Hacker Newsnew | past | comments | ask | show | jobs | submit | widdakay's commentslogin

These are two different products. It's like SpaceX launching satellites for competitive satellite internet services. They didn't care that they were providing launch capabilities for a competitor and neither should Anthropic. What if Apple stopped allowing you to use an iPhone if you worked at Google?


"They didn't care that they were providing launch capabilities for a competitor"

Yeah because another internet provider did not have SpaceX reusable rocket technology

its not really quite the same you know


> What if Apple stopped allowing you to use an iPhone if you worked at Google?

This wouldn't be remotely comparable. This is targeting of a competitor's employees, not targeting a competitor's subsidiaries.

If you want to go the Apple-Google route a better comparison would be that this is like Apple refusing to allow you to hook up an Air Tag on an Android phone. Which is something that they do, in fact, do.


Another way this isn't comparable is that the training data is critical for the services each provides. Seeing answers Claude gives (at scale, for free) would be a huge competitive advantage to OpenAI, whereas Apple Mail seeing the email you send via Gmail wouldn't convey any competitive advantage in email, for example.


I think the nature of your two examples, along with the one from the news, is too different from each other for the analogy to really hold. These situations can only be judged on a case-by-case basis.


If SpaceX was launching every rocket at a loss then they would absolutely care about a competitor taking advantage of it.


Do you understand that data can be used for training?


Is this what is used in the new Tesla Model Y?


Repost of https://news.ycombinator.com/item?id=42862262 during actual hours when people are awake.


Can someone find more about why GEICO did this? It makes no sense. Furthermore, they could just increase rates if they found accidents too common.

Also, there are much less safe cars out there for pedestrians such as the Hummer EV (complete behemoth) or Rivian (weighs more than CT and has significantly higher frontal profile which is shown to be the largest contributor to pedestrian safety above basically anything else). If it's about breakdowns, that money comes from Tesla's wallet so would make no sense. Even so, I see Cybertrucks driving daily and haven't seen a broken down one yet.


My understanding is that any damage to the cybertruck unibody frame, no matter how minor, becomes a catastrophicly expensive repair. There are many laws surrounding fair pricing of insurance rates, and Geico might not want to go through the legal work of justifying massive rate hikes specific to the Cybertruck over just dropping such a tiny volume of vehicles, especially if their "profitable" pricing just drives all their customers away anyways. Also keep in mind that in the meantime they're locked into their current rates until each state approves a higher rate.


Rivian is likely heading this way too, there's plenty of stories about a fender bending costing 42000 to fix

https://www.theautopian.com/heres-why-that-rivian-r1t-repair...


It looks like the Cybertruck is more like a Delorean than I thought!


Insurers work with data/metrics and generally make decisions based on that alone[1]. I often feel that it's unfair that I pay more as a male, because I am an extremely responsible and defensive driver - fact is, men cause more expensive accidents more often. That's the data, that's what insurers care about.

At some point Geico likely did insure super cars, up until they started becoming highly anomalous in the data. The same has happened to cybertrucks, whether your highly restricted sample demonstrates it or not.

Insurance companies don't turn away profitable customers. Cybertrucks became a problem for Geico, but they are being tight lipped as to why; it might not be a reliability issue.

[1]: https://en.wikipedia.org/wiki/Quantitative_analysis_(finance...


> I often feel that it's unfair that I pay more as a male, because I am an extremely responsible and defensive driver - fact is, men cause more expensive accidents more often. That's the data, that's what insurers care about.

Same - I remember a few friends complaining about this but an older friend basically explaining how insurance worked from the perspective of the company. Almost everyone will swear that they’re a great driver, and they can’t tell the difference until you’ve been driving for years. The only alternative would be the kind of monitoring + speed limiters that most drivers get extremely upset about so it’s unlikely to change before we get L5 self-driving.


> they could just raise rates

Well, they (insurance companies in general ) also dropped a lot of homeowners in certain states, rather than simply raising the rates.

Perhaps they have done some market research and determined that there is an inflection point beyond which raising rates would actually reduce profits due to reduced competitiveness


Car insurance in general is a race to the bottom with competition. A good quarter has 70% of incoming premium going out to settle claims.

When you have fender bender claims costing 20-40k USD to repair, how do you price that risk?


> When you have fender bender claims costing 20-40k USD to repair, how do you price that risk?

This is a solved problem. Ask any actuary who specializes in casualty insurance, or read a standard textbook about non-life insurance mathematics.


Shush, I want to see a YC25 auto insurance startup that uses a pile of overheated video cards as a crystal ball to make underwriting decisions.


most people buying specialty casualty insurance also have MUCH deeper pockets than even most people buying a cyber truck


Not really, car insurance rates need to go through alot of state level validation in order to be approved.

