I think the problem is ultimately that LED retrofit bulbs are a fundamentally terrible idea.
They offer terrible thermal performance, require the power supply to be in a tiny case together with the hot LEDs, don't make good use of the directional nature of LEDs, and perhaps worst of all they compress most of the technical complexity of lighting into a tiny and super cost sensitive commodity item that is then put into some cheap fixture.
Lighting is a complex thing and LEDs in properly designed lights can be fantastic but we need to get to a point where there are good offerings of purpose made lamps that integrate LEDs in a sensible fashion rather than crappy retrofits that are barely good enough because they use an amazing technology and then squander it's potential entirely.
and no having the LEDs be integrated doesn't necessarily mean they are not replaceable. COB LEDs come in standardized sizes that have solder free sockets so you could totally make a lamp where the light source could be replaced when it dies or, much more likely, be upgraded when an even nicer one comes around.
These are not comparable situations.
Ideally everyone should do a good job, sure, but cops are in an entirely different position in terms of power and responsibility.
as part of the executive arm of government they are tasked with holding the monopoly on violence and egregious mistakes in doing so are a fundamental violation of basic rights and the peaceful order of society.
When they screw up it's not a bad product being delivered, it's trust in society being eroded, lives being destroyed and the door to tyranny being opened just a little further.
severe deterrence against misbehavior, loosing a job honestly seeming VERY weak compared to the scale of responsibility at play here, should simply be par for the course in this context.
True, but how does firing the whole department solve the problems? I don't think just firing the bad apples is a good long term solution. Mainly because these types of mistakes continually happen in multiple departments in multiple counties and countries. Whenever things keep happening over and over again, it is likely the system itself that causes the problems. Perhaps we should look at countries with the lowest police mistakes and emulate them.
I try to solve problems from a systems perspective. For example the common automobile stop and warrantless search allows officers to more easily abuse power and escalate situations quickly as cops no longer required a warrant from a judge. This was by design by the Supreme Court who ruled in 1925 to allow this exception in times of prohibition when an individual ran away from police with alcohol and they didn't have time to go get a warrant. Stop and frisk was an extension of this case law.
As part of the executive arm of government, they are under the responsibility of the government. When they screw up it reflects on the ability of the government to organize, regulate and fund that executive arm, and also the trust that people have put into those politicians.
If people no longer trust that executive arm of government, then hold those in power responsible for it. Start at the top and replace those politicians. An elected politician can only remain in office if they continue to receive support by voters.
Sure politicians should also be held accountable for policy failures but I don't think it absolves the people enacting an injustice from responsibility.
"just following orders" is not a valid excuse and "just vote different" is not a valid way to hold people accountable (in all cases).
When everything works in the spirit of the law and the output is still bad it makes sense to just vote for different laws/people to make new legislation, sure, but when rights are being violated there need to be harsh consequences.
It is important that those responsible for the police force can not just simply wash themselves clean of the problem by firing the lowest level employees. I do not disagree that individuals need to be held responsible for their own actions, through most blames should lie on those who hold power to create real change in the organization. As an generalization, a good model is one where blames travels upward and praise travel downward.
Just voting differently is naturally not the only way a government can be held responsible. Political assigned jobs is a job description based on trust, and there should be mechanism to remove individuals who do not fulfill that trust. Governments should also be liable to their citizens, so when injustice occurs people should have the option to demand compensation from the government. This in turn put pressure on politicians to avoid injustice caused by the police.
A police force that protects citizens rights should be viewed as an asset to local government, and the opposite should be a harsh liability that local governments can feel.
Sure, replace the politicians at the top who aren't willing to cull and replace dysfunctional and corrupt members of the police force. And if it's systemic, the entire police force, which is what's being suggested.
Wine is interesting because, while there is definitely some pretentious nonsense going on, the other extreme is hilarious as well.
There are people out there insisting that it's impossible to differentiate red from white wine by blind taste test.
