Given that tokens are supply constrained right now for Anthropic and OpenAI (especially a problem for Anthropic), stepwise efficiency advances for either would give it a leg up on the other. It would also help them better compete on price with Chinese models.
Given that neither company releases parameter counts, that sort of information would be slow coming out anyway. The most important thing is improvements in actual performance/ benchmark numbers, which allow them to preserve their price points as much as possible.
The best thing to do is to tell the model itself what you'd like from it, and to help you craft a set of user preferences. Every so often I'll have a new suggestion for it, and it will revise the entire set for me. My Claude user prefs is around 1000 words and probably overkill, but I don't think it hurts besides using a few more tokens to start a chat.
Here's my guess- your post reflects your honest opinion on the matter, with some LLM help. It elaborated on your smartphone analogy, and may have tightened up the text overall.
Because historically we haven't handled mental injuries as well as the physical ones. I don't completely disagree with your original points. I think depth, nuance, and accuracy of the conversation matters most of all. There is plenty of superficial, influencer-level chatter in both realms.
If you are interested in learning more, there is a podcast series by New York Magazine called Cover Story: Power Trip that goes pretty deeply into MAPS. It is just one perspective on it, but a very valid one. Many of these concerns have been raised within the psychedelic community for years but basically brushed aside by MAPS. It all came out yesterday though.
(my personal opinion is that psychedelics are or can be of great therapeutic benefit, but the MAPS/Lykos clinical trial was definitely problematic.)
There might be some deliriant that produces altered states of mind, such that a patient with no MDMA/psychedelic experience could be blinded. That could also potentially work for the therapist or facilitator. If the control was essentially a sugar pill, I think any reasonably experienced therapist or facilitator would easily recognize whether a patient is under the influence of a psychedelic.
It only makes sense for Deckard to be a replicant — something that Roy would know, and choose not to tell.
Maybe I should have recognized this earlier, but I really like this point and feel it adds depth- whether or not the screenwriters or Ridley Scott even intended it. It's not necessary for this to be the case for Batty to spare Deckard, it's also possible that recognizing his end was near he wanted to pass something on and/or communicate his life in some way to another.
I think the movie is definitely set up with the premise that humanity is a stabled fixed point of reference, with Deckard representing that viewpoint. But once you had Deckard falling in love with Rachel, and then the ending with Deckard and Batty, I think the movie is asking viewers to question this initial premise.
Within the context of a big budget Hollywood sci-fi actioner, I think it's about as reasonably deep as you could expect. None of its ideas are profound or new, but the cinematography and set design were.
Replicants can try to fake the appropriate emotional reaction — might even be programmed to fake it to themselves, such that even they don’t realize what they are — but there is always a split-second delay, which the trained operator can detect.
This makes sense with replicants who know they are replicants and are trying to avoid detection (I'm not sure why they would ever be 'programmed' to try to fake detection, the simpler explanation is that they know what the test is and are trying to avoid identification and death.) But it's different for Rachel. She thinks she is human and wouldn't be trying to game the test, yet it takes a long time for Deckard to assess her. So why is that? Either way...
...leading one to suspect that the Voight-Kampff test may not be measuring pure, unadulterated empathy in quite the way everyone seems to think it is....But in Ridley Scott’s preferred version of the film, the central theme falls apart, to be replaced with psychological horror’s equivalent of a jump scare: “Deckard himself is really a replicant, dude! What a mind fuck, huh?"
Those 2 statements contradict themselves, because Deckard himself being a replicant (who actually is never tested himself anyway) is an extension of blurring the lines between human and replicant, and what the Voight-Kampff test actually is testing (which already happens when Deckard tests Rachel.)
I feel like the idea of next-gen replicants like Rachel (and possibly Deckard) that blur the lines further is thematically consistent with the rest of the movie. I don't know what kind of social life Rachel has led, but if she has any friends or social interaction at all some sort of empathy would be needed to pass off as human. These kind of things are not explicit in the movie, which I think is mostly fine as the movie is more about the big picture of what makes us 'human', or what that even means ethically or experientially. The early replicants are basically psychotic so it's easy to have no ethical concerns about them even if they have agency and consciousness. But then what about succeeding versions that do come closer and closer to experiencing regular human existence?
These guys were literally just filming themselves playing the piano, when they were approached by people completely outside the view of the camera telling them to stop.
Besides wanting to keep all the revenue, the NCAA talks up all that sanctimonious crap to justify their non-profit status so they don't have to pay taxes on it either.
