|
rotor posted:i thought this was a neat article. I think the article is sort of true but the prompts were stuff like: quote:Example prompt (2nd Photo): "A group of male Norse, Dane, and Vikings huddled together and is taking a group selfie picture together in 793 CE. They are drinking ale at a feast in a Viking longhouse. They are all wearing traditional Viking armor and helmets. Everyone smiling directly at the camera. The image is photorealistic, has natural lighting, and is taken with a front-facing phone selfie camera by one of the Vikings. --ar 3:2 --s 1000 --no phone --v 5 --q 2" E: to be clear the author of the article didn't write the prompt, that's from the OP on Reddit. Do not read the reddit discussion it's awful
|
#
¿
Apr 8, 2023 09:14
|
|
|
|
| # ¿ May 14, 2024 20:01 |
|
|
I don't understand why that even works - if a computer generated it shouldn't it be particularly easy to classify as well?
|
|
#
¿
May 23, 2023 14:00
|
|
|
Or is it actually just to generate/score training data
|
|
#
¿
May 23, 2023 14:01
|
|
|
The image manipulation (rather than pure creation) tools like that and the photoshop fill one definitely seem like the clearest "this will be extensively used by professionals" cases buying clothes online from brands you don't already know is going to go from hard to impossible
|
|
#
¿
Jun 6, 2023 21:56
|
|
|
Filtering suivellence footage seems like a real use case for ai https://www.bbc.co.uk/news/uk-england-cornwall-66508840.amp
|
|
#
¿
Aug 19, 2023 20:48
|
|
|
I used to work for a company that had a very specific problem, that ai solved nicely. We had a bunch of photos of flats/hotel rooms/houses all untagged and were showing hero images on a website by picking a random one (first alphabetically by however they were named). Unfortunately a non zero amount of the time this was a photo of the accomodations toilet, and our product people objected whenever they got a search or filter page full of toilet pics. The solution was to feed all our photos into our cloud providers generic image tagger, and mark any that came back as "toilet", "bathroom" etc. as not suitable for hero images. Worked perfectly!
|
|
#
¿
Sep 21, 2023 17:07
|
|
|
akadajet posted:what does tencentbot think about the tiananmen square massacre? I was kind of curious about this but the Baidu chatbot requires a chinese phone number. the economist claims that it says covid started with american vape users which imo makes it the best chatbot yet https://www.economist.com/business/2023/09/03/meet-ernie-chinas-answer-to-chatgpt
|
|
#
¿
Jan 11, 2024 16:16
|
|
|
https://www.newyorker.com/magazine/2024/02/12/a-teens-fatal-plunge-into-the-london-underworld Very long article in the new yorker about some dodgy London types. When it's revealed that one of them was into crypto and then AI it makes total sense
|
|
#
¿
Feb 6, 2024 22:07
|
|
|
I'm less instinctively hostile to the idea of "ai" than I think most people itt (although I haven't actually found a use for the current gen of llms) but at both my partner's and my own workplace there are people who seem to have been given a remit to introduce AI and are now going around looking for use cases. Not in a like "what are the orgs problems, and how can we solve them" way, but in a "how can we use AI to do something" way. It's less bad at mine since they are fairly segregated and leadership said at the start that it was mainly a marketing thing (not that that hasn't stopped the team self selecting into the most fervent true believers) but at hers they have actual power, and are really fudging the data to try and make it look useful. Like taking stats for work that was done were the ai tool was interacted with (because people are told to) and counting that as ai-produced work, even if the llm's output was entirely disregarded. They're trying to take a set of teams that are producing world leading output and get something that is mostly only capable of producing blogspam to replace them. e: spoke to some people and read some things and it turns out Gartner analysts absolutely love "ai" branded things, even for obviously nonsense use cases. distortion park fucked around with this message at 17:32 on Feb 7, 2024 |
|
#
¿
Feb 6, 2024 22:12
|
|
|
|
| # ¿ May 14, 2024 20:01 |
|
|
https://www.packback.co/ I did some quick research on reddit and you'll never guess who has to pay for it it's the students lol e: wow 
|
|
#
¿
Feb 8, 2024 15:40
|
|