I watch a lot of crafty Youtube's and a fair amount of the crafters only film what they're working on, and never show their face (mostly women). We still get yoyo's asking for 'face reveals' and 'oh, but I want to see your emotion as you're working'. No! Absolutely not needed! I want to see the work
December 28, 2025 at 6:39 AM
I watch a lot of crafty Youtube's and a fair amount of the crafters only film what they're working on, and never show their face (mostly women). We still get yoyo's asking for 'face reveals' and 'oh, but I want to see your emotion as you're working'. No! Absolutely not needed! I want to see the work
That's the point at which humans started thinking they did the domestication.
December 25, 2025 at 1:57 AM
That's the point at which humans started thinking they did the domestication.
Better than Costco?
December 25, 2025 at 1:37 AM
Better than Costco?
Are you kidding? You sound like you're operating from that old timey "well, women don't really want to have sex" mindset. Those social taboos around premarital sex were largely enforced against women, not men, and there are indeed lots of women who enjoy sex without marriage.
December 24, 2025 at 2:22 AM
Are you kidding? You sound like you're operating from that old timey "well, women don't really want to have sex" mindset. Those social taboos around premarital sex were largely enforced against women, not men, and there are indeed lots of women who enjoy sex without marriage.
Small boats for small hands.
December 23, 2025 at 12:54 AM
Small boats for small hands.
German designed. That explains a lot.
December 22, 2025 at 9:48 PM
German designed. That explains a lot.
Like I said, common elements in the construction of the models doesn't mean they are the same thing. The two things that end up coming off the assembly line are very different.
December 22, 2025 at 5:26 AM
Like I said, common elements in the construction of the models doesn't mean they are the same thing. The two things that end up coming off the assembly line are very different.
I get it. You're talking about the training, and this thread is about the models themselves. I.e. the reason why nobody is using LLM's (or any sort of GenAI) to detect cancer cells. Why LLM's hallucinate so much to the point of uselessness, while ML inference has proven very useful.
December 22, 2025 at 5:26 AM
I get it. You're talking about the training, and this thread is about the models themselves. I.e. the reason why nobody is using LLM's (or any sort of GenAI) to detect cancer cells. Why LLM's hallucinate so much to the point of uselessness, while ML inference has proven very useful.
Yes. I am retired now, but I've been frustrated by the "AI experts" with superficial knowledge and a bunch of buzzwords who don't know anything about the history and mechanics of how these things work.
December 22, 2025 at 4:48 AM
Yes. I am retired now, but I've been frustrated by the "AI experts" with superficial knowledge and a bunch of buzzwords who don't know anything about the history and mechanics of how these things work.
That's like saying a Segway is a plane because they both have engines.
December 22, 2025 at 4:41 AM
That's like saying a Segway is a plane because they both have engines.
Absolutely they are not.
December 22, 2025 at 4:40 AM
Absolutely they are not.
Nobody is using an LLM to identify cancer cells. You're talking about ML (machine learning), which is nothing like an LLM. There is no way for an LLM to be able to label "potential hallucination" because there is no way to distinguish those pathways in an LLM.
December 22, 2025 at 4:37 AM
Nobody is using an LLM to identify cancer cells. You're talking about ML (machine learning), which is nothing like an LLM. There is no way for an LLM to be able to label "potential hallucination" because there is no way to distinguish those pathways in an LLM.
Fwuh?! If a reasonable candidate can't even get traction in a Democratic primary, why do you think all of a sudden it would be great in the general?
December 22, 2025 at 3:46 AM
Fwuh?! If a reasonable candidate can't even get traction in a Democratic primary, why do you think all of a sudden it would be great in the general?
Well, by Chinese/Japanese standards, neither can the English.
December 20, 2025 at 9:23 PM
Well, by Chinese/Japanese standards, neither can the English.
judgement that these Black people didn't have cancer when they did, because that's how the training data was given to it.
December 20, 2025 at 8:27 PM
judgement that these Black people didn't have cancer when they did, because that's how the training data was given to it.
Their cancers will be more advanced when samples are taken, and they are more often judged (by humans) to not have cancer when they do. The ML (not AI, they're just using a buzzword) model is not at fault, it detected a difference that was in the training data, and followed the human
December 20, 2025 at 8:27 PM
Their cancers will be more advanced when samples are taken, and they are more often judged (by humans) to not have cancer when they do. The ML (not AI, they're just using a buzzword) model is not at fault, it detected a difference that was in the training data, and followed the human
Likewise, when put into practice, the lab samples that were being evaluated didn't have demographic data attached. They say in the article that both the training samples and lab samples of Black patients had higher anomalies, and my guess is that's because Black people tend to get worse healthcare.
December 20, 2025 at 8:27 PM
Likewise, when put into practice, the lab samples that were being evaluated didn't have demographic data attached. They say in the article that both the training samples and lab samples of Black patients had higher anomalies, and my guess is that's because Black people tend to get worse healthcare.
It's because the training data was different for Black patients. No identifying demographic data (age, sex, race/ethnicity, etc.) was provided during training. But because the selection of samples (made by humans) for training was different, the model was able to detect that difference.
December 20, 2025 at 8:27 PM
It's because the training data was different for Black patients. No identifying demographic data (age, sex, race/ethnicity, etc.) was provided during training. But because the selection of samples (made by humans) for training was different, the model was able to detect that difference.
That's not true. Jay Inslee ran heavy on climate change, and the fact that you don't remember that says it all.
December 20, 2025 at 7:41 PM
That's not true. Jay Inslee ran heavy on climate change, and the fact that you don't remember that says it all.
Strangely, everytime a friend or relative comes to visit, it's always the best weather we've ever had. Definitely won't last.
December 20, 2025 at 12:42 PM
Strangely, everytime a friend or relative comes to visit, it's always the best weather we've ever had. Definitely won't last.
Bee-dee, bee-dee.
December 17, 2025 at 4:31 PM
Bee-dee, bee-dee.
I suspect he is referring to Venezuela nationalizing its oil and gas industries in the 70's. In other words, they "stole" the oil that American companies had previously "owned". But at best they might have a claim to the facilities and equipment of those companies, they never "owned" the oil.
December 17, 2025 at 1:28 AM
I suspect he is referring to Venezuela nationalizing its oil and gas industries in the 70's. In other words, they "stole" the oil that American companies had previously "owned". But at best they might have a claim to the facilities and equipment of those companies, they never "owned" the oil.
Better than the ones where applicant proudly touts his "projects" on github, which consist solely of the exercises from a popular "learn to code" course, and only about 2/3's are actually completed.
December 15, 2025 at 9:29 PM
Better than the ones where applicant proudly touts his "projects" on github, which consist solely of the exercises from a popular "learn to code" course, and only about 2/3's are actually completed.
It's a whole thing these days to make YouTube videos that are just someone reading Reddit posts.
December 15, 2025 at 8:27 PM
It's a whole thing these days to make YouTube videos that are just someone reading Reddit posts.
Wasn't sure if this was still the case. When I graduated in the 80's, the first year (and much of the second) was a lot of math background. Given the deficiencies noted in graduates these days, it would not surprise me if they ditched a lot of the math reqs. Vibe coding FTW!
December 15, 2025 at 10:59 AM
Wasn't sure if this was still the case. When I graduated in the 80's, the first year (and much of the second) was a lot of math background. Given the deficiencies noted in graduates these days, it would not surprise me if they ditched a lot of the math reqs. Vibe coding FTW!