more of a better auto-completion tool than something mind-blowing. I'll keep trying to use it though, maybe I'll report back that it's great in a few weeks.

I think Juniors are going to have a hard time in general. The market is horrible even for seniors, let alone people with little experience. And I don't think it's going to get much better in the long-term.

I do think that there'll be an "AI fixer" role soon after people who have no clue about what they are doing start cranking out crap made by AI that barely holds together.


That's my same thought. I see some people on Twitter over-exaggerating about Copilot, and I can only thing most of them are writing trivial apps. Even a CRUD feature is a struggle to build with this tool.

Agreed simple CRUD type stuff is a waste of time with it because your currently going to have to feed it more prompt than writing by hand.

But start adding custom utils and things like that your flows and it can figure out the pattern and copy your other code to keep that pattern as it goes (hard to explain in words vs some sort of example).

My main use of it so far has been things like "yo take a look at my nodemon config and my npm scripts - give me 5 things I can check that can cause X to happen" - saved me clicking around an hour on stackoverflow when it ends up being one of the 5 for instance. Another example is - "rewrite this so there isn't a race condition" when you know it has one and don't even want to waste time figuring out what it is - you can see by what it suggests very quickly if it found it and suggested a good fix. Just some ideas.

I haven't used it on the front end but I imagine its good there just for an initial component start like "using styled components and react make me a functional component that takes x,y,z, props in and renders A when B happens, etc." then you take what it spits out and modify it to your liking.

 

[srodrigo]

Agreed simple CRUD type stuff is a waste of time without because your currently going to have to feed it more prompt than writing by hand.

But starting adding custom utils and things like that your flows and it can figure out the pattern and copy your other code to keep that pattern as it goes (hard to explain in words vs some sort of example).

My main use of it so far has been things like "yo take a look at my nodemon config and my npm scripts - give me 5 things I can check that can cause X to happen" - saved me clicking around an hour on stackoverflow when it ends up being one of the 5 for instance. Another example is - "rewrite this so there isn't a race condition" when you know it has one and don't even want to waste time figuring out what it is - you can see by what it suggests very quickly if it found it and suggested a good fix. Just some ideas.

I haven't used it on the front end but I imagine its good there just for an initial component start like "using styled components and react make me a functional component that takes x,y,z, props in and renders A when B happens, etc." then you take what it spits out and modify it to your liking.

Interesting use case, I'll try it when there's something that needs fixing (although I ended up fixing what it created, lol).

I tried it on the frontend this morning, and I wasn't too happy with the results. Maybe it was me not guiding it, but it was for something quite simple that I know how to implement anyway. So I don't know. But I'll keep testing it.

What I found it more useful for was for outlining a feature. It didn't get it right by any means, but it gave me some ideas/starting point. I also tried asking how to improve a piece of code and it made decent suggestions. It's giving me more or less satisfactory results for the high level ideas and for the auto-completion bits, but not for things in between.