djspiewak.fosstodon.org.ap.brid.gy
@djspiewak.fosstodon.org.ap.brid.gy
@dpiponi @skewray …and obviously, in some contexts, that's extremely correct! Even for regular person scale problems and hardware, people generally underestimate the impact of page faults (in my experience, they dominate over almost all else in most backend network services), and obviously there […]
Original post on fosstodon.org
fosstodon.org
November 7, 2025 at 6:23 PM
@dpiponi @skewray …and obviously, in some contexts, that's extremely correct! Even for regular person scale problems and hardware, people generally underestimate the impact of page faults (in my experience, they dominate over almost all else in most backend network services), and obviously there […]
@dpiponi @skewray Well, it's a tradeoff, right? Obviously you're changing the log base on the amortization, which hits particularly hard early on (while conversely the 30% absolute magnitude is much higher with more growth cycles), so the optimum really has to be very workload-dependent.
November 7, 2025 at 5:48 PM
@dpiponi This also assumes that we monotonically advance to the fill point. Another interesting case to examine is when you have a workflow with stochastic ups and downs in its buffer usage, filling from the front of the array. Reasoning not only about the expected occupancy ratio but also the […]
Original post on fosstodon.org
fosstodon.org
November 7, 2025 at 4:46 PM
@dpiponi This also assumes that we monotonically advance to the fill point. Another interesting case to examine is when you have a workflow with stochastic ups and downs in its buffer usage, filling from the front of the array. Reasoning not only about the expected occupancy ratio but also the […]
So what the quantum gravity paper is really proving is that the physical laws which govern our universe are uncomputable *within our universe*, which is not really that surprising of a result. If we imagine our universe is some projection of a higher-order computational domain, then clearly that […]
Original post on fosstodon.org
fosstodon.org
November 4, 2025 at 6:03 PM
So what the quantum gravity paper is really proving is that the physical laws which govern our universe are uncomputable *within our universe*, which is not really that surprising of a result. If we imagine our universe is some projection of a higher-order computational domain, then clearly that […]
Simpler example: the Halting Problem on finite Turing Machines is uncomputable by a finite Turing Machine, but it *is* computable for an infinite one. Meanwhile, the Halting Problem for infinite Turing Machines is uncomputable by an infinite Turing Machine, but presumably it could be computable […]
Original post on fosstodon.org
fosstodon.org
November 4, 2025 at 6:03 PM
Simpler example: the Halting Problem on finite Turing Machines is uncomputable by a finite Turing Machine, but it *is* computable for an infinite one. Meanwhile, the Halting Problem for infinite Turing Machines is uncomputable by an infinite Turing Machine, but presumably it could be computable […]
@acowley the simulated reasoning models actually do a decent job at exactly this class of problem, since they have more cross checking and self validation. The tradeoff (aside from tokens) is they sometimes also spiral.
October 30, 2025 at 10:55 PM
@acowley the simulated reasoning models actually do a decent job at exactly this class of problem, since they have more cross checking and self validation. The tradeoff (aside from tokens) is they sometimes also spiral.
@dpiponi I really disliked some of the liberties they took with the conclusion of the book, particularly with how they expanded on the bit where it is uh, not itself. But overall I thought it was a superb adaptation and quite faithful to Wells’ vision.
October 21, 2025 at 1:16 PM
@dpiponi I really disliked some of the liberties they took with the conclusion of the book, particularly with how they expanded on the bit where it is uh, not itself. But overall I thought it was a superb adaptation and quite faithful to Wells’ vision.
@ross was hit (fortunately only at 10 ish mph) by someone who ran a stop sign. Bystander (a little old lady) did indeed kick the driver’s ass. I appreciated the ass kicking more than I can express, but I still would have preferred to avoid the ER and the surgery.
September 25, 2025 at 2:12 AM
@ross was hit (fortunately only at 10 ish mph) by someone who ran a stop sign. Bystander (a little old lady) did indeed kick the driver’s ass. I appreciated the ass kicking more than I can express, but I still would have preferred to avoid the ER and the surgery.
@dpiponi I actually kinda like the notification summaries. They're not awesome, but the bar here is "truncated version of the most recent message", which it easily clears. Image recognition generally works well, also, at least for all the plants in my neighborhood. Beyond that…
September 16, 2025 at 6:41 PM
@dpiponi I actually kinda like the notification summaries. They're not awesome, but the bar here is "truncated version of the most recent message", which it easily clears. Image recognition generally works well, also, at least for all the plants in my neighborhood. Beyond that…
@ross Flash drought is actually a thing that happens in the Midwest! The most recent one was in 2023 I think, when we had that really absurdly dry June/July. The problem was it was not only hot and dry but also the humidity was exceptionally low, so the soil rapidly transpirated, which in turn […]
Original post on fosstodon.org
fosstodon.org
September 11, 2025 at 9:54 PM
@ross Flash drought is actually a thing that happens in the Midwest! The most recent one was in 2023 I think, when we had that really absurdly dry June/July. The problem was it was not only hot and dry but also the humidity was exceptionally low, so the soil rapidly transpirated, which in turn […]
@ross I exported mine as a tarball when they opened up that functionality, so it just lives in my local storage somewhere. Same thing with twitter.
