hunter2_

> you'd still have to pay your electricity bill.

I should only have to pay 1/3 of what it adds to my bill, given that it's 1/3 as efficient as a heat pump. And that coefficient is to be adjusted as outside temp changes (assuming air source heat pump; water source would have a stable COP).

Easier said than done, but if you water-cool the GPU just upstream of your domestic water heater, it wouldn't be a bad thing. Perhaps using coolant and a counterflow heat exchanger, rather than the potable water, but the point stands. Would just need a secondary tank as a buffer (able to soak up heat from the GPU at all times, even when there's no demand for hot water) which then flushes in when demand exists.

"repeated use" seems to be doing the mitigation work here, though it does seem unusual that someone offended by the repetition would be unoffended by a one-off.

What is the "good" when considering the lack of aggressive truncation/overflow concerns here?

While that might be the case, best to normalize that: Title Normalizers are Crap.

So long as the "getting" doesn't consist of circumventing something that is ostensibly a security mechanism, such as a URL that had never been presented to you by the server but is fairly easily guessed.

In a place where recording audio requires consent from all parties, I don't think it matters whether "all the processing is on-device" given that the law was probably written with tape recorders in mind. If I buy a device that can record audio, and place it in my home where unsuspecting guests are recorded by it, whether that recording is transmitted to the manufacturer or stays within my home seems irrelevant.

On the one hand, I wouldn't expect too many variations here (the vast majority of devices probably use 48 kHz and 24-bit output, a few use 44.1 kHz and/or 16-bit, etc.) but just like DPR and all the other properties with a very small set of popular values in practice, you only need a bit or two from each measurement to eventually have a high quality fingerprint.

This is a huge stretch, but if this problem exists in not only the PC versions of Chrome/Firefox but also the Android/iOS versions, then theoretically the app store reviewers could flag the browsers for facilitating this behavior against app store guidelines. In practice, apps of such caliber as popular browsers might be a bit above such reviewers' pay grade, so to speak.

Does anyone else get tripped up by the word "avoid" being positioned to apply to each item in the series, but then logically it must only apply to some (the first and second out of 3) items?

{ avoid meanings, avoid affiliations, avoid fitting }

My solution is generally to repeat the word for each item that needs it, but I can see how the repetitiveness isn't great either.

No, I've never seen Google Photos reflected in any Google Drive interface (web, OS integration, etc.) which on the one hand makes sense (think of Google Photos more like a database-driven black box app than a hierarchical filesystem) but on the other hand Google Drive does show gdocs, gsheets, etc. (which is also sort of a black box app, as the "files" seen in GDrive are merely lightweight stubs).

Only in storage quota views do Photos and Drive (and Gmail, etc.) commingle.

> the web applications like Google Photos are annoying but they take away all of this

They minimize it but don't take it away. A content creator must personally place their creation into multiple storage solutions, regardless of whether those solutions are cloud apps or local disks or anything else. In a world where disk failure is the only failure mode, sure, you can outsource the work of disk redundancy to a company, but it's not. The person you outsource to could decide to deny you access! Losing your cloud account (credential theft, alleged ToS violation, bankruptcy, etc.) is also a very real failure mode, albeit far less frequent than disk issues I suppose.

The creator must be the fork in the road.

Your 4th hypothesis is equivalent, not "other," to mine. But I agree that it could be anything you mentioned, as all of them have crossed my mind as possible reasons why I experience "verifying you are human" and "find the motorcycles / busses / traffic lights" more often on my residential ISP than on my cellular or work ISPs, and the unknown botnet/exit node seems most probable (by far) when considering my own traffic. I never have the 403 problem discussed here, luckily.

Perhaps the bad behavior is failure to remove a compromised (botnet) device on your network that has been attacking one or more sites that use a popular security solution.

"Broken link" is not an accurate way to describe this situation.

This reminds me of the pregnant man emoji [0] but it turns out they decided to handle that in an unusual way [1] so not really.

[0] https://emojipedia.org/pregnant-man

[1] https://www.reddit.com/r/technology/comments/u7x3l9/comment/...