I have been to two vegan sushi restaurants (I live in Los Angeles), one was okay and the other is surprisingly good. I remember the second place having a surprisingly good salmon substitute, but it obviously wasn't from this new startup since they apparently haven't launched anything yet. I don't need these meat substitutes to taste exactly like the real thing, I just want them to taste good and provide a similar taste experience. The author mentions the 'uncanny valley' of taste- that doesn't bother me, as long as it generally tastes good in a similar setting. I think that makes for an easy substitute so restaurants can provide more non-meat menu items.
I have only been vegetarian for a short while, and was previously a long time meat lover. I don't need meat substitutes in my life (oyster and king mushrooms are a big love of mine now) and they will never be a basic staple, but it is nice to have the option. If you're at or having a BBQ, it's easy to also grill an Impossible burger for example. I used to love sushi and will occasionally surrender to a spicy scallop roll or something like that, so appreciate having vegan options.
My main reason for becoming a vegetarian is avoiding industrial farming as much as possible, with probably sustainability being the second factor. I don't have an ethical objection to humanely raised (or wild) and humanely slaughtered meat, there are some good sources for that although I don't generally find myself seeking them out.
I am 100% for decriminalization, but understand it's not a black and white issue. I don't disagree with your point, at the same time there's a part of me that thinks, "You can't outlaw stupid." I feel for the guy in the article, but whether psychedelics are legal or not this guy was just not getting the message (I don't mean metahphysically, I just mean realizing that yes psychedelics are a powerful thing):
> "So my first psychedelic experience was when I was 20 years old in in Amsterdam. I took mushrooms and I had a hell of a trip like because I wasn't aware of dosages and don't know what type of dosage I took."
And then a few years later we get to this-
> "I was planning to do an ayahuasca retreat in Peru, and a friend of mine said why not smoke DMT here. I didn't know that I was doing one of the strongest psychedelics on earth. I didn't have any preparation."
Like I said I really do feel for the guy and hopes he finds resolution, but what was he thinking? I also hope lots of people read his story.
I also think that 'net-zero trauma' poster/slide is pretty cringey. Like you said, it's not that simple.
There's also the situation in American football where the team on defense purposely lets another team score a touchdown rather than stop them for a field goal, in order to have some time left to try to regain the lead. Also vice versa, where the player on offense will purposely fall before going into the end zone.
This one though is at least somewhat of a double-edged sword, as it's not a 100% given that the team's kicker will make a field goal (probably around 90-95% success rate.)
Yeah, I can't think of a plausible rationale for this rule. FTA:
--------------------------------------------
No match could end in a draw; if the teams were tied at the end of regular time, they would go into sudden death extra time. But! Any goal scored in extra time would count as two goals. This was presumably done because this tournament, like many, used goal difference to break ties in the qualifying groups. (Goal difference = total number of goals they’ve scored minus the number of goals they’ve conceded.) So that extra time “golden goal” would give a team an edge in the overall competition. Little did the organisers know that it would also lead to one of the strangest football games ever seen.
--------------------------------------------
Such a rule has no impact within a game, it doesn't change the basic premise that a tie game goes to sudden death and next goal wins. But potentially weird scenarios are actually pretty easy to think of if you just consider the rule for a couple of minutes.
> BlackBerry has everything Apple's Tetris film lacked: human drama grounded in actual history, without the need to spice things up with car chases and fantastical storytelling. On the face of it, the rise and fall of Research in Motion's keyboard-equipped smartphone may not seem inherently compelling. But the brilliance of the film — directed by Matt Johnson, who also co-wrote it with Matthew Miller — is that it makes the BlackBerry's journey feel like a genuine tragedy.
Everything you guys say rings true to me. It ended up not adding a great amount of value for either businesses or consumers, and without true added value there's no revenue bite for Groupon.
The stigma of using a coupon didn't bother me so much, most restaurant servers were always cool about redeeming them (and of course I always tipped on the pre-Groupon bill.) But I would sometimes get an (unintentional) negative vibe from higher ups, like great another money-losing customer that I'll never see again. That reaction is what discouraged me from using Groupon, if it just ends up generally being a bad proposition for businesses I'd rather just avoid the whole thing.
Oh man, this reminds me I still have a much of credit at Restaurant.com, which I stopped using a few years ago too...
I agree with your point, just wanted to add that 'implicate' is not even the right term to describe the paper in question. Metaphors and analogies don't prove or even implicate anything, but can provide new perspectives for considering a subject. Looking at the original paper, that seems to be why steam locomotives are discussed.
Given that neither company releases parameter counts, that sort of information would be slow coming out anyway. The most important thing is improvements in actual performance/ benchmark numbers, which allow them to preserve their price points as much as possible.