Little Things That Make You Smile 7: The Joy Department

[Asura]

one thing that really used to p me off was the endless clipping paths id have to create to remove the background from an image, its one click now. They have taken a really mundane part of my job and made it a breeze which im all for they havent replaced me they just made my job easier. I'm not against these sort of things that make your life easier, take the mundane tasks away and give me more time for the actual creative stuff.

AI's a nuanced issue, which is part of the problem.

Photoshop has, for years now, had the "content-aware-fill" tool which is a sort of AI-based tool to speed up workflow. But that only uses data from within your image (or what you supply) and has to be used in the correct way, which takes some skill. It had no chance of "replacing" any creative; it mainly just allowed creatives to get a better result in a bit less time.

Adobe is closer than anyone else to getting past many of the ethical issues because of Adobe Stock; they're dealing with a huge dataset of people who have given permission to use their stuff. But I've used Adobe Stock prior to this and I know it has huge blind spots.

I'm seeing a lot of debate about this stuff online; same for ChatGPT - of whether AI is going to move in jumps, rather than a constant drip of improvement like most technologies. There's been the suggestion that ChatGPT might be nearing the "best it can be" for that type of language model, and while improvements are possible, they could offer diminishing returns until someone figures out the next big jump, and there's no idea when that will be.

AI generated imagery, especially for character illustrations from scratch, is the same; in terms of generating stuff from scratch, I keep up to date with this stuff, look at imageboards etc. where people post AI "art", and honestly, already, it's thousands and thousands of really quite similar images. In the anime space, you just see pages and pages of weirdly similar-faced anime women, usually with large boobs, rendered with backgrounds that look like knock-offs of Makoto Shinkai's work, all rendered in the same painterly style. And they all look dead; as-in, there's no through-line of action in the images. The character is just standing there.

Conversely, I'm seeing stuff which is interesting where people fold these tools into their workflows; but again, that's skilled work.

[Dogg Thang]

I bought a house and moved in a couple of months ago. Previous owners hadn’t cleared out the attic so I eventually tackled that. Total mess, full of crap. But buried among the rubbish was a bag of PS3 and XBox 360 games. I might never play them but they’re mine now.

[kryss]

Motherlode! Pics or it didn't happen. Also, the games too.

[Neon Ignition]

An AI expert was on the radio last week discussing the rise of online AI such as ChatGP and the stuff used for art and he said that the main area of focus for those concerned should be the tech giants who are publicly quick to call out the dangers of the current wave of AI growth but are working on advanced AI behind closed doors. The argument was that whilst we're seeing a surge of AI art, chat etc currently that appears scary it will ultimately subside because it's fundamentally flawed AI. Because much of what the AI is currently working from is internet based, it is pulling from a huge amount of unreliable data. As the data is based on human created data there's an unreliability factor it has to deal with, then as the years go on the internet is being flooded with content that AI itself is creating which dilutes the data massively. Basically, AI eventually eats itself too much. This is compared to the tech giants using it as a distraction tactic whilst they privately work on curated powerful AI which is the more likely to cut through entire employment sectors. Though, the argument there is that you get a trade off - like cancer research companies able to calculate new cures in a tiny fraction of the decades it takes humans to work through them etc.

[Lebowski]

the argument there is that you get a trade off - like cancer research companies able to calculate new cures in a tiny fraction of the decades it takes humans to work through them etc.

So where all gonna live longer but Ai and robots will be taking over all our jobs, I'm really not looking forward to the "lifestyle choice" of living in tent.

[importaku]

Pinball finally waved farewell to it and it magically turned into a housebrick sized pile of £20 notes, must resist urge to just head to airport and book a day trip to Japan for shopping

Will head to bank later to deposit it

Cha ching, most i have ever held in my hand at one time.

[Dogg Thang]

Motherlode! Pics or it didn't happen. Also, the games too.

I’ll send you a pic. It’s not a great collection though. The likes of Saints Row, GTAs, Modern Warfares and so on. That sort of gamer. Would probably fetch me €1 in Cex, which is where I think most of them came from in the first place.

[Zaki]

https://mediaassets.lex18.com/cordillera-network/wp-content/uploads/sites/4/2019/01/27095458/poster27.jpg

Cha ching, most i have ever held in my hand at one time.

