So he was on a leave of absense, and now is no longer on such a leave? This is only an attempted coup if you think the leave of absense was a pretense, but they give no evidence of that in the article. I saw it suggested that he usually has wild ideas after returning from burning man, and the board wanted to insulate the company from the effects of that. That he's now back is evidence for a theory along those lines, and contradicts a theory that the board was trying to get rid of him. But I have no insight beyond that.
So murderers put their guns in a bag and get away with it, normal people who have "nothing to hide" leave their guns out in the open and get the police called on them.
Idk if soundcore is the same quality but it's pretty good and half the price of the base airpods. Hard to justify paying twice as much and not getting any ANC.
Fine tuning is not necessarily oriented at making a model better at generic "intelligence" tasks. In fact, the goal is often specifially to improve model performance at narrow tasks while recognizing it will get worse at other tasks.
I clicked through the leaderboard. It seemingly isn't sorted by best or worst performing, but rather by absolute change in performance - improved and worsened models are interleaved. Regardless, I selected the model that worsened the most, and investigated what they claimed their fine tuning accomplished; in fact they claimed nothing, their model sheet was the default trl model sheet without even the training set listed. The second worst performing was the same, but they at least listed their training set. It was a set of millions of word math problems. So we would expect the model to get better at those problems and worse at everything else.
On the other hand, the most improved model on the table had an actual model sheet, where they claim the model has been trained to be "more helpful". They claim that by ablating model refusal the model ends up doing better on benchmarks as well. That claim seems to have been borne out.
Obviously if you train a model for some purpose it will not necessarily do well at that purpose. But I would expect popularily used models to be ones that actually do what they say they do. I don't think people need to be told that sometimes models are badly trained. And the claim that this project actually has data to support, that finetuning on one task does not necessarily improve performance on another task, is so obvious that it doesn't need to be stated.
Nothing but blood pressure and heart rate was significant, and heart rate only barely so. I wouldn't read anything into the other numbers. It was a 20 day study, you aren't going to get long term outcomes from that.
One seemingly deliberately conflation this post commits is betwee "gypsy" and "roma". The first as a set of actions, the second as an ethnicity divorced from any action of behaviour. Dhh refers to the first, and uses the first word. The fine article interprets it to mean the second, and correspondingly substitutes the second word. But dhh makes it clear what he means by "gypsy", so one can only interpret this conflation as deliberate.
reply