"Pay for their own electricity" is just not an actual thing that seems to be doable at this scale without those externalities. You're talking about sites that are, say, 800MW or 1GW or more... these use more power than tens of thousands of homes, and require entirely new power plants, interconnects, lines run, transformer costs, staff, etc. Typically, power companies amortize those costs and spread them over ratepayers, and when it's just part of the normal induced load of population growth, that's just how things go, and the system works.
In these cases, the rates that these DCs are paying for power are nowhere close to being able to fully absorb or offset the additional CapEx that the power companies are suddenly tasked with, even if they put up, say, the capital for IC, which is usually what's required. So the remaining new shortfalls get spread over the remaining ratepayers, ie, everyone else. If the demand wasn't induced by the build of these massive sites that, strictly speaking, aren't "necessary," then the rates wouldn't have to climb to accommodate them.
Frankly, there should be laws against power companies raising ratepayer rates to accommodate infrastructure investments driven solely by DC/fab load inducement.
Would it be possible to potentially expand the size options? Large just isn't large enough for me on my MBP; I'd like it to be nearly 2x the size, or roughly equivalent to the 150% 4k scaling factor I have on Windows. (I like the icon-only mode!)
On the default XDR M2 Pro MBP display I'm on, I have it set to the default scaling for reasons, but I'd really like to be able to scale the BoringBar to be... quite a lot larger... maybe a scale bar, or maybe add an XL and XXL option?
Wow okay - I thought the large option was already quite large. Will experiment with an extra large option in some time and push out an update if it appears stable.
One other feature request: allow users to click and drag pinned (or any) icons on the taskbar. I very much arrange my pinned apps spatially so I know where to look/find/launch them. I know that I can effectively be deliberate in my pinning to try to get them in the right spot, but that's obviously quite limiting.
Just bought a license, though, and really enjoy this! Well done!
It's also routinely failing the car wash question across all models now, which wasn't the case a month ago. :-/
Seeing some things about how the effort selector isn't working as intended necessarily and the model is regressing in other ways: over-emphasizing how "difficult" a problem is to solve and choosing to avoid it because of the "time" it would take, but quoted in human effort, or suggesting the "easier" path forward even if it's a hack or kludge-filled solution.
it does feel something in the hidden system prompt makes it try less hard, so many times in the past several weeks I have found divergences with what was in plan and looking back at the jsonl it's always some variant of "doing it this way would be too complicated, let me take this hardcoded way out". If asked to review the change, it will find it, and it will say also yeah I agree prompt said not to do this, but I did anyways, not sure why.
As others have said, anthropic is between a rock and a hard place, you can't scale compute as quickly, and the influx of new accounts has definitely made things tough for them: I think all the "how is claude this session 1/2/3/4" questions that keep coming up must be part of some a/b on just how far to quantize / lower thinking while still maintaining user satisfaction.
> over-emphasizing how "difficult" a problem is to solve and choosing to avoid it because of the "time" it would take
I heard a while back Claude refused to attempt a task for days, saying it would take weeks of work. Eventually the user convinced it to try, and it one-shotted it in 30 seconds.
Awesome, I didn't know about the car wash question.
Totally true, also tokens seem to burn through much faster. More parallelism could explain some of it but where I could work on 3-5 projects at once on the max plan a month ago, I can't even get one to completion now on the same Opus model before the 5h session locks me up..
They're just words. It's not a person. It doesn't "understand" anything. (I sound like the bad guy in a robots-have-feelings movie)
I've also tried giving LLMs religion to much more limited success (haven't figured out the right way yet).
I'm manipulating a language model, not a person. "fuck you" translates into a vector in a really big space, and it has different results than being polite about it.
In that prompt I'm reenforcing a directive in five different ways
- idgaf about risk
- you coward
- waste some time
- just do it
- stop bitching
This cluster of instructions are all related but in slightly different directions, are unambiguously strong, attention grabbing, and direct and the model does not argue or get confused about intent
In this particular instance this was the fifth time I had given a particular instruction only to have it subverted by the model that had decided "that's too hard I'm going to do something else instead" in four separate ways.
Abusive cursing did indeed work better than any other form of urgency or insistence.
This is a bit like saying a hangnail and a gangrenous amputation are "not equally bad but they are both very bad". One is literally chopping things off to permanently alter them. The other is, at times, uncomfortable and frustrating.
The false equivalence of doing the "both bad!" song and dance serves to so radically under-emphasize the absolute wanton, orders-of-magnitude-worse levels of corruption and evisceration of norms of one side by reducing it to "more bad than the other but they're both very bad." It allows the window to shift to normalize the sort of destruction of systems we're seeing by hand waving away how "the other guys aren't great, either!" It's borderline discourse malpractice at this point, and should be called out as such.
What does it say about us, as a society, or just as _humans_, where the scale and magnitude of this problem is so great and only growing? Where and how are we failing ourselves that the sort of mental illness that percolates and drives this sort of behavior festers, amplifies, and converts into actual, illicit action?
These numbers are mind-boggling, and while I understand that a "few (extremely) bad apples" are probably responsible for an outsized amount of production, AND that AI-generated imagery is flooding the zone disproportionate to the amount of actual human children being physically harmed, it's still absolutely wild to me that we collectively are producing and consuming so much of this content, despite it being largely universally considered essentially the most abhorrent thing possible.
