[Technology] Artificial Intelligence

[Wozza]

It's funny to all of a sudden we see people becoming worried or scared about AI even though it's been chugging along in the back ground for years now.

Nobody was giving a toss about it 3 months ago but now people are starting to say shut it down

Well, ChatGPT 4.0 is pretty good, but I preferred their earlier stuff.

 

[Couldn't Be Hyypia]

I was going to post this. Jaw dropping

 

[Poojah]

I was going to post this. Jaw dropping

Saw this posted in another thread, and I’ve since watched the full two hours. It’s an interesting discussion, absolutely, but I can’t help but feel that it errs on the sensational.

I’ve been using AI, or more specifically GPT, commercially, for about 2 and a half years. Without being incredibly technical, I have a relatively good handle on how it works, what it can do, and the pathway it is taking is on.

Without doubt, it is very clever technology, and it represents a seismic step akin to if not greater than the creation of the internet, and it will have ramifications, some positive, some negative, for almost everyone.

That said, what we have right now is something that is capable of ingesting huge amounts of data being able to manipulate that data in various ways, some useful, some less so. For example, if you asked it to write you the most recent BHA match report in the style of Shakespeare, it could do so in a matter of seconds.

In effect, it knows what Brighton & Hove Albion is, what sport you are referring to, what their last match was, who Shakespeare was and what his writing style was like. More than that, it is able to combine those pieces of information to respond to the original request, and that is groundbreaking.

But it is not sentient, nor is it on a direct pathway to sentience, or a position where it might enslave humanity. Imagine the shift from calculator to spreadsheet - that represented a huge increase in mathematical power. I see this as the shift from Google to AI, only Google (and search in general) covers a subject matter much broader than numbers and is therefore many times more profound with millions more applications.

Will that mean job losses? Sadly it inevitably will. I once thought, arrogantly, my own job was too skilled, too nuanced to be replaced by artificial intelligence. I am no longer convinced. Will the world change drastically over the course of the next 10 years or so? Yes, even faster than the last 10. But are we on a road to destruction? A path slipperier and more treacherous than climate change? I honestly don’t think so.

This bloke, captivating to listen to though he is, has a book to sell.

 

[Couldn't Be Hyypia]

Saw this posted in another thread, and I’ve since watched the full two hours. It’s an interesting discussion, absolutely, but I can’t help but feel that it errs on the sensational.

I’ve been using AI, or more specifically GPT, commercially, for about 2 and a half years. Without being incredibly technical, I have a relatively good handle on how it works, what it can do, and the pathway it is taking is on.

Without doubt, it is very clever technology, and it represents a seismic step akin to if not greater than the creation of the internet, and it will have ramifications, some positive, some negative, for almost everyone.

That said, what we have right now is something that is capable of ingesting huge amounts of data being able to manipulate that data in various ways, some useful, some less so. For example, if you asked it to write you the most recent BHA match report in the style of Shakespeare, it could do so in a matter of seconds.

In effect, it knows what Brighton & Hove Albion is, what sport you are referring to, what their last match was, who Shakespeare was and what his writing style was like. More than that, it is able to combine those pieces of information to respond to the original request, and that is groundbreaking.

But it is not sentient, nor is it on a direct pathway to sentience, or a position where it might enslave humanity. Imagine the shift from calculator to spreadsheet - that represented a huge increase in mathematical power. I see this as the shift from Google to AI, only Google (and search in general) covers a subject matter much broader than numbers and is therefore many times more profound with millions more applications.

Will that mean job losses? Sadly it inevitably will. I once thought, arrogantly, my own job was too skilled, too nuanced to be replaced by artificial intelligence. I am no longer convinced. Will the world change drastically over the course of the next 10 years or so? Yes, even faster than the last 10. But are we on a road to destruction? A path slipperier and more treacherous than climate change? I honestly don’t think so.

This bloke, captivating to listen to though he is, has a book to sell.

I take your point. But the key take out for me wasn't that the robots will kill us all, but rather the damage someone malicious could do with AI when it gets to a certain point.

His key point was that it is people that are the issue, and not AI deciding humans are a problem. When that penny drops, it all becomes a lot more terrifying.