When the Pandora music service launched, I remember reading about how it worked (predicting what you might like next) and it explained how they don't need to know anything about the songs at all (such as genre, tempo, instrumentation, place of origin, year of release, language, etc.) to predict what else you'll like. They simply identify other users with a similar history to you, and deliver to you some of the stuff that those other users had listened to that you hadn't listened to yet. If you abandon a song (suggesting that you don't like it much) then you become a bit less associated with the users who play that song a lot.

At least nobody has figured out my password yet.

I just had this happen, and this HN post was the first result when searching "what is discover.aigc" on google. The specific Discover Feed entry that led me to see this domain (within an ERR_NAME_NOT_RESOLVED error) would've been (if the link actually worked) an Allrecipes article entitled "Over 12,000 pounds of bacon recalled" which, in my case is not not some shock-value nonsense (like I assume is the case for OP's experience) because there is, indeed, this Allrecipes article [0] along with this .gov article [1].

[0] https://www.allrecipes.com/maple-leaf-foods-bacon-recall-jul...

[1] https://www.fsis.usda.gov/recalls-alerts/maple-leaf-foods-in...

> just read an extra second [...] accidentally trigger watchdogs

Well, this seems to touch heavily on the concern: if your job normally takes 0.1s and costs $100, it could accidentally be considered to have taken 1.1s and therefore cost $1100. This is quite contrived, but if you consider that some people put very infrequent cron jobs at the top of the hour (which isn't typically the best idea, but I digress) and it might start just ahead of what the billing system considers to be the top of the hour, theoretically there's something to this concern.

Yes, but a mitigation suggestion like "keep using it, except don't do X specific sequence" (for example, send to a Yahoo address via the Reply button, or whatever the case may be) could be helpful as well, since it seems that bad actors (and/or good actors spilling the beans) will figure it out sooner than later anyway.

As someone who doesn't rely on this feature, I'd love to know now as well, but perhaps the etiquette in public would be to align ourselves with:

> we will not discuss or disclose the details of the exploits until they're fixed.

But if there's a public forum where the cat's already out of the bag, then game on. Perhaps this:

https://www.reddit.com/r/apple/comments/1ukilw1/apple_hide_m...

...which makes it seem like perhaps the attack surface is limited to scenarios involving a Yahoo/Sonic address (assuming that Apple only sends X-Sonic-* headers when talking to those providers that want to see it), which might be a small percentage of users.

I fully understand you, but then, copying copyrighted data can't be piracy (because literal piracy is stealing).

When you're already going 10 mph and you're about to add another 10 mph, you can only "call that zero" (i.e., go from 0 mph to 10 mph again) if your point of reference (i.e., the ground) also begins moving with you at that point. Since the ground is stationary, you're definitely about to increase from 10 mph to 20 mph relative to the ground, not from 0 mph to 10 mph, and that's harder to do. But if you're on a treadmill that was stationary for the first change, and then suddenly starts moving at 10 mph right before the second change without affecting your speed relative to the ground, then you can "call that zero" and you'll be able to add another 10 mph (ending up at 10 mph relative to the treadmill and 20 mph relative to the ground) with the same ease as the first go.

That's a good question, and I suppose the mgh formula isn't a suitable answer, so my answer would be something like: if you lift an object to some height, and then you repeat that action (lifting it from there to twice the height), you've done twice the work, and doing twice the work requires twice the caloric intake.

Brilliant. For those wanting more numbers [0], the ball on the 10ft ladder hits the ground at (I'll stick with imperial units) 17.296 MPH, the ball on the 20ft ladder hits the ground at 24.46 MPH or 41.42% faster, and the ball on the 40ft ladder hits the ground at 34.59 MPH or 100% faster.

