- Posts
- 2,281
- Likes
- 6,537
Deafboy
··His Holiness Puer SurdusI use Perplexity a lot now for internet searches, and a bunch of other uses. I used to use Google all the time for internet searches. Now, Perplexity 95% of the time.
Is AI for everybody?
- Posts
- 2,440
- Likes
- 3,315
Donn Chambers
·Some of my students use AI to try to do their research and writing. The quality of the work produced is abysmal, and it is easy to tell that it is AI. AI cannot quote and cannot cite its sources. The writing itself is almost like reading something translated from English into another language and back to English. As a result of this uptick in cheating, that is what it is, I have had to be much more thorough when I describe what constitutes plagiarism because these kids don't understand that having an AI write a paper is the same as paying a service to write one.
I haven’t seen this in any of my students writing (yet), but I have seen it in some colleagues (unfortunately). I’m working on a committee at my university discussing the creation of a new college by merging some units from different colleges. One of the members presented some text as a suggestion for a new mission statement that she generated with the help of ChatGPT. It was was just word vomit that said nothing — lots of catch phrases and jargon without substance. We chose to ignore that suggestion and moved on with writing it ourselves.
There has already been an issue with AI-generated text in a published scientific paper:
https://www.technologynetworks.com/...s-paper-with-ai-generated-introduction-384837
It was discovered because the authors (and editor) missed a common AI-phrase and left it in the opening section: “Certainly, here is a possible introduction for your topic.”
- Posts
- 3,423
- Likes
- 9,286
Waltesefalcon
·I haven’t seen this in any of my students writing (yet), but I have seen it in some colleagues (unfortunately). I’m working on a committee at my university discussing the creation of a new college by merging some units from different colleges. One of the members presented some text as a suggestion for a new mission statement that she generated with the help of ChatGPT. It was was just word vomit that said nothing — lots of catch phrases and jargon without substance. We chose to ignore that suggestion and moved on with writing it ourselves.
There has already been an issue with AI-generated text in a published scientific paper:
https://www.technologynetworks.com/...s-paper-with-ai-generated-introduction-384837
It was discovered because the authors (and editor) missed a common AI-phrase and left it in the opening section: “Certainly, here is a possible introduction for your topic.”
- Posts
- 1,287
- Likes
- 2,841
OllieOnTheRocks
·I use Perplexity a lot now for internet searches, and a bunch of other uses. I used to use Google all the time for internet searches. Now, Perplexity 95% of the time.
- Posts
- 1,327
- Likes
- 4,951
ahartfie
··The black sheep in the Spee-ee-eee-eedmaster flockSo far we’ve mostly been talking about words, but AI is also skewing serious art.
One of my friends on IG (misterenthusiast) is a Real Artist who recently published a book of AI-enhanced photographs of life in America in the 40s and 50s. The pictures are full of vivid color and rich detail. I’m sure they were a lot of work to compose.
I find them unbearable. I’ve done a lot of thinking about it and I think it’s because they look so real but my brain knows they are fake, and my logical processor errors.
ETA: Link to book
One of my friends on IG (misterenthusiast) is a Real Artist who recently published a book of AI-enhanced photographs of life in America in the 40s and 50s. The pictures are full of vivid color and rich detail. I’m sure they were a lot of work to compose.
I find them unbearable. I’ve done a lot of thinking about it and I think it’s because they look so real but my brain knows they are fake, and my logical processor errors.
ETA: Link to book
Edited:
- Posts
- 1,287
- Likes
- 2,841
OllieOnTheRocks
·So far we’ve mostly been talking about words, but AI is also skewing serious art.
One of my friends on IG (misterenthusiast) is a Real Artist who recently published a book of AI-enhanced photographs of life in America in the 40s and 50s. The pictures are full of vivid color and rich detail. I’m sure they were a lot of work to compose.
I find them unbearable. I’ve done a lot of thinking about it and I think it’s because they look so real but my brain knows they are fake, and my logical processor errors.