 

[MJsGhost]

AI can help people with basic software engineering skills develop applications far beyond their ability (already here with low/no code products such as Power Apps not even labelled as AI).
Actually knowing how to code will soon be obsolete in many situations, because it's possible to describe what you want using natural language and AI will create the code. It still takes some knowledge (currently at least) of how it all fits together, but the nuts and bolts of coding won't be necessary for many 'development' roles (though, as with using a calculator, having an idea of how to do it 'manually' will help spot where it has gone wrong rather than blindly accepting the result. See also: SatNav).

View attachment 161654

The coding is done. AI can already do more than most of the best coders in the word. You ask it for something, anything, and it will write you the relevant code.

Opinion | It’s the End of Computer Programming as We Know It. (And I Feel Fine.)

Don’t bother learning to code.

www.nytimes.com

 

[KZNSeagull]

I take your point. But the key take out for me wasn't that the robots will kill us all, but rather the damage someone malicious could do with AI when it gets to a certain point.

His key point was that it is people that are the issue, and not AI deciding humans are a problem. When that penny drops, it all becomes a lot more terrifying.

The key point I took out from it was the mass unemployment that may occur from AI in a short space of time, which would cause all sorts of societal problems. Interestingly, I heard on the radio this morning that there may be a trial of Universal Basic Income in England soon, if funding can be found. Will be interesting to see the result of the Welsh trial.

 

[Couldn't Be Hyypia]

Opinion | It’s the End of Computer Programming as We Know It. (And I Feel Fine.)

Don’t bother learning to code.

www.nytimes.com

Is anyone else finding that a 12 foot ladder is less effective than it used to be. Are walls taller now?

 

[beorhthelm]

Opinion | It’s the End of Computer Programming as We Know It. (And I Feel Fine.)

Don’t bother learning to code.

www.nytimes.com

article by someone who didnt understand software engineering (he says he dropped it went to writing), telling us how software engineering will be changed

coding is a small part of development/engineering. understanding the problem, design of solution, testing, debugging, optimisation, deployment etc all take far more time. so if AI changes a small part of the process, it leaves more time for the rest. AI will mean non-developers able to do more, better in fields outside core software engineering.

 

[MJsGhost]

article by someone who didnt understand software engineering (he says he dropped it went to writing), telling us how software engineering will be changed

coding is a small part of development/engineering. understanding the problem, design of solution, testing, debugging, optimisation, deployment etc all take far more time. so if AI changes a small part of the process, it leaves more time for the rest. AI will mean non-developers able to do more, better in fields outside core software engineering.

Are you trashing the whole article because the journalist isn't an IT expert, or did you miss that a big chunk is simply reporting the views of Matt Welsh (former Professor of Computing at Harvard & currently CEO of an AI start-up). Welsh's full "End of Programming" article for the ACM (association for computing machinery) is here: https://cacm.acm.org/magazines/2023/1/267976-the-end-of-programming/fulltext. I'd say he has fair credentials to be commenting on this (granted he has skin in the game, so a pinch of salt is required as ever)

The rest of your post is no different to what I've already said (except that I am less optimistic that there will just be jobs magically created for all of the people whose roles are reduced or replaced by AI).

You are right - coding is just a part of the software development process. What can't be ignored is AI's potential impact on the rest of it as well. Those redundant developers are going to have to pivot to something else, but so are some BAs, PMs, PMO staff, Testers, DevOps Engineers, Trainers...

understanding the problem - natural language prompts from business users to an AI DevBot (© MJsGhost 2023) won't replace business analysts completely (not for a while), but will speed BAs requirements gathering & analysis up, reducing headcount
design of solution - made for AI. Teach it your Enterprise Architecture standards, design artifacts, coding standards etc and your architects will save a whole heap of time, reducing headcount
testing - test automation has been around for years and will be accelerated by AI.
debugging - made for AI, whether it's hunting them out, or preventing bugs through automated code review / predictive analysis in the first place
optimisation - AI will help with predictive and dynamic resource scaling and allocation, network or database optimisation...
deployment - we're a long long way from boot disks already. Automation/orchestration centrally is done by the push of a button now. We're not really going to need a person to push the button (or even write the orchestration scripts)

The project manager overseeing the above is going to have a load of AI-powered software to reduce the time it takes them (and/or reduces PMO support headcount). Training chatbots and/or video user guides generated by AI will replace some of the work done by IT Trainers...

