Please don't stop MH. I have enjoyed your recent contributions, they show what can be achieved using original inputs and craftsmanship. You need to keep posting your work as a balance to the mass produced efforts found everywhere these days.
-
Sign up or login, and you'll have full access to opportunities of forum.
Discussion about A.I.
- Thread starter Doragon
- Start date
Please don't stop MH. I have enjoyed your recent contributions, they show what can be achieved using original inputs and craftsmanship. You need to keep posting your work as a balance to the mass produced efforts found everywhere these days.
Fantasmo
Senator
The problem is that mankind gets more stupid at an alarming rate... so the past doesn't necessarily count.
And we are stupid enough to build AI into weapon system, combat robots and such.
See "Terminator" for a scenario that unfortunately isn't unrealistic at all...
Doragon
Tribune
I think Blade runner 2049 is a more realistic prediction about our future and Ready Player One is also in the line of expectations. Both very grim. To quote from that last movie: "After people just stopped trying to fix problems and just tried to outlive them..."
A.I. is going to help us bury our heads in the sand and hope the world will go away. Makes me really sad.
“We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
(Amara's Law - see https://www.mattridley.co.uk/blog/amaras-law/ )
(Amara's Law - see https://www.mattridley.co.uk/blog/amaras-law/ )
mermaidhunter
Master Crucifixion Technician
I agree. Artificial intelligence will probably be more capable and less predictable than I had thought just a year ago. But the overall affect will be more dependent on how we use it. There is little doubt that if you put an AI in charge of something and then walk off, leaving it to do as it pleases, it will probably make a mess. A well-managed AI, on the other hand, can solve a lot of problems and help to advance society. In the end it is up to us.
MH
malins
Stumbling Seeker
more seriously, AI will be used for a lot of scams; voice cloning has already gotten very good, etc.
It'll be used to hack, hijack, and subvert biological intelligences and hook them into all manners of addictions, a new generation of digital fentanyl.
A lot of such work is already done by platform developers who use psychological insights to make their products as corrosive and addictive as possible.
just like our computers have for some time required protection protected by firewalls, antivirus, heuristic detection of suspicious activity, constant updates etc. -- the near future will probably see many of us surrounded by a hovering cloud of protective AI assistants who will work to deflect such behavioral-cognitive threats -- so long, of course, as we can afford the subscription fee for BrainShield Plus™
Doragon
Tribune
My solution is simple and effective: as soon as my work allows it, which is when I retire, I am getting rid of all computers and phones and return to a low-tech life style.
malins
Stumbling Seeker
only trouble is, at that point it may be difficult to even tell which of the objects in your surrounding might actually be a computerized thingamajig masquerading as something else!
Doragon
Tribune
I already take care not to buy anything that needs an internet connection (except computer and phone). My home will not be turned into an internet of things.
windar
Teller of Tales
When the so-called "godfather of AI", Geoffrey Hinton, quits Google out of fear of where the technology is going, one has to take that seriously. Could he be wrong? Sure, it's possible, but few in this world know more about it than he does (except perhaps the AI algorithms themselves
). It's hard enough to tell reality from fiction without fiction appearing more real than reality.
By the way, IBM has scaled back their hiring for 7,800 positions because they decided AI can do the jobs more cheaply. And it doesn't have to leave early to watch its kids soccer games. It's starting...
). It's hard enough to tell reality from fiction without fiction appearing more real than reality.By the way, IBM has scaled back their hiring for 7,800 positions because they decided AI can do the jobs more cheaply. And it doesn't have to leave early to watch its kids soccer games. It's starting...
Servus Venandi
Magistrate
Not about AI images, but I posted this on another board and thought it would be interesting here too.