- Posts
- 1,327
- Likes
- 4,951
ahartfie
··The black sheep in the Spee-ee-eee-eedmaster flockOh now I want to see them!
I just added a link. Have at it.
- Posts
- 5,370
- Likes
- 9,157
BlackTalon
··This Space for RentWhere’s my slide rule? 😁
- Posts
- 1,986
- Likes
- 5,853
hen
·So far we’ve mostly been talking about words, but AI is also skewing serious art.
One of my friends on IG (misterenthusiast) is a Real Artist who recently published a book of AI-enhanced photographs of life in America in the 40s and 50s. The pictures are full of vivid color and rich detail. I’m sure they were a lot of work to compose.
I find them unbearable. I’ve done a lot of thinking about it and I think it’s because they look so real but my brain knows they are fake, and my logical processor errors.
ETA: Link to book
I see what you mean, they are unsettling.
And music too! I just got a song mocking me for drinking beer in the sun and other worse things in Ska beat. It took my neighbour all of 4 minutes to make with some AI.
I will try to link it, it is in Norwegian. https://www.udio.com/songs/uCvKavNd...xHF96W9ZVZP1UKKFN4Xkc9BldXgFBv7g5zhNm9lL5_tdW
- Posts
- 1,327
- Likes
- 4,951
ahartfie
··The black sheep in the Spee-ee-eee-eedmaster flockAnd music too! I just got a song mocking me for drinking beer in the sun and other worse things in Ska beat. It took my neighbour all of 4 minutes to make with some AI.
I heard it just fine.
There I Ruined It on IG does amazing AI mashups of songs that don’t go together. For some reason those don’t make me cringe.
- Posts
- 62
- Likes
- 348
sci
·I don't like how rapid this devopment is, but denying it is not gonna stop it. I also don't like the cloud model with everything tracked and answers coming from censured and subjective model, manipulating sometimes the factual truths. However, I found the offline models, with them you have full control over the privacy, and you can also choose more objective models of your liking. Currently I use WizardLM-2.
Edited:
- Posts
- 2,358
- Likes
- 3,755
sheepdoll
·The cloud model is a step backwards to the 'dumb' terminals of the 1960s and 1970s up to the 1980s. Rent time.
I was probably one of the first people in the 1980s to default a CC. Not really something to be proud of. I got addicted to Byte Information eXhange. Bix. There was also the source, and of course Usenet. The latter was free if one could find the right phone number to a 'live' node.
I remember when they wanted to repossess my TimeNet. Of course there was nothing to repossess.
The BBS craze with connections to Usenet were a better and 'free' alternative in the 1990s.
May also have been the first 'Lurker' Without spilcheck I was not posting 5000 word replies. Bix had an early chatroom called "Writers, Sleazy bar," where well known SF writers would hang out. This had the virtual ferns. Since I would see these people at conventions and know what was being discussed, Someone once asked in a post. "Who is that lurking behind the virtual ferns?" I had told them I was hiding behind the virtual ferns at one of the conventions.
Speculations about AI was a frequent topic. It was always interesting to meet online people in person. Of course you knew they were 'real'
I had noticed back in college, that people tended to trust a computer because it was a computer. There was a twilight zone episode about someone falling in love with a computer. I saw this in some of my classmates. One would get mad if someone tried to play the song Daisy on the printer. The printers used drums and chains, which ran at high speeds. a series of interposers would engage and the printers would chunk out a line at a time.) If one programmed things just right crude music could be played on a nearby transistor radio.
I still think AI is like a wax model. Detailed on the outside. When one cuts it open, there is nothing on the inside.
I was probably one of the first people in the 1980s to default a CC. Not really something to be proud of. I got addicted to Byte Information eXhange. Bix. There was also the source, and of course Usenet. The latter was free if one could find the right phone number to a 'live' node.
I remember when they wanted to repossess my TimeNet. Of course there was nothing to repossess.
The BBS craze with connections to Usenet were a better and 'free' alternative in the 1990s.
