AI in the Workplace

[RaleighBlue]

They should work in tandem. You're not so much replacing humans as enhancing their capabilities. Humans have always invented new tools to make life easier. The internet, factories, the wheel. They all came about to make things faster and easier. AI is the same, it's just the new kid on the block.

Mmmm.

I think this ignores the reality of it though. It's not actually been proven at any scale to add real value. No one is making money here, it's just investment in the expectation that this will lower costs somehow. Otherwise the investment doesn't make much sense, because there is no world where they could recoup the investment in increased revenues.

AI isn't the internet. It's not a revolutionary way to share information. It's just a new spin on how to access that information, nicely preloaded, filtered, cleaned, and without insight into it's specific sourcing....so basically worse in every way from a truth-value standpoint. The thing is, it was the possibilities engendered by the sharing of human ideas that drove the past 25 years of growth.

AI doesn't do that. It just repackages what is out there, it adds nothing new, and I don't see how it enables adding anything new in the long game.

I'm just not sure it actually makes anything easier that's truly worth doing. Everything I've created with it that felt good at the time... when I was done I was left with a sour taste in my mouth and the feeling I should have either done the thing by hand, or the feeling the output itself had no real inherent value.

It's a mildly amusing thing with some specific values*

*Again, aside from the highly valuable use-case of specially training LLMs, which are very cool statistical technologies, which is where the groundbreaking AI headlines all come from. I'm just talking about the big public model LLMs like ChatGPT/Claude/Grok.


Editing, because here I am.

What I mean, is that yes it's good at creating a research paper, but the value of creating a research paper is doing the research, and learning the subject deeply enough to write about it. Using AI creates the output, but none of the value.

In my experience, that is pretty much all of what it's good at; stuff where the value of the output is the actual input that should go into it, with the output just being a side effect. Using AI is, in that world, inherently counterproductive while still appearing to be highly productive to those that are only observing the outputs.

This would function like a time bomb in the innards of all companies working this way, where one day no one knew what anything was doing anymore and that happened slowly, like a frog in a pot.

I dunno. I've been working with it pretty deeply for a while now, and I'm just really not convinced.

 

[Bkk Andy]

Mmmm.

I think this ignores the reality of it though. It's not actually been proven at any scale to add real value. No one is making money here, it's just investment in the expectation that this will lower costs somehow. Otherwise the investment doesn't make much sense, because there is no world where they could recoup the investment in increased revenues.

AI isn't the internet. It's not a revolutionary way to share information. It's just a new spin on how to access that information, nicely preloaded, filtered, cleaned, and without insight into it's specific sourcing....so basically worse in every way from a truth-value standpoint. The thing is, it was the possibilities engendered by the sharing of human ideas that drove the past 25 years of growth.

AI doesn't do that. It just repackages what is out there, it adds nothing new, and I don't see how it enables adding anything new in the long game.

I'm just not sure it actually makes anything easier that's truly worth doing. Everything I've created with it that felt good at the time... when I was done I was left with a sour taste in my mouth and the feeling I should have either done the thing by hand, or the feeling the output itself had no real inherent value.

It's a mildly amusing thing with some specific values*

*Again, aside from the highly valuable use-case of specially training LLMs, which are very cool statistical technologies, which is where the groundbreaking AI headlines all come from. I'm just talking about the big public model LLMs like ChatGPT/Claude/Grok.

Well one of my teams has AB tested their tasks with and without AI over a 4 month period and with AI has made ~1m USD more than without. It's not a magic wand, and each one has it's own strengths and weaknesses, but there is undeniable value if used in the best areas.

 

[RaleighBlue]

What

Well one of my teams has AB tested their tasks with and without AI over a 4 month period and with AI has made ~1m USD more than without. It's not a magic wand, and each one has it's own strengths and weaknesses, but there is undeniable value if used in the best areas.

I won't argue that there is undeniable value.

I see value.

I don't see anywhere near the value that justifies the valuations of AI companies. And I see more pitfalls from the hype around the tech than value from the tech itself.

I also am a jaded dev with many years at...enough...companies and I know that they often damage their long term prospects for short term gains. If you're integrating it effectively and with proper oversight, that's really cool. I am not going to deny the possibilities here. 2005-2010 google search was basically a cheat code on life too compared to the competition.

 

[NeilMorepies]

It all depends on what your job is.
Let see Ai unblock a drain or change the oil on
Car.

Any job that can be done by
AI isn’t a real job

 

[Nymzee]

The people refusing to use it are putting themselves in a position where they have the least knowledge. That in itself leaves them open to being replaced.

If you give people a crutch for too long they inevitably forget how to walk without one.

 

[Hinchcliffe's Corner]

AI does my wank for me.

Cheers Al

 

[Hinchcliffe's Corner]

Davek was one of the first AI programs I've interacted with.

 

[Tony Hibbert's Wife]

It all depends on what your job is.
Let see Ai unblock a drain or change the oil on
Car.

Any job that can be done by
AI isn’t a real job

I think the idea that anything that requires physical labour will be insulated from these kinds of technological advances is pretty short sighted. The human element will never be removed but that doesn't mean the expertise involved cannot be replicated by a form of AI in the future. How far are we from being able to set up a phone to walk you through the steps of an oil change in real time, while it watches and corrects you? Or any similar job? If that happens, what differentiates the artisan from the average joe? Having the correct tools to hand?

Consider also the 'Internet of things'. How long before everything you own is connected to and monitored by AI? What about proprietary rights? Your washing machine goes on the blink. AI diagnoses the issue, orders the part and books a fix with the manufacturer (or authorised agent) for a fraction of the price of calling out a local tradie. Is there a future where nothing in your house or car can be fixed without the requisite diagnostic tools, closely controlled by intellectual property rights and terms of purchase for the benefit of the company that made the product?

Workers of the world united, my friend. To a capitalist there are no 'real' jobs, just cost and profit.

 

[Bkk Andy]

If you give people a crutch for too long they inevitably forget how to walk without one.

A bit like hunting for food or making fire for heat...

 

[Nymzee]

A bit like hunting for food or making fire for heat...

Well yes, exactly like that. If the world ever descended into a post-apocolyptic state most people would just die out because they don't know how to hunt or make a fire.

 

[Bkk Andy]

Well yes, exactly like that. If the world ever descended into a post-apocolyptic state most people would just die out because they don't know how to hunt or make a fire.

You're absolutely right. Survival of the fittest. Those who adapt/learn fastest stand the best chance of survival. Whether we like it or not, new tech is exactly the same. We embrace it or get left behind.

 

[magicjuan]

Because I don't want to hire and pay 2 people when 1 will do? That's just commonsense. Maybe you can take half pay to encourage your mingebags to hire more people.

The dig was at management speak, but it seems you're a bit spikey about that. Perhaps a meeting, or a working group could help? Or just ask Chatcrap. I'm sure you'll deal with your non social practical environment collegue unawareness dissolution procedural efficiences

 

[Nymzee]

You're absolutely right. Survival of the fittest. Those who adapt/learn fastest stand the best chance of survival. Whether we like it or not, new tech is exactly the same. We embrace it or get left behind.

And if that tech ever fails, or becomes privatised, what you're left with is a bunch of people who have no idea how to do anything in your workplace. Sounds like a fun scenario.

 

[Bryan+]

My company has just raised $150m off the back of it.

 

[GoldenToffee]

There will be trouble when companies start charging for it and those who have heavily relied upon it cannot afford to keep using certain tools.

Still, it's remarkable technology and the limits are seemingly limitless.