[0] https://www.omnicalculator.com/physics/free-fall

I'll admit, many maneuvers can be described as cutting someone off. One so crazy that it would likely cause a collision should never be performed, but I was talking about the low-speed "bumper to bumper traffic" scenario common in lanes dedicated to making a left turn: the reason the gaps are too small for a non-cutoff lane change is a direct consequence of the low speeds involved, say 0-10mph. Imagine cars are nearly stopped for a red light, with gaps varying between 0.2 car length and 0.5 car length (this is much tighter than other parts of the US where gaps while stopped for a red light are typically 1-3 car lengths). I would line the nose of my car to the very front of the 0.5 gap (taking the majority of the risk), signal, cross the lane line within a flash or two, forcing the gap to expand to just over the length of my car, all at just a few mph. This is the type of cutting off I was referring to. And the driver who was cut off has no opportunity to fight; they would if I signaled too early!

Indeed, the extremely minor inconvenience of creating a gap in that scenario is very common even in the northeast US. It's the other half of the spectrum, where it's a more severe inconvenience, that it's uncommon for a northeastern driver to accommodate, from which the perception arises that we fight people who are signaling. The signaler is the one picking the fight, and we simply don't entertain it.

> I wouldn't turn on my signal until after they had passed. If I did it before they pass, they might think my intent is to merge in front of them, and slam on their brakes

This is precisely the point I'm making, which I perceive as more typical in the northeastern US than other parts of the US where people use the turn signal to beg for an opportunity to be created that doesn't yet exist (at which point declining to create is seen as actively fighting when it's really not).

> Also really strange to me to be prioritizing fuel economy over safety.

If I see that someone is actually changing lanes (in terms of lateral position changing within their lane) then of course I would get out of their way, as it would be unsafe not to. But if they're postponing their lane change (despite their signal already being on) because I'm about to be in their way, that's the scenario I'm talking about when I say I'm going to continue on my way rather than lose momentum/fuel: they are waiting for me to not be in their way, they will continue waiting until I'm gone, and they should've waited until I'm gone before signaling (by "should" I simply mean if they want to assimilate into this driving style). The example scenario in my other recent comment will help illustrate this.

> leave a big enough buffer with the car in front of you [...] leaving a healthy gap

I realize that my original comment wasn't clear on this, but the "gap" I'm talking about is between cars in different lanes (again, see the example scenario in my other comment). We can assume no lead car at all.

> I also can't comprehend how if you're "scared" by someone signaling when you are in their blind spot, the best course of action could be to put yourself directly in their path vs giving a little extra space to safely merge.

I won't say it's the best, it's just what I've noticed. But it's not "their path" if they're just hoping to be let in without actually moving. When they start moving, it's a whole different story, and I would get out of there (ideally lane change, but brake if needed). I guess "stealing my attention" would've been a better way to say it than "scare."

> How do you handle an upcoming left turn (assuming right hand driving) during heavy traffic?

If I am in the middle lane (lane 2), and I realize I need to get into the leftmost lane (lane 1) to make a turn, but lane 1 is too full for me to simply move into it without affecting others, then I would have no choice but to cut someone off. I would minimize the effect in two ways: by trying to cut off whoever has left the largest gap in front of them (hunting for a gap that might not be the largest now, but will be the largest when I actually use it), and by assuming as much of the rear-end-collision risk as possible until the lane change is complete. Only once my position is optimized to begin the lane change would I signal, because signaling from a suboptimal position could scare people (or give them an opportunity to fight my ability to change lanes). If I can remain in the optimal position for a couple of blinks without any downside, I absolutely will, but in the very heavy traffic we're discussing, typically the tires hit the lane line between first and second blink -- very much not an "ask."

> Does "momentum towards closing the gap" just mean that you're keeping a higher speed than the car in front of you?

No, I was referring to the gap between the car signaling for a lane change and the car that ends up preventing the lane change, which are in two different lanes. Suppose I'm in lane 2, and a car is in lane 1 a few car-lengths ahead of me. Suppose the car in lane 1 is going slower because they just merged from a left-side entrance ramp. Due to our speed difference, after a moment they're now only 2 car-length ahead of me. Their right turn signal comes on. Now they're 1 car-length ahead of me but they haven't yet changed lanes. Now they're 0 car-lengths ahead of me (i.e., the gap is closed) and cannot change lanes. I did not "let them in" upon seeing their signal, because that would ruin my momentum.