"But increasing productivity can only be good... they can concentrate on higher-value tasks" I hear you say...

All well and good, but the reality is that once all of the use cases for AI are considered, it's pretty clear that the net result is very likely to be a huge loss of jobs, because it can/will be used pretty much everywhere.

The speed at which that could happen isn't going to leave much time for economies/societies to adapt. I don't personally think it will herald the end of civilization as we know it, but I do think that there's a significant risk of major trouble ahead unless some brakes are applied somehow (sadly I don't have the answer to how and chatGPT doesn't seem willing to help for some reason)

Looking on the bright side, we might have a societal flip, where there's only enough 'work' for people to do 2 days a week each. 5 day weekends as standard and enough UBI for a comfortable life...? BRING IT ON!

 

[GingerBeerMan]

I'm surprised by the number of people sleeping on the fact that major breakthroughs in AI are being reached every month. To the clued in observer it seems humanity as we know it will be drastically changed in as little as five years.

And AI isn't just coming for the blue collar jobs, rather it's the white collar jobs that will go first. The world economy is not yet ready for a huge rise in unemployment for programmers, paralegals, web developers, marketers, accountants, call center works, translators etc.

I'll post some cool use cases and AI-related news over the next few days. Hopefully I can get a few more NSCers aboard the AI hype train.

Edit: I'll use the existing thread, I've asked for this one to be deleted.

 

[Wozza]

Some previous discussion here...

[Technology] - ChatGPT

My grandson says that it's more intelligent than humans is it? What does this mean to mankind? I worry about the future employment of youngsters. Who is the expert that can tell us exactly where this is going? Thanks for your help.

www.northstandchat.com

[Technology] - Artificial Intelligence

This potential disaster seems to have a crept up on us and is a far bigger threat than most of us realise. If some of the experts are to be believed, we have created a monster that is going to kill us all, and it will be sooner than we expect. It’s seriously worrying and I’m not sure what to...

www.northstandchat.com

 

[Bakero]

I think a lot of people are well aware of what's coming. Problem is, no one really knows what to do about it.

NSC does have one quite vocal doubter, who I'm sure will be along soon...

 

[GingerBeerMan]

Some previous discussion here...

[Technology] - ChatGPT

My grandson says that it's more intelligent than humans is it? What does this mean to mankind? I worry about the future employment of youngsters. Who is the expert that can tell us exactly where this is going? Thanks for your help.

www.northstandchat.com

[Technology] - Artificial Intelligence

This potential disaster seems to have a crept up on us and is a far bigger threat than most of us realise. If some of the experts are to be believed, we have created a monster that is going to kill us all, and it will be sooner than we expect. It’s seriously worrying and I’m not sure what to...

www.northstandchat.com

Ah, thank you!

I did know the ChatGPT one existed, but I searched for an Artificial Intelligence thread and nothing showed up (ironically).

I will use the existing thread to post some cool updates.

 

[TomandJerry]

Some of us here on NSC struggle with the most basic of technology to be worrying about AI

 

[GingerBeerMan]

This is pretty cool - draw and prompt in real time

 

[Wozza]

Ah, thank you!

I did know the ChatGPT one existed, but I searched for an Artificial Intelligence thread and nothing showed up (ironically).

I will use the existing thread to post some cool updates.

The secret is to search on Google:

site:northstandchat.com "subject here"

 

[raymondo]

The biggest threat to humanity is humanity, cos we are indeed effing stupid. I am waiting for AI to solve global warming...if it can't then it's a waste of space and resources.

 

[KZNSeagull]

The biggest threat to humanity is humanity, cos we are indeed effing stupid. I am waiting for AI to solve global warming...if it can't then it's a waste of space and resources.

It won't take long to work out that the easiest solution is to kill all the humans.

 

[GingerBeerMan]

The level of complexity here is absurd. Think of the different aspects of this, the neural programming that allows the AI to understand the written prompts, the understanding of graphical representation, and then being able to turn that all into a functioning game by writing it's own script. The most amazing thing is that ChatGPT was only released 12 months ago, imagine what we can have in the next 12 months.

 

[The Clamp]

The capability of AI is nowhere near as advanced as people think, or boffins would have you believe.