Interesting, but one has to ask, were they testing the hypothesis or the participants? The participants were describes as both "students" and a "panel of experts".
except the study does not test for that. it somewhat interestingly shows that distinct vocabulary gets used for reds and whites once that context is established by a visual cue.
that doesn't however mean no systematic difference between reds and whites can be detected.
This doesn't seem to work very well at all.
I frequently search for a particular vendor of something by name only to be served another similar vendor as the top result.
It feels like it should be doable to serve these results in some sort of "you may also like these similar vendors" box or something like that which would massively improve the user experience
It doesn't work if the real company isn't an advertiser, or if they are an advertiser but aren't in this auction due to their targeting rules or being out of budget.
Because film has a color grade built in to some extent ewhile digital does not.
The "look" people built into the film back in the day because they couldn't just do it in post can still be a great look today and is pretty orthogonal to objective quality of the information captured.
I don't think taken to this extreme this is true either.
Sure for a given film setup you could come up with a better digital setup but that doesn't mean every digital setup is strictly better now than what you'd use if digital wasn't an option.
and you could always construct some monstrosity of a camera that just throws more film at the problem to get a better result, though IMAX seems to be the limit of what people are willing to do on that front.
I'm pretty sure at this point high end digital cinema cameras can straight up outperform film in all but the most extreme cases (like giant IMAX negatives maybe) but that's also a much more recent development than most people would think.
More freedom in post allows more freedom to do things audience may or may not like but I don't think it's technical limitations holding things back anymore.
Similarly I'm a bit dubious on the LED lighting comment by the anonymous redditor.
Yes LED color rendering used to be pretty horrible but these days it really doesn't have to be. all the major LED manufactures have high end white parts for color critical applications and even color mixing is actually good now with the advent of phosphor converted broad spectrum colored LEDs.
All this stuff increases the available creative space but at the same time it might not be easier to find the good stuff in a larger space
I find all the arguments about why it's bad to anthropomorphize here a bit tiresome.
It seems clear as day anthropomorphization is more appropriate to this subject matter than to almost any other subject matter where we commonly do it without giving it a second thought.
It's a chatbot! the whole point is anthropomorphism!
if it lies, it lies.
We know it when we see it and pretending we don't won't get us any closer to nailing down a better definition than that.
I think it's better to just acknowledge the fluidity of language and our difficulty with nailing down hard definitions in general than to pretend this particular subject matter is somehow unique in this sense
I'm a bit worried about this kinda stuff being passed of as "AI safety".
no, making your LLM actively more deceitful and less aligned with user intent is not the way to make AI safe.
It would be very interesting to know how chatGPTs censorship engine is implemented though. are they retraining the whole thing all the time to fix new jailbreaks?
The paper makes a valid point in general but I feel it makes unjustifiably definitive and general statements and puts up odd goalposts.
The section on emergence makes a very convincing point about how such systems might, at least in theory, be doing absolutely anything, including "real" cognition, internally and then goes right ahead and dismisses this entirely on the basis of the system not having conversational intent.
who cares if it has conversational intent? If it was shown to be doing "the real thing" (how ever you might want to define that) internally that would still be a big deal wether the part you interact with gives you direct access to that or not.
Then it goes on to argue that these systems can't possibly actually believe anything because they can't update believes.
Frankly I'm neither convinced that the general use of the word "believe" matches the narrow definition they seem to be using here nor that even their narrow definition could not in principle still be taking place internally for the reasons laid out in the emergence section.
I agree people should probably be mindful of overly anthropomorphic language but at the same time we really shouldn't be so sure that a thing is definitely not doing certain things that we can't even really define beyond "I know it when I see it" and that it sure looks like it's doing.
beyond that I'm not even really sure there is a good philosophical grounding for insisting that "what's really going on inside" matters, like, at all.
The core thing with the turing test isn't the silly and outdated test protocol but the notion that, if something is indistinguishable by observation from a conscious system, there is simply no meaningful basis to claim it isn't one.
all that said the current state of the art probably doesn't warrant a lot of anthropomorphizing but that might well change in the future without any change to the kinda of systems used that would be relevant to the arguments made in the paper