What would fixing this at the root cause even start to begin? How do we apply whatever combination of therapeutic intervention or further societal pressure or whatever might work to reduce the incidence of people having these urges, exploring them, feeding them, and sometimes acting on them? We see signs in every airport bathroom telling us to look for signs of trafficking. Trafficking intervention training is a huge deal in the travel industry in general. There are early intervention and detection systems for social workers and case workers.
But has anyone spent any real time looking at this from the other side: the side of the offender? I imagine there's research on the typical chain of how someone gets "onboarded" here: it probably starts with some early abuse, or if not that, early exposure or early curiosity, and then snowballs from there. I'm just thinking out loud about how large the magnitude of the problem is on the offender side if we're talking about this volume of images, and how we might be able to evaluate things from the "ounce of prevention worth a pound of cure" side of things, because damn is this depressing.
Images are interesting though. You can have a massive amount of images for only a few consumers.
I would be interested in statistics related to the percent of adults who would be considered child predators. I have zero scope on how large this issue is by percent of population.
If we're talking about 3% of everyone who is sexually attracted to children, that's one thing, but if it's .0000001% then the issue really is just the producers of content.
Does anyone here know of any studies or statistics? My basic googling hasn't really turned up anything trustworthy.
That's what I'm getting at with the "few bad apples" reference: it's _possible_ (and I'd hope) that the percentages are very small... but the insane volume of things like _grooming_ and other behaviors, to say nothing of just how many women report some form of sexual assault or abuse by the time they reach adulthood being in, what, the high 30%s?... it's not great.
As per Wikipedia there is really bad/no data on this because almost all research relies on convicted pedophiles and going around making “are you a pedo, perchance?” surveys in the general population simply does not work.
Germany has an anonymous support programme for people who feel paedophilic urges but don't wish to offend. I believe they've used that network for research, but I think it's probably quite a limited, and potentially biased, sample.
It's also worth considering just parent taking photos of the child would hit the positive on classifier. And it can be CSAM and not CSAM at the same time, because it is fine to be on the device of the parent, but it can also be stolen and distributed by maliciosu actor.
> What does it say about us, as a society, or just as _humans_, where the scale and magnitude of this problem is so great and only growing?
That the people in power have too much power and they get away with it often enough that there is actual money to be made supplying them.
I remember when the official terminology changed from "child porn" to "child sexual abuse material", and how this was meant to emphasize that it was produced by actually abusing an actual child.
That's Phoenix, it's here. Waymos commit to nominally keep the speed at the speed limit but it is _extremely noticeable_ that that's the case because literally NO ONE drives 65 on the freeways here. Everyone is at minimum at 74. It's a rite of passage in Arizona. It's not even a speeding ticket until 75. Goes back to the 70s with the feds trying to force speed limit laws or threatening to revoke highway funding. Arizona said "fine, but it's not a speeding ticket. it's 'misuse of a finite resource.'"
So you'll see the Waymos kind of puttering along at 65 as everyone zooms around them. They DO say they'll occasionally exceed speeds when it's safer to do so, but it's obvious they don't want a narrative of them being speed demons and flying around exceeding the speed limit.
FullStory just tried to pull this with their renewal. We had a mult-year contract that started with a two-page order form, on which the words "renewal" or "cancellation" never once appear. During negotiations, it was never discussed that the plan would renew, or that there was a cancellation window. Instead, buried at the very bottom of the form (which they send via CongaSign, and wasn't clickable or obvious), was a line about their subscription agreement being linked to their terms and conditions page. On THAT page, they mention the plan will auto renew and must be cancelled with 60 days notice.
We cancelled at T-45 or so days before renewal, having determined it wasn't a fit for our client anymore, and they insisted "well, actually, you've renewed anyway!" which, no, we haven't. Absolutely absurd to try to "clickwrap" buried renewal terms in a 20+ page T&C/privacy document rather than as a material point of fact on the actual order form being executed.
Feels like the height of absurdity to try to bully your client into forcing them to use your services against their will when they still gave ample notice that they were cancelling and when there was no material loss to the business, but it's always felt like their revenue team has been unhinged in general: exploding offers, insane terms, super high-pressure sales... part of the reason we left them in the first place.
100F days are fine, cakewalks, even, especially with misters + shade. We had 70+ days of 110°F two years ago, and over 20 days 115°F+. They are not the same. Those days are unbearable nightmare fuel, and worse, they turn into insanely miserable nights where the low temperature rarely dips below 95°. It is absolutely awful, dry or not.
In these cases, the rates that these DCs are paying for power are nowhere close to being able to fully absorb or offset the additional CapEx that the power companies are suddenly tasked with, even if they put up, say, the capital for IC, which is usually what's required. So the remaining new shortfalls get spread over the remaining ratepayers, ie, everyone else. If the demand wasn't induced by the build of these massive sites that, strictly speaking, aren't "necessary," then the rates wouldn't have to climb to accommodate them.
Frankly, there should be laws against power companies raising ratepayer rates to accommodate infrastructure investments driven solely by DC/fab load inducement.
reply