terms like "oversocialized" suggest you spend too much time on imageboards and you would do well to get off those sites. same with "humiliation ritual"
You can get neotame online, including on Amazon. I really dislike it though. Its sweetness has a very strange attack and decay, if that makes sense, so that there isn’t really a right amount to use where it won’t taste some combination of too sweet and not sweet enough over the course of a bite or sip.
"One of the common solvents, ethyl acetate, comes naturally in many foods and beverages. It’s considered a safe chemical for decaffeination by the Food and Drug Administration."
speaking from a chemist's perspective, ethyl acetate is basically non-toxic. it decomposes into ethanol and acetic acid (vinegar). dcm (methylene chloride) is toxic, but should be completely removed before it ever reaches you.
the nice thing about volatile organic solvents is that they're...volatile, so they are easily stripped from coffee beans.
it does. that's what the "direct preference" part of DPO means. you just avoid training an explicit reward model on it like in rlhf and instead directly optimize for log probability of preferred vs dispreferred responses
What is it called when humans interact with a model through lengthy exchanges (mostly humans correcting the model’s responses to a posed question to the model, mostly through chat and labeling each statement by the model as correct or not), and then all of that text (possibly with some editing) is fed to another model to train that higher model?
I don’t think that process has a specific name. It’s just how training these models works.
Conversations you have with like chatgpt are likely stored, then sorted through somehow, then added to an ever growing dataset of conversations that would be used to train entirely new models.
Aaron has some biases but could be worth watching his reality check report from december 2023 "OpenAI, Inc. Video Reality Check: Decoding Non-Profit Tax Returns and Business Structure": https://youtu.be/PdQxPQf_1EQ
"They are smart people. About that I have no doubt. But I’ve seen how they treat rules. I know what they think of ethics.
Good idea?
Enter Larry Summers, who was just added to OpenAI’s Board of Directors—now that all the women are gone—to be Sam’s babysitter, so far as I can tell. Why? Because with Sam at the helm, OpenAI clearly violated its non-profit charter, repeatedly. The company is likely looking at an impending investigation by the California Attorney General, which has authority over tax-exempt organizations in the state, and perhaps an uncomfortable audit by the Internal Revenue Service as well. By having Larry on the Board, the company can plead that there’s now an “adult” in the house"
i get the opposite take from this info. the fact that, in just 10,000 years, the gene spread to over 1/3 of the population shows how useful of an adaptation it is.
Of course it’s useful. Particularly while kids are young, it’s a nutrient-dense food source that helps fuel immune systems being uniquely taxed by our post-agrarian densification (and cultivation of farm mammals immunologically similar to us, thereby giving rise to plagues).
The fact that it’s new means it’s necessarily loosely integrated. So if you aren’t in a calorie-deficient dietary environment, it follows that it can be assumed to not be an ideal nutrient source (even while remaining an ideal food source). And we see that. Milk consumption causes all manner of inflammatory responses. Because of course. It’s, in essence, another mammal’s filtered blood.
Please, for the love of God, can people please stop making me pry their point from their hands? What about astroturfing makes reddit, a platform where you can build your own walled garden, a platform where you cannot find content that isn't filtered through US hegemony?