[importaku]

Just don't turn on the news

[Asura]

An AI expert was on the radio last week discussing the rise of online AI such as ChatGP and the stuff used for art and he said that the main area of focus for those concerned should be the tech giants who are publicly quick to call out the dangers of the current wave of AI growth but are working on advanced AI behind closed doors. The argument was that whilst we're seeing a surge of AI art, chat etc currently that appears scary it will ultimately subside because it's fundamentally flawed AI. Because much of what the AI is currently working from is internet based, it is pulling from a huge amount of unreliable data. As the data is based on human created data there's an unreliability factor it has to deal with, then as the years go on the internet is being flooded with content that AI itself is creating which dilutes the data massively. Basically, AI eventually eats itself too much. This is compared to the tech giants using it as a distraction tactic whilst they privately work on curated powerful AI which is the more likely to cut through entire employment sectors. Though, the argument there is that you get a trade off - like cancer research companies able to calculate new cures in a tiny fraction of the decades it takes humans to work through them etc.

This is true, to a degree.

It's like I said elsewhere; ChatGPT doesn't understand what it's saying. Those image generators don't know what they're drawing. So getting the AI to do anything with what it makes is presently impossible, without a lot of human intervention.

Just like those demos of NPCs using ChatGPT in Skyrim; they're ALL bull****. I mean, they're not lying, they do work inasmuch as you see, but they don't actually function as far as the game is concerned. The characters can't connect the words they're saying to the things they're seeing/doing with any degree of reliability, and any you see which look like they're doing that are just careful editing and prompting to get the AI to say something that looks good in a YouTube video.

[Hirst]

I use ChatGPT to write response Emails at work, but specifically only to people I have no real respect for and don't really care about. I have been doing this for the last few months with great success.

My usual prompts are along the lines of "Write an Email response to a manager who was critical of the way you handled [situation] due to [issue], mitigating circumstance being [this]. You are a British office worker. Do not accept any blame or promise anything. General feeling of "it is what it is" but do not openly state it. Friendly, but non-committal. Three paragraphs max. More of a statement than an invitation for further dialogue".

It mostly dishes out complete word salad, enough to fob off idiot middle-managers as it sounds like one of them. Would I use it for any kind of creative process? No, but I did use it to write a scene from Carry On Camping where Sid James goes nuts and starts smashing all the tents up with his bare fists.

[charlesr]

https://www.bbc.co.uk/news/health-67295526

[Asura]

https://www.bbc.co.uk/news/health-67295526

This is a bit of an aside, but it reminds me of back when the Prince of Wales Innovation Prize used to be on TV; it's an award given out annually to British inventors who have made innovative stuff.

But I remember the year that Smart Water was invented. It's a chemical dye put into sprinklers which sprays out when burglar alarms go off, which is harmless, but dyes a person's skin with colours that can only be seen under a specific wavelength of UV light, and doesn't fully wash off for up to a year. The idea is if the police arrest a person, they can shine the light on them to see if they have the dye on their skin, and it can be tinted with certain colours so you know they were covered at that specific break-in. A lot of very high-value facilities use it, like NHS drug storage and the like.

But then the next exhibitor was showing the first version of those prosthetic legs that used AI computer controls and sensors to match the user's gait. You could see from the videos that compared to older false legs, this was an absolute revolution; a person could walk with one while clothed and you would literally have no idea they were wearing a prosthetic.

And I literally remember the presenter getting a bit editorial saying that he felt for the Smart Water guys as their thing was doubtless going to be a big success... But they didn't literally make the lame walk again so the winner was pretty obvious

[Dogg Thang]

This is true, to a degree.

It's like I said elsewhere; ChatGPT doesn't understand what it's saying. Those image generators don't know what they're drawing. So getting the AI to do anything with what it makes is presently impossible, without a lot of human intervention.

Just like those demos of NPCs using ChatGPT in Skyrim; they're ALL bull****. I mean, they're not lying, they do work inasmuch as you see, but they don't actually function as far as the game is concerned. The characters can't connect the words they're saying to the things they're seeing/doing with any degree of reliability, and any you see which look like they're doing that are just careful editing and prompting to get the AI to say something that looks good in a YouTube video.

Firstly, this only describes the current state of AI, which is evolving at a pretty incredible rate so it won’t be true for long. Secondly, the point about its understanding quickly becomes irrelevant when it functions to a degree where that understanding no longer matters. Google translate doesn’t actually know what it’s translating. It doesn’t understand the words or the context or think about the content in any way that could be considered understanding and yet it doesn’t matter for most use cases. Yes, a person who does have that understanding and fluency in two languages will do a much better job as a result but, for most use cases, Google translate wins due to speed and the sheer amount of languages it can have a go at. Give it time and it will likely be improved to do a way better job, and yet still without actually understanding what it’s doing. That won’t matter at that point.

When it’s getting the results, the level of understanding is really only a point of curiosity or something to be debated.

[Neon Ignition]

If they want to signal the dangers of AI and slow its progression/adoption then they need to rename it to something that easily and concisely communicates the danger.

I put forward - Skynet