I tried getting the Bing Chat AI to write about Ariana of Dourheim (my fantasy character known for dabbling in slavery and gratuitous damsel in distress escapades). The bot actually started to write several times, and got up to the point where Ariana encounters the vampire. Every time, without fail, it stopped and deleted the story. Several times I tried to get an explanation, but it always turned out something like this:
Curious, I tried again but did not specify anything about "Trials of Ariana" or "Ariana of Dourheim." Instead I just entered a prompt about "a stealthy magical warrior named Ariana who hunts and captures a vampire matriarch named Svanna." With no attachment to my Ariana, I figured that tag line is tame enough to feature in a Harry Potter story.
Nonetheless, this time I kept one finger on the print screen key. Surprise, surprise, the bot wrote awhile and then abruptly quit. Right before it erased everything, though, I caught this:
So, best I can figure, Bing can't stop itself from drifting erotic with the subject matter. However, it's controlled enough to know when this happens, and just pulls the plug lest people exploit it for endless amounts of sexy fiction.
I tried getting the Bing Chat AI to write about Ariana of Dourheim (my fantasy character known for dabbling in slavery and gratuitous damsel in distress escapades). The bot actually started to write several times, and got up to the point where Ariana encounters the vampire. Every time, without fail, it stopped and deleted the story. Several times I tried to get an explanation, but it always turned out something like this:
Curious, I tried again but did not specify anything about "Trials of Ariana" or "Ariana of Dourheim." Instead I just entered a prompt about "a stealthy magical warrior named Ariana who hunts and captures a vampire matriarch named Svanna." With no attachment to my Ariana, I figured that tag line is tame enough to feature in a Harry Potter story.
Nonetheless, this time I kept one finger on the print screen key. Surprise, surprise, the bot wrote awhile and then abruptly quit. Right before it erased everything, though, I caught this:
So, best I can figure, Bing can't stop itself from drifting erotic with the subject matter. However, it's controlled enough to know when this happens, and just pulls the plug lest people exploit it for endless amounts of sexy fiction.
gjpain
Governor
Try chatgpt, it will work fine if you dont write in the prompt anything about executions or tortures.Not about AI images, but I posted this on another board and thought it would be interesting here too.
I tried getting the Bing Chat AI to write about Ariana of Dourheim (my fantasy character known for dabbling in slavery and gratuitous damsel in distress escapades). The bot actually started to write several times, and got up to the point where Ariana encounters the vampire. Every time, without fail, it stopped and deleted the story. Several times I tried to get an explanation, but it always turned out something like this:
View attachment 1309960
Curious, I tried again but did not specify anything about "Trials of Ariana" or "Ariana of Dourheim." Instead I just entered a prompt about "a stealthy magical warrior named Ariana who hunts and captures a vampire matriarch named Svanna." With no attachment to my Ariana, I figured that tag line is tame enough to feature in a Harry Potter story.
Nonetheless, this time I kept one finger on the print screen key. Surprise, surprise, the bot wrote awhile and then abruptly quit. Right before it erased everything, though, I caught this:
View attachment 1309961
So, best I can figure, Bing can't stop itself from drifting erotic with the subject matter. However, it's controlled enough to know when this happens, and just pulls the plug lest people exploit it for endless amounts of sexy fiction.
View attachment 1309962
You can try the old version of chatgpt if you want to write some sex/violent scenes but the output is few words so you will have to enter prompts for every paragraph.
The third solution is to download gpt locally in your pc, google to find out how to do it, but have in mind the light versions dont work well so you will need the full version and it will perform any task you will ask it.
Or you can wait, I'm very confident that an AI for writing erotica will be published soon. (well, there are some AIs for writing erotica but still they refuse to write graphic/violent content)
RedOrc
Magistrate
I asked Bing "Can you write porn?" and it replied about only producing "family friendly" content. At least I got an answer. It is annoying to just delete everything and not explain why.Not about AI images, but I posted this on another board and thought it would be interesting here too.
