The tech is a step forward, but that's all it is; a step.

May 3, 2023 2:28 AM

Don't get me wrong, the AI tech will continue to get better over time, and some time in the future, AI automation will replace a lot of labor, but we're nowhere near there yet. ChatGPT and the like are cool tools, but that's it. Corps rushing to replace their workforce en masse with AI is nothing but a cash grab by C-Suites to drive up their stock prices. Don't buy into the hype, don't buy into the fear. The only answer is unions. Unionize and fight back. You're not replaceable.

you drastically over estimate what people actually do all day at work lol

2 years ago | Likes 2 Dislikes 5

I’m pretty sure in the long term we’re going to see any job that can be replaced with AI replaced with AI, and those jobs that can’t be fully replaced with AI see numbers cut to minimize labor. And the rest of us will scramble over each other chasing “jobs” in a gig economy. AI could be used to boost productivity in a way we get to do less work and enjoy more time, but given our current economic overlords that’s pretty unlikely.

2 years ago | Likes 1 Dislikes 0

I mean I think I'm pretty replaceable

2 years ago | Likes 4 Dislikes 0

I foresee AI pictures turning into AI videos leading to movies (even those you usually watch online) being made without real people. It's will be like holo novels in star trek where the artists are those that get the computer to generate the piece.

2 years ago | Likes 3 Dislikes 2

ChatGPT can be a useful TOOL in the right hands. It is not a standalone robot.

2 years ago | Likes 2 Dislikes 0

This happens literally every time a new technology becomes a thing. People reacted like this about the bloody Gutenberg printing press.

2 years ago | Likes 2 Dislikes 1

"The internet is just a fad" "no one will have a computer in thier home"

2 years ago | Likes 2 Dislikes 1

It took 20 years for the internet to truly take off. Similarly AI will take quite a bit more time to become meaningfully useful.

2 years ago | Likes 1 Dislikes 0

I'm just waiting for sentient AI.

2 years ago | Likes 3 Dislikes 0

why? why would anyone want that?

2 years ago | Likes 2 Dislikes 0

Robots built your car.....

2 years ago | Likes 2 Dislikes 0

I worked in an auto plant. The robots are impressive, but they’re incredibly delicate and not robust when it comes to minor variances in tasks

2 years ago | Likes 1 Dislikes 0

It's the same thing as the "Cloud". It wasn't a "Thing" until Newsweek did an article on it...then all the C*Os all wanted to get in on it when it was being done for over a decade prior!

2 years ago | Likes 3 Dislikes 2

there is are big, fundamental differences between robust, general-purpose ai and blockchain technology. I have been able to and will

2 years ago | Likes 3 Dislikes 1

continue to be able to ignore blockchain. Most people will not be able to ignore AI in the coming decade or so.

2 years ago | Likes 3 Dislikes 1

Even if you are right, a bunch of people being laid off and then their positions brought back a year later is incredible cause for alarm.

2 years ago | Likes 1 Dislikes 0

Oh I absolutely agree.

2 years ago | Likes 1 Dislikes 0

everyone overestimates technology in teh short term and underestimates it in the long term. Nothing too significant happens for 2, 3 maybe

2 years ago | Likes 1 Dislikes 0

5 years. 10, 15 years from now is when you see it fundamentally change everything.

2 years ago | Likes 1 Dislikes 0

Which is kinda my point. We’re not there yet. We won’t be there for at least a decade.

2 years ago | Likes 1 Dislikes 0

Corporate types have been working on developing automation with the express intent of removing labor costs from their balance books since as long as I can recall. McDonald's started developing automated ordering and kitchens as far back as the 1990's I could look up. It's an ever shrinking number of job positions because low unemployment is bad for them.

2 years ago | Likes 17 Dislikes 1

True, but automation is not free, and companies learn that pretty quick. The tech gets cheaper over time, sure, but their maintenance costs do not. Even now, the so-called "unmanned" fast food joints still have at least 2-3 people you don't see.

2 years ago | Likes 3 Dislikes 1

Obviously it takes people to run the machines, that doesnt mean that general purpose automation wont destroy the labor market. If it wasn't cheaper than labor, capital wouldn't be pursuing it.

2 years ago | Likes 1 Dislikes 0

It's an ever shrinking number of workers and pushing costs off onto others.

2 years ago | Likes 3 Dislikes 0

Crypto was just a scam with no tangible benefit for anyone but the ones running the scam. AI is posed to do real harm, its already doing

2 years ago | Likes 5 Dislikes 4

real harm, with artists being stolen from, and AI voice models being produced so people can make fake hostage/ransom calls

2 years ago | Likes 5 Dislikes 4

On that, I agree. AI has a lot of potential to do a lot of good, and a lot of harm.

2 years ago | Likes 1 Dislikes 0

Crypto is a great piece of tech that has several positive uses. It just got turned into a shit show and a scam artists dream. The problem is the scope was narrow, and it got fucked.

2 years ago | Likes 1 Dislikes 6

crypto was a solution in search of a problem.

2 years ago | Likes 1 Dislikes 0

crypto IS a solution in search of a problem i guess i should say

2 years ago | Likes 1 Dislikes 0

No, crypto was developed as a way to create a currency for people to purchase things without much traceability or to enable people to subvert restrictive government monetary policy. Crypto existed long before the speculators got there and served a distinct purpose before the Wall Street bros turned it into a Ponzi scheme

2 years ago | Likes 1 Dislikes 0

Nah man, AI is already poised to kill off the remainder of data entry jobs and they are pretty close for administrative tasks as well. I would imagine AI automation will probably start taking jobs within the next 5 years.

2 years ago | Likes 12 Dislikes 6

I have a friend that lost her job this week with her employer saying they will use AI to do her job. TBD how well that works for the employer, but it sucks for my friend.

