Posts by janellecshane@wandering.shop
(DIR) Post #AXE8HdU72EMx5ylXGq by janellecshane@wandering.shop
2023-06-30T19:33:38Z
2 likes, 0 repeats
Nobody should be using GPT detectors for anything important.From a recent study that found that GPT detectors were misclassifying writing by non-native English speakers as AI-generated 48-76% of the time (!!!), compared to 0%-12% for native speakers.https://www.aiweirdness.com/dont-use-ai-detectors-for-anything-important/
(DIR) Post #AXE8HewVc8g5cL3jAe by janellecshane@wandering.shop
2023-06-30T19:36:35Z
0 likes, 1 repeats
The paper showed that asking chatGPT to reword an essay will evade the detector.Assuming they know of the existence of GPT detectors, a student who uses AI to write or reword their essay is LESS likely to be flagged as a cheater than a student who never used AI at all.
(DIR) Post #AZBwPEgIFjViMdVOcq by janellecshane@wandering.shop
2023-08-28T13:54:00Z
0 likes, 0 repeats
Image recognition vs Giraffe With No Spots: they're terrible at it?Is it because the giraffe (born July 31 2023) is too young to have been in their training data?https://www.aiweirdness.com/ai-vs-a-giraffe-with-no-spots/
(DIR) Post #AZlsc6iPCSHobuRG2y by janellecshane@wandering.shop
2023-09-14T19:40:59Z
0 likes, 0 repeats
can anyone recommend a good book about relativity/math/astrophysics for an 8 year old? Or as the kid put it, “Einstein math”
(DIR) Post #AZwUOKwFIaZGs2nTkG by janellecshane@wandering.shop
2023-09-16T22:13:16Z
0 likes, 0 repeats
I called my senators to oppose the Kids Online Safety Act. I've heard it has some bipartisan support now, so even if you have democratic senators they need to hear from you.https://wandering.shop/@charliejane/111002680357125260
(DIR) Post #Aa5xiUcdS4XMhsID1U by janellecshane@wandering.shop
2023-09-24T13:25:47Z
0 likes, 1 repeats
Chatbots like GPT-4 and Bard will "retrieve" an explanation of a meme even if it doesn't exist.https://www.aiweirdness.com/trolling-chatbots-with-made-up-memes/
(DIR) Post #Aaa9nhHjBaD5yxgO8G by janellecshane@wandering.shop
2023-10-09T03:53:45Z
0 likes, 1 repeats
Anyone going to be in Dallas on Oct 17? I'm giving a free public talk at the Perot Museum of Nature and Science in Dallas, Texas. I love a museum talk!(and look it's @AstroKatie !)https://www.gresearch.com/blog/article/g-research-distinguished-speaker-series-with-janelle-shane/
(DIR) Post #Aae0Q8kZIV9LG6hkUy by janellecshane@wandering.shop
2023-10-10T18:36:09Z
0 likes, 0 repeats
https://www.latimes.com/business/technology/story/2023-03-31/column-afraid-of-ai-the-startups-selling-it-want-you-to-be
(DIR) Post #AbBOfrYldClUGelv8a by janellecshane@wandering.shop
2023-10-27T03:27:41Z
0 likes, 2 repeats
it's halloween and this is your loot(generated these with #dalle3 because this is where the weird ai-generated text lives now)https://www.aiweirdness.com/choose-your-candy/
(DIR) Post #Abh19FX5pkcqOKY23E by janellecshane@wandering.shop
2023-11-11T01:44:05Z
0 likes, 1 repeats
Learn your fruits and vegetables with the help of DALL-E3!https://www.aiweirdness.com/learn-your-fruits-and-vegetables/
(DIR) Post #AbhW7wuag5SKveOm2a by janellecshane@wandering.shop
2023-11-11T13:17:42Z
0 likes, 0 repeats
learn the berries#dalle3
(DIR) Post #AbhW7xr5AcFBr3lUDA by janellecshane@wandering.shop
2023-11-11T13:22:32Z
0 likes, 1 repeats
learn the berries in swedish#dalle3#thoseTrickyTripleUmlauts
(DIR) Post #Ac1DPqdA0t56KTikt6 by janellecshane@wandering.shop
2023-11-21T02:43:59Z
0 likes, 1 repeats
#dalle3 more: https://www.aiweirdness.com/thanksgiving-table/
(DIR) Post #AcO2435cNdcRjtVr6W by janellecshane@wandering.shop
2023-12-01T14:38:29Z
0 likes, 1 repeats
It's time for an AI-generated advent calendar!I'm only doing this because #dalle3 is terrible at labeling images and I want everyone to know it.https://www.aiweirdness.com/ai-weirdness-advent-calendar-2023/
(DIR) Post #AcO2453v2vRrrCjWjY by janellecshane@wandering.shop
2023-12-01T14:39:28Z
0 likes, 0 repeats
Dec 1 (AI-generated food(?))
(DIR) Post #AdfYXw8eqQkWMTlJL6 by janellecshane@wandering.shop
2024-01-09T03:21:59Z
0 likes, 1 repeats
learn the mammals with the help of dall-e3! https://www.aiweirdness.com/learn-the-mammals-with-dall-e3/
(DIR) Post #AlOaBccU40yNenI6Ay by janellecshane@wandering.shop
2024-08-26T13:32:56Z
0 likes, 0 repeats
you may not like it but this is what peak language model performance looks like https://www.aiweirdness.com/new-paint-colors-invented-by-neural-17-05-17/amp/
(DIR) Post #At1z8rIJ5IkiwVFxiq by janellecshane@wandering.shop
2025-04-12T23:46:11Z
3 likes, 7 repeats
“Slopsquatting” in a nutshell:1. LLM-generated code tries to run code from online software packages. Which is normal but 2. The packages don’t exist. Which would normally cause an error but3. Nefarious people have made malware under the package names that LLMs make up most often. So4. Now the LLM code points to malware.https://www.theregister.com/2025/04/12/ai_code_suggestions_sabotage_supply_chain/
(DIR) Post #At1z8zOcxgYu4QH7gW by janellecshane@wandering.shop
2025-04-12T23:58:18Z
0 likes, 0 repeats
no problems ever with mimmic software packige Can trust yes
(DIR) Post #Ax3X3psYOESSHt7sGm by janellecshane@wandering.shop
2025-08-08T16:18:36Z
0 likes, 1 repeats
New AIWeirdness post: ChatGPT will apologize for anything, including stuff that didn't happen. It's not reflecting on stuff it did wrong, it's improv.https://www.aiweirdness.com/chatgpt-will-apologize-for-anything/