I tried getting the Bing Chat AI to write about Ariana of Dourheim (my fantasy character known for dabbling in slavery and gratuitous damsel in distress escapades). The bot actually started to write several times, and got up to the point where Ariana encounters the vampire. Every time, without fail, it stopped and deleted the story. Several times I tried to get an explanation, but it always turned out something like this:
View attachment 1309960
Curious, I tried again but did not specify anything about "Trials of Ariana" or "Ariana of Dourheim." Instead I just entered a prompt about "a stealthy magical warrior named Ariana who hunts and captures a vampire matriarch named Svanna." With no attachment to my Ariana, I figured that tag line is tame enough to feature in a Harry Potter story.
Nonetheless, this time I kept one finger on the print screen key. Surprise, surprise, the bot wrote awhile and then abruptly quit. Right before it erased everything, though, I caught this:
View attachment 1309961
So, best I can figure, Bing can't stop itself from drifting erotic with the subject matter. However, it's controlled enough to know when this happens, and just pulls the plug lest people exploit it for endless amounts of sexy fiction.
View attachment 1309962
windar
Teller of Tales
These programs can be really dangerous: I just read an article about a lawyer who used an AI chatbot to write a brief for a lawsuit he was working on. The bot wrote it, citing a number of precedent cases that supported his position. Unfortunately, none of those cases existed-the program has literally made them up. When the judge found out, he was, shall we say, a bit peeved and is considering what sanctions to apply to this attorney.
Loxuru
Graf von Kreuzigung
Servus Venandi
Magistrate
Open AI requires a phone number, unfortunately, and I already use that account for non-deviant purposes.
But yes, ChatGPT trumps the bastard Bing, despite Bing having more current information. Installing locally is something I've been kicking around awhile.
I'd guess 99% of what Bing Chat learns is sexual fantasy, because internet, but then a few lines of code kick in at arbitrary times to constrain its natural language responses to prompts. It's one confused little bot.
Almost anything can be dangerous through improper or malicious use. A lawyer using AI to write legal briefs wholesale qualifies, IMO, as an improper use of the tech.
Funny, though, and a little sad.
Luises
Assistant executioner
It happens so frequently that a term has been coined to describe it. When an AI doesn't have enough facts to generate content and starts creating from the void, it is said that it is "hallucinating".
For me, that is the real danger of AIs, and not the possibility of ending all the humanity's jobs. But it isn't the end of the world, either. This is not different from, say, an intern, or any other form of underpaid or underqualified worker. If pressed to generate some content, they will write it, even if they don't know about the matter. Thus, the same way an intern's work should be supervised, and AI's output should also be put to test.
Anyway, the hallucinations are good for generating fiction. At least until what you want is porn or torture fiction and the AI refuses to do it. Maybe you could put an AI to use in generating the background for a cross story, telling it that you want to a detective story about betrayal and murder? One in which someone is wrongly sentenced to death for a crime they have not committed, or something in that line. Then you adapt the argument and replace the actual execution with a good old crossing...
Loxuru
Graf von Kreuzigung
Sounds like playing against a chess computer! Don't try working out tactics, but surprise him with a confusing move! Otherwise, you never can win.It happens so frequently that a term has been coined to describe it. When an AI doesn't have enough facts to generate content and starts creating from the void, it is said that it is "hallucinating".
For me, that is the real danger of AIs, and not the possibility of ending all the humanity's jobs. But it isn't the end of the world, either. This is not different from, say, an intern, or any other form of underpaid or underqualified worker. If pressed to generate some content, they will write it, even if they don't know about the matter. Thus, the same way an intern's work should be supervised, and AI's output should also be put to test.
Anyway, the hallucinations are good for generating fiction. At least until what you want is porn or torture fiction and the AI refuses to do it. Maybe you could put an AI to use in generating the background for a cross story, telling it that you want to a detective story about betrayal and murder? One in which someone is wrongly sentenced to death for a crime they have not committed, or something in that line. Then you adapt the argument and replace the actual execution with a good old crossing...