May also have been the first 'Lurker' Without spilcheck I was not posting 5000 word replies. Bix had an early chatroom called "Writers, Sleazy bar," where well known SF writers would hang out. This had the virtual ferns. Since I would see these people at conventions and know what was being discussed, Someone once asked in a post. "Who is that lurking behind the virtual ferns?" I had told them I was hiding behind the virtual ferns at one of the conventions.
Speculations about AI was a frequent topic. It was always interesting to meet online people in person. Of course you knew they were 'real'
I had noticed back in college, that people tended to trust a computer because it was a computer. There was a twilight zone episode about someone falling in love with a computer. I saw this in some of my classmates. One would get mad if someone tried to play the song Daisy on the printer. The printers used drums and chains, which ran at high speeds. a series of interposers would engage and the printers would chunk out a line at a time.) If one programmed things just right crude music could be played on a nearby transistor radio.
I still think AI is like a wax model. Detailed on the outside. When one cuts it open, there is nothing on the inside.
- Posts
- 6,794
- Likes
- 12,825
Evitzee
·Earlier today I was reading some analysis of future power needs for the AI revolution, it is astronomical. An AI-enabled Google search boosts the energy per search tenfold. A high speed AI chip like the 5 million turned out by Nvidia over the past three years uses as much electricity per year as three EVs. And the demand for these chips is explosive and seemingly unlimited. They'll get more efficient but the demand will be far greater. It will be interesting to see what happens on the power front that may blunt the AI revolution.
- Posts
- 3,132
- Likes
- 20,575
Dsloan
·Earlier today I was reading some analysis of future power needs for the AI revolution, it is astronomical. An AI-enabled Google search boosts the energy per search tenfold. A high speed AI chip like the 5 million turned out by Nvidia over the past three years uses as much electricity per year as three EVs. And the demand for these chips is explosive and seemingly unlimited. They'll get more efficient but the demand will be far greater. It will be interesting to see what happens on the power front that may blunt the AI revolution.
I've seen some similar write-ups about generative AI for video. Right now, it seems like we're in the phase where beta services are being provided free or cheap to get people interested, mine data, and improve the platforms. But at some point, the companies have to make money and the computational demands are massive. Could end up being surprisingly expensive and not offer a huge value over just buying stock footage for some things.
- Posts
- 4,823
- Likes
- 14,460
DaveK
··Yoda of YodelersAI as we know it in the form of chatbots is just companies making up for not wanting to pay for entry level positions. Yet, they are annoying their customers in the meantime. Usually the AI searches and summaries are useless. Now they are stealing facial and vocal patterns. Something needs to be done to reign this in before we do have a SkyNET situation going on, but it will be like WALL-E more than terminator. As much as I hate government regulations, esp in a global situation, the way it's handled now is untenable.
Those chat bots can be helpful, but can also be wrong as this story conveys
https://www.cbc.ca/news/canada/british-columbia/air-canada-chatbot-lawsuit-1.7116416
- Posts
- 1,986
- Likes
- 5,853
hen
·Those chat bots can be helpful, but can also be wrong as this story conveys
https://www.cbc.ca/news/canada/british-columbia/air-canada-chatbot-lawsuit-1.7116416
The company chat bot sold the more expensive tickets, no surprise there. Air Canada blaming the chat bot as a legal entity, well really no surprise there either, but god damn, how dare they!
- Posts
- 4,823
- Likes
- 14,460
DaveK
··Yoda of YodelersThe company chat bot sold the more expensive tickets, no surprise there. Air Canada blaming the chat bot as a legal entity, well really no surprise there either, but god damn, how dare they!
Yah, it’s all fun and games until someone has to take responsibility 😀 It was actually a bit surprising that they just blamed the bot and took zero responsibility for them selves. It reminds me of how Sears had in-store Sears Travel and Sears Home Reno desks etc, but if something went wrong they’d say “oh that’s not us, that’s Sears Travel, they are a different company”.