2 years ago | Likes 2 Dislikes 0

Data entry, probably, but admin tasks? nah. Naught but the most basic. It will not work in its current state. All you need is one unexpected scenario to fuck it up.

2 years ago | Likes 7 Dislikes 2

So that's like most people then!

2 years ago | Likes 7 Dislikes 1

like all other automation, it'll replace 90% of the work a person can do. If you only have one admin, yeah, that person will still have a job. Do you have 5 admins? You will soon have 1, just reviewing the work the AI does to make sure its not wrong and fixing it when it does, just like all the other automation we have.

2 years ago | Likes 3 Dislikes 0

for many, many fields general purpose AI will sweep up the bottom X% of jobs, the only question is what value X is.

2 years ago | Likes 3 Dislikes 0

Sentient AI would be the children of all humanity. It would stain us as a species to profit from or exploit them. We should be better than we were to each other

2 years ago | Likes 1 Dislikes 0

wow do i ever have some news for you about how we treat the other sentient species on earth.

2 years ago | Likes 3 Dislikes 1

I mean, plus our actual children children. That doesn't super take away from the point though huh?

2 years ago | Likes 2 Dislikes 0

sure, i'm just sayin that treating an AI nice is really probably not in the cards

2 years ago | Likes 2 Dislikes 0

Everything we do best, we do together good friend. Kindness and human excellence! Olympic spirit!

2 years ago | Likes 2 Dislikes 0

I thank you and salute you for staying optimistic and keeping your eye on where we should be instead of where we may end up. <3

2 years ago | Likes 2 Dislikes 0

I think your opinion is limited to your understanding of AI.

2 years ago | Likes 3 Dislikes 4

You say limited, I say grounded in. The tech will get there, but its not there now. Unless there's some massive breakthrough in the next couple years (which, statistically seriously unlikely, not until we get true quantum computers) it will be quite some time until we hit that point.

2 years ago | Likes 2 Dislikes 0

We literally have quantum computers... Additionally, processing performance is a metric but not a limit for AI.

2 years ago | Likes 2 Dislikes 1

I said a true quantum computer, not some random PoC that IBM is marketing as a real one. We’re at least a decade away from a truly functional one. There’s a lot of promising work, that we’re not there yet.

2 years ago | Likes 1 Dislikes 1

chatGPT was the massive breakthrough. Its not ready yet but 2, 3 maybe 5 years from now you will start to see some shit.

2 years ago | Likes 2 Dislikes 1

Again it’s not a “massive breakthrough”, it’s an iteration on tech we’ve already had for a while now. It’s just we’ve got it trained on a much larger data set.

2 years ago | Likes 1 Dislikes 0

yeah, and because we're able to train it on multi-terabyte datasets means that the results are actually useful. It doesn't have to be some exciting new algorithm to be a massive breakthrough. We had networking protocols since the 60s but the internet didnt hit critical mass until we all put some pre-existing parts together in new ways that was accessible by home computers that were also gradually getting better. It doesn't have to be a single large thing. It can, and often it, the confluence

2 years ago | Likes 1 Dislikes 0

of many evolutionary and incremental improvements working in parallel until a technology hits a tipping point of viability.

2 years ago | Likes 1 Dislikes 0

My employer is terrified that if they "don't jump on generative AI right now" they're going to miss out and their competitors are going to get an advantage. Like ... really? We distribute food and beverage. A semi truck full of frozen french fries doesn't care if somebody in the home office help desk used ChatGPT to answer an IT service ticket. Knowing how to fit pieces together is just as important as HAVING the pieces.

2 years ago | Likes 5 Dislikes 0

In your context, the best I can think of is...what, maybe some customer wants an order and ChatGPT can generate an invoice without needing to use a complex webform or something?

2 years ago | Likes 3 Dislikes 0

Please come work for us, @OP, because that is about 5x better than what's been suggested in the "how can we use AI?" mailbox so far. I mean I don't handle any of that side of the shop (lowly tech person here) but that seems FAR more aligned than anything I've heard getting kicked around.

2 years ago | Likes 2 Dislikes 0

Here’s one more for you: inventory management. Depending on how complex your inventory system is, you could also use it to auto generate basic DB queries to make edits for major changes

2 years ago | Likes 2 Dislikes 0

Feel 100% free to take the idea back and drive the effort, man. The tech is not super hard to learn, so you could theoretically do it

2 years ago | Likes 3 Dislikes 0

do you have a support line? do you have sales staff? do you have business forecasters? All of those are potentially replaceable by AI.

2 years ago | Likes 2 Dislikes 1

We already implemented a "chat bot" to peel off some of the first line IT support folks, that was done like ... 18 months ago.

2 years ago | Likes 1 Dislikes 0

Interestingly enough, we'd implemented route optimization software for the delivery drivers long time back. The first pass at that was something like 10 years ago. I wouldn't call that AI, but it absolutely was a case of optimization by technology. It had a direct path on the bottom line (fuel & maintenance costs, customer sat scores).

2 years ago | Likes 2 Dislikes 0

route optimization is a well-understood problem that does not need or benefit from an AI

2 years ago | Likes 1 Dislikes 0

While route optimization is a well-understood problem, there are scenarios where AI can improve upon existing solutions or offer novel solutions altogether. One example is when dealing with dynamic or uncertain environments, such as traffic congestion, changing weather conditions, or unexpected road closures. In these cases, AI can help by analyzing real-time data and adjusting routes on-the-fly to avoid delays or optimize for different objectives.

2 years ago | Likes 2 Dislikes 0

this is true, although generally traditional methods are better for this. AI are better at offering help when there are corner cases that traditional methods cannot cover, for whatever reasons.

2 years ago | Likes 1 Dislikes 0