Other models are not considered insurable by GEICO as well. So likely small pool of policyholders + exorbiant claim payments == not worth the headache


The reasons listed in the article are spot on. Ridiculously high repair costs is enough. With ridiculously high repair costs come extended repair times (rental car and claims management). There comes a point where just increasing the premium price can't offset risk.

There is nothing practical about Cybertruck ("Cybertruck," really?). It's a collector's vehicle.


> Can someone find more about why GEICO did this? It makes no sense. Furthermore, they could just increase rates if they found accidents too common.

Very few insurers will insure literally anything (back in the day, Lloyds of London were notably unusual in that they would write a policy on basically anything, though you mightn't like the cost). Most conventional insurers will have a line after which they say "this is too risky, we'll leave it to specialist insurers". Ask anyone who's ever tried to get insurance on a non-conventional-construction house.

And it's a pretty niche vehicle; if they do find it unacceptably risky, then dropping it is presumably a fairly easy decision.


Good point I had not thought about the remaining parts of the economics if they did go though with something like a rate hike.


Look at the weight of some cars for comparison. Modern EVs aren't too far off as you say when compared with comparable performance cars: Tesla model 3: 3920lbs (SR) 4056lbs (LR) Kia ev6: 4017-4255lbs Average new US vehicle: 4100lbs BMW M3: 3,840-3,990 lbs BMW M5: ~5300lbs Audi A5: 3,682-3,990 lbs Toyota Camry: ~3500lbs Toyota Prius: 3,097-3,340lbs Dodge charger: 5100-5900lbs


Yes, those BMWs are insane weights for an "M" ("Motorsport") series.

Especially considering that it's common knowledge in motorsports (although evidently less common than it should be) that more power only increases acceleration performance, but lighter weight increases performance of acceleration, braking, cornering, and tyre wear! Plus, reducing weight is compounding; take out one chunk of mass, and other components can also be made lighter, e.g., reduce the body mass, and you can fit lighter suspension components, brakes, wheels, tires, etc., and the light-weighting compounds.


I'm surprised this isn't more common. The only difference between a heat pump and standard air conditioner is a reversing valve. These are usually $50-$100, and just require one more wire to the thermostat. In colder weather, defrost and fancier controllers are needed, but for mild climates the reversing valve is really all that is needed.


You are dealing with different pressures on the refrigerant lines, but honestly that shouldn't really matter all that much. You also need a bit of logic for if the condenser starts freezing over to temporarily reverse the flow (and turn off the home fan) to defrost.

But otherwise, yeah, almost identical and a little crazy they'd cost much more over a typical install.


Off the shelf heat pumps will have a defrost control board too, but you make a solid point.

The hard part is that you have to recover the refrigerant and refill, which takes HVAC/R equipment and and EPA certification to do legally.


PGE goes up to 66c/kwh during peak now and in San Diego I think is >70c/kwh. Off peak base is 34c/kwh for PG&E. Gas is $4.5-$5/gal so break even I think ends up being near 50mpg ish for a 300wh/mi EV.


I'm not in CA, so genuinely curious. When you quote 66 cents/kwh is that the marginal rate for an addition kwh or is that averaging in fixed costs? I pay about $20/mo just to have an active service line to my house, even if I were to shut off the breaker and not use any power. But my marginal rate is about 18 cents/kwh.

I would call my rate 18 cents but not sure if we are doing an apples to apples comparison.


Marginal rate. SDGE is insanely expensive.


I find my (NorCal) PG&E bill extremely (and deliberately?) inscrutable, but for the most recent statement, the Off-peak net price was $0.427/kWh and $0.466/kWh on-peak.


Yes there are some. Eg: Acura MDX. It uses the mobileye chipset and can barely make a highway turn. The car is good otherwise, but it in no way compares to the Tesla ADAS (I have daily driven both). The Tesla has done several trips to LA and back with >1hr between touches on California i5. my commute on bay area 280 feels far more natural on Autopilot when cars cut me off etc. I think performance depends significantly on what roads people drive.


Do we need TSA for interstate highways?


You mean like the random checkpoints set up by border patrol in the southwest where they stop all the freeway traffic? Been around for years.


They already tried it once in Tennessee. IIRC, it was only for truckers and after a good bit of blowback, they claimed it had been voluntary (though the truckers didn't know that) and strictly informational.


They're sorely needed at the entrance to my apartment building.


This is an opinion that is unironically held by a lot of people on Nextdoor.


South Park had a pretty appropriate episode about something like this [1].

[1] https://southpark.fandom.com/wiki/Toilet_Safety_Administrati...


Furthermore, if I press the button when near two Teslas, both charge ports open up, so the ability for the person to replay the signal yielded a "duh" to me.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: