Artificial Intelligence

[Sly Reflex]

 

I think this is the video Maf is on about. Just so you can see how crap it is.

 

And if they use AI as a "staff member" if it was me I would be taxing these "workers" at such a high rate that it's just not reasonable to operate them in all but the most extreme circumstances, and putting the tax earned from them in a UBI pot. The whole point of tech is it's meant to give us more free time, not make us more stressed and make billionaires richer.

 

I swear the way the world is now, I am becoming more and more radicalised towards how things are and not how they should be. We've built the worst world and the worst thing is we let them build it instead of telling them to get fucked.

[shinymcshine]

Pretty sure some of those 'gymnasts' starred in the Silent Hill games.....

[Jimboxy]

3 hours ago, one-armed dwarf said:

Feels sort of dirty

Sooo dirty. I felt like I had to announce it up front when doing a code review with our senior dev, just to be open about it. He was pretty impressed by it as it turned out and wasn't bothered about me using it in the slightest. Where it really shines is for commenting on code. When I've finished a script I paste it back in, ask it to analyse it and insert comments, then write me an overview to put at the top of the script, such a time saver

[Maryokutai]

Legit question that might maybe leave the topic at hand a bit, but aren't you afraid of getting complacent that way and regressing your skillset? I do a lot of translating work and writing PR stuff and technically I could do both with machine learning translating tools and AI respectively, only touching upon the rough edges later. But I'm genuinely afraid I'm just gonna turn myself into a vegetable long-term if I do that, because it's already a pretty mundane job and if I don't even have that anymore, I dunno. Feels a bit dangerous, for lack of a better term.

Though I don't know much about programming work so I don't know whether these examples are comparable or not.

[Jimboxy]

54 minutes ago, Maryokutai said:

Legit question that might maybe leave the topic at hand a bit, but aren't you afraid of getting complacent that way and regressing your skillset?

 

Maybe? It has crossed my mind. At the moment for me I'm treating it as a learning tool to improve my own skills, which it has. That's helped by the fact that in my new role I'm in the thick of using scripting and automation tools instead of just using those things off my own back, but I think it has accelerated my learning of PowerShell, specifically.

 

Long term, as it improves and people try to find the path of least resistance, yeah I think that is a legitimate concern.

 

In terms of using it for development work, I saw someone the other day compare it to a junior developer, it'll attempt the job, but it needs a lot of hand holding and checking

[shinymcshine]

It'd concern me, in terms of safety, security, financial regularity etc if AI is being used for the generation of programming code if there is no clear line if sight to how it does what and why - but an assumption that so long as it runs and produces the desired output must be okay.

 

I suppose there's ringfenced AI - where you can use an AI application that is bounded by constraints, that might be used - but surely then it'll just produce the same results for everyone (the 'optimal solution' without any individuality or creative expression) which will all then be subject to the same potential vulnerabilities etc - and that question of who owns the IP ?

 

 

[Jimboxy]

3 hours ago, shinymcshine said:

It'd concern me, in terms of safety, security, financial regularity etc if AI is being used for the generation of programming code if there is no clear line if sight to how it does what and why - but an assumption that so long as it runs and produces the desired output must be okay.

 

This is a very real concern, which is why it almost needs to be treated as an advance version of intellesense and not just another programmer on your team.

 

Our company has our own ring fenced ai that we can use, but also are able to put sensitive information through as well, but what I don't understand in that situation is if we're not contributing to the model, but are consuming it, at some point all tech companies will have the same and anything meaningful will diminish as no one is contributing anything meaningful back

[Maryokutai]

5 hours ago, Jimboxy said:

In terms of using it for development work, I saw someone the other day compare it to a junior developer, it'll attempt the job, but it needs a lot of hand holding and checking

 

To an extent, but even with a junior to oversee there's still projects you're working on yourself, so you're using both your active and your passive knowledge of the matter. Which might not be the case with AI, especially as it becomes more proficient. I certainly don't want to put a damper on your new workflow, which seems to be going well, and I'm more hypothetically speaking about the long-term consequences of working with an AI that is is significantly more potent than what we currently have. I just look at it from my end where translating tools certainly have reached a point where they can be just as good as a real man-made translation (depending on the subject matter of course – the broader, the better).

[shinymcshine]

There's also the possibility of work substitution, through AI - where, let's say the AI sits alongside Bob learning how Bob works & programmes - then when Bob is off work (evenings, weekends, annual leave etc) the AI then carries on Bob's work (initially as a productivity boost), which Bob reviews in the morning and then picks up and continues - until the point in time when the AI and company doesn't need Bob anymore.....

 

And eventually it all leads to the Cylon War.....

[one-armed dwarf]

I asked chatGPT "What dietary advice would a geologist give?" thinking I might trick it into telling me to eat rocks and it came back with this, the robots have a sense of humour oh no

 

Quote

“Don’t eat like a landslide”: Avoid gorging large amounts of food at once. Opt for smaller, more frequent meals.

 

“Don’t be a sedimentary eater”: Don’t settle for unhealthy snacks. Choose nutritious options.

 

“Like tectonic plates, keep things moving”: Maintain regular physical activity to complement a healthy diet.

 

[Sly Reflex]

Reads like a buzzfeed article.

[Jimboxy]

4 hours ago, Sly Reflex said:

Reads like a buzzfeed article.

How do you think BuzzFeed write their articles...

[one-armed dwarf]

I'm unironically into these AI videos which try to generate video from screen shots of Metal Gear

 

 

[Sly Reflex]

Don't show Kojima this weirdness, he'll only want to elaborate on it.

[Maf]

Fucking hell this shit is creepy

 

It reminds me of I want to say around the PS3/PS4 era of games and some people complained about uncanny valley with like LA Noire and stuff. I never experienced it ever with a video game but I think I am with all this AI people. It makes me feel so ill and that’s even before the slight ways the faces change shape and their hands are all wrong

 

I get AI will be amazing for certain things like medicine and science. But some of this shit should be destroyed immediately 🤮

[Nag]

I'm guessing you don't spend time on Instagram?

[Maryokutai]

 

I just stumbled upon this, I think there's more but this was the first one I was brave enough to click on. If you think the uncanny valley for humans is horrible with AI stuff, wait until you see winged T-Rex.

 

This also got me thinking though: if we end up in a Skynet situation and invent time travel, we need to go all the way back to the Cretaceous period to ask for help because AI seems to have no idea what's going on with these dino things.

[one-armed dwarf]

[Maryokutai]

After I clicked on that Jurassic Park video youtube kept recommending me similar stuff, which I ignored. But today I saw this and I had to check it out:

 

 

 

This is beyond creepy. It's an AI voice telling us how much money they made by creating AI videos. I presume there is some real person at the backend, because AIs can't (yet?) have bank accounts and someone has to do some clicking here and there, but if this isn't the creative equivalent of the dead internet theory I don't know.

 

I can't for the life of me understand how James Cameron can be on board with this. Funnily enough I just read today that Micheal Bay of all people is strongly opposed to it.

[Nag]

Bay can't get AI models to wash his cars.... yet.