There was a similar situation last summer over a prediction market on whether Iran's Fordow nuclear facility would be destroyed by a certain date. That one was resolved as "yes it got destroyed" after the air strike on the facility. A lot of people on the other side of that bet were complaining because it seemed like an arbitrary guess: All we can really tell from publicly available info is that it was hit. The actual effect may have been anywhere from superficial light damage to comprehensive destruction, with no way to be sure without access to the underground facility.
I didn't bet on that one, but I'd seen something about it on Twitter & gotten curious how they could come to a firm conclusion one way or the other. AFAICT the market didn't have a solid way to be sure & were just taking a White House press briefing that said it was probably destroyed at face value.
I'll never buy a car manufactured after about 2014 for this reason. I'm planning to just keep getting repairs & upgrades done on my model year 2006 for at least the next 10-20 years. By then perhaps I will want to switch to electric, but I'll do it by electrifying something older.
Cars from around 1998-2014 usually have side curtain airbags & adequate rollover durability. The only improvements since then that I'd even want at all are better EV batteries & marginal efficiency gains for IC engines, but those can be retrofitted &/or aren't worth the anti features they also added IMO.
If car companies want my business they'll have to remove the telemetry & automatic updates.
I don't care if I end up paying more to drive an old car eventually, but this approach has also been saving me money so far.
No thank you. I will take predictable handling and a steering wheel that responds to my inputs. Loss of traction situations are exactly where I don’t want any systems helping. I need to countersteer and feel the car. Speaking as someone who was raised in winter driving and encouraged to find the limits of handling in snow and ice covered parking lots.
Of course if you are one of those drivers who removes their hands from the wheel in a stressful situation (there are many), these systems will help somewhat.
It really depends on the situation and the car. I’ve had it really help and not take over too much (very modern Porsche in the mountains), and systems where it was actively making the situation much worse by alternately locking the brakes on individual wheels. That was down a long hill which turned icy a third of the way down in a borrowed 2013 BMW F30, and I still consider it luck that I kept it on the road and nothing was coming the other way.
I have a car from 2017 that is perfectly dumb. It had been a rehash of a car being produced since 2010 though.
All other models of the same year by the manufacturer had telemetry, mobile app start etc. All those models are now dumb though since for those earlier years they used 3G wireless which is now a dead spectrum.
> An AI may not produce information that harms a human being, nor through its outputs enable, facilitate, or encourage harm to come to a human being.
This part is completely intractable. I don't believe universally harmful or helpful information can even exist. It's always going to depend on the recipient's intentions & subsequent choices, which cannot be known in full & in advance, even in principle.
They wanted to keep accurate global positioning as a US military exclusive capability. It's definitely useful for guided munitions, & alternative satellite positioning systems didn't exist or were less mature at the time, so US GPS was the only system one could realistically use for that. A missile able to hit a target within a 3 meter radius is vastly more effective than one that can only hit within 100m, for instance.
There are still some restrictions around this sort of thing: IIRC a GPS receiver for sale to the public isn't allowed to give accurate data if it's too high up &/or moving too fast, to prevent unauthorized usage in ICBMs & other similar weapons. I think there would be a lot of red tape involved if you wanted to buy an unrestricted GPS device without this limitation.
0°F to 100°F spans the full range of temperatures I'd go out in & not consider it "extreme weather", so it's rather intuitive in that you can think of it as "how hot is it on a scale of 0-100". It feels very human centric & convenient for everyday usage IMO.
I see two main types of 'AI safety': (a) Safety for the business providing the model. This includes a censorship layer, system promoting, & other means of preventing the AI from giving offensive/controversial/illegal output. A lot of effort goes into this & it's somewhat effective, although it's often useless or unhelpful to end users & doesn't address big-picture concerns. (b) The science fiction idea of a means to control a hypothetical AI with unbounded powers, to make sure it only uses those powers "for good". This type of safety is still speculative fiction & often assumes the AI will have agency & motivations, as well as abilities, that we see no evidence of at present. This would address big-picture concerns, but it's not a real thing, at least not yet.
It remains to be seen whether (b) will be needed, or for that matter, possible.
There are a lot of other ethical questions around AI too, although they mostly aren't unique to it. E.g. AI is increasingly relevant in ethical discussions around misinformation, outsourcing of work, social/cultural biases, human rights, privacy, legal responsibility, intellectual property, etc., but these topics predate LLMs by many years.
It sounds like a fun & challenging record. Falling off a scooter at 110 mph would certainly shake you up, but it'd probably be a little safer than moto gp if you used similar safety gear. I wonder how it feels at high speed though - I've never ridden one anywhere near that fast, but the scooters I've ridden didn't feel like they'd be very stable at high speeds.
When I'm snowboarding on an overcast day, it can sometimes be hard to see the exact shape & conditions of the snow ahead so I have to slow down to make sure I don't catch an edge on a 'hidden' mogul. I'd like an AR system that used LIDAR/FLIR/etc. to augment my vision to see these features better.
I'm also bad at learning & remembering a lot of people's names at once in social settings, so I'd like a discrete pair of AR glasses that used a local model to add virtual nametags to people in certain situations. (Assuming I controlled the data - I wouldn't like it if this meant data about my acquaintances would be sold behind my back).
So there's at least two potential AR applications I'd be interested in, assuming they could be made to work in a trustworthy & reliable manner for under $1k.
In terms of microplastics, I would think 100 of the old flimsy single use bags would be much worse than 5 reusable plastic bags, even if the total mass is the same. The heavier reusables have less surface area per mass, so they'll be degraded more slowly by the sun. They also are less easily blown by the wind, so it's more likely someone will dispose of them properly or that they'll naturally end up buried somewhere that does a better job of containing the eventual microplastics. Fewer bags in total would probably be better for sea turtles than thinner bags as well.
I'm not sure if that makes the reusables better overall, but I don't think we can say they're 10-100x worse based on weight alone.
I agree with most of your comment, except that microplastics come from paint, tires, and washing synthetic garments, not plastic bags, and I'm dubious about your photodegradation point.
Are you sure about that? AFAIK effective laser drone defenses are not yet widely deployed proven technology, but I don't think small beam size is a limiting factor. Getting enough power onto the target to disable it is a big challenge, but part of that is fighting the natural tendency of the beam to spread out & be attenuated by the atmosphere - not that the beam affects too small of a spot on the drone.
Having a laser that spreads out to e.g. 30cm radius at 500m is not hard to do if you need an area of effect weapon & can push enough power (ie. your laser is powerful enough, but not so intense that it ionizes the air & blocks itself). Reflections seem like a bigger problem: If the most effective defense includes guys with shotguns &/or there are a lot of unprotected personnel in the area, how do you make sure stray reflections don't end up blinding them?
The point is that a focused laser will put a hole through the drone, much like an armor piercing round, but that is often insufficient to disable the drone. A larger ballistic projectile (think a solid shell or a rock) is much more effective. Alternative energy weapons based on microwaves and SPL also work well.
Already tested. Success rate is too low. A great deal of aenergy gets wasted.
Remember, this is about asymmetric warfare. If the number of rounds or amount of energy required costs more than the drone it shoots down, then it's not an effective deterrent. Militaries are looking for single-shot weapons to take down drones. Fire once and move on. It's the only way to deal with a swarm. Think about it for a bit and it will become very obvious.
I didn't bet on that one, but I'd seen something about it on Twitter & gotten curious how they could come to a firm conclusion one way or the other. AFAICT the market didn't have a solid way to be sure & were just taking a White House press briefing that said it was probably destroyed at face value.
reply