August 14, 2025 at 1:11 PM
@ross I exported mine as a tarball when they opened up that functionality, so it just lives in my local storage somewhere. Same thing with twitter.
@ross I actually think it's mostly the domain expiry that rate limits this process. Startups where the founders dumpster-dive the domain registry (or more likely, it falls to them since they initially set it up on their personal registrar and never migrated) tend to have a longer, if not […]
Original post on fosstodon.org
fosstodon.org
July 29, 2025 at 8:34 PM
@ross I actually think it's mostly the domain expiry that rate limits this process. Startups where the founders dumpster-dive the domain registry (or more likely, it falls to them since they initially set it up on their personal registrar and never migrated) tend to have a longer, if not […]
@ross I like this quite a lot! Excellent examples. However, minor correction: with type constructors, the lack of variance annotation is not *invariance*, but rather *polyvariance*. More precisely, it is invariant at the use-site and polyvariant at the call-site (you can instantiate F[_] with […]
Original post on fosstodon.org
fosstodon.org
July 29, 2025 at 2:14 PM
@ross I like this quite a lot! Excellent examples. However, minor correction: with type constructors, the lack of variance annotation is not *invariance*, but rather *polyvariance*. More precisely, it is invariant at the use-site and polyvariant at the call-site (you can instantiate F[_] with […]
@dcz Having been enmeshed in exactly that kind of situation for decades, I know exactly what you mean, and this is what led me to initially rebut my colleague a bit. However, thinking about it, the bankrolling is only effective because of the cachet established by the individual contributors […]
Original post on fosstodon.org
fosstodon.org
July 21, 2025 at 4:25 PM
@dcz Having been enmeshed in exactly that kind of situation for decades, I know exactly what you mean, and this is what led me to initially rebut my colleague a bit. However, thinking about it, the bankrolling is only effective because of the cachet established by the individual contributors […]
@emc2 I feel like sociologists (and as you said, philosophers!) are really missing out by not studying this whole beautiful mess in real time. I have no doubt that future generations will look back on its emergence within this era with a great deal of fascination.
July 17, 2025 at 9:53 PM
@emc2 I feel like sociologists (and as you said, philosophers!) are really missing out by not studying this whole beautiful mess in real time. I have no doubt that future generations will look back on its emergence within this era with a great deal of fascination.
@dpiponi I recently finished the Mistborn trilogy and had a blast despite the fact that Sanderson's characters essentially only exist as a vehicle to carry his own voice to the reader. I find myself extremely annoyed by this now, after the fact, but as I was reading it seemed like a non-issue.
July 17, 2025 at 9:45 PM
@dpiponi I recently finished the Mistborn trilogy and had a blast despite the fact that Sanderson's characters essentially only exist as a vehicle to carry his own voice to the reader. I find myself extremely annoyed by this now, after the fact, but as I was reading it seemed like a non-issue.
@dpiponi I've made the discovery somewhat recently that I'm remarkably tolerant of flaws in books that I'm actively reading. Once I'm done with it, I can take a step back and analyze and realize how crappy things were in various ways, but while I'm actively reading something I usually just […]
Original post on fosstodon.org
fosstodon.org
July 17, 2025 at 9:44 PM
@dpiponi I've made the discovery somewhat recently that I'm remarkably tolerant of flaws in books that I'm actively reading. Once I'm done with it, I can take a step back and analyze and realize how crappy things were in various ways, but while I'm actively reading something I usually just […]
In a real sense, OSS is one of the most democratic things humanity has ever done, while also collectively representing far and away the most complex thing it has ever created. That’s actually kind of incredible. The optimism my coworker feels is not about OSS specifically, much less its present […]
Original post on fosstodon.org
fosstodon.org
July 11, 2025 at 12:02 AM
In a real sense, OSS is one of the most democratic things humanity has ever done, while also collectively representing far and away the most complex thing it has ever created. That’s actually kind of incredible. The optimism my coworker feels is not about OSS specifically, much less its present […]
I think about this a lot. When he first said this to me, I felt it was a little overly rosy. After all, most decisions in OSS tend to be made by a pretty small cabal of well connected maintainers. But at the end of the day, even maintainers like Linus or Guido would mean nothing without the […]
Original post on fosstodon.org
fosstodon.org
July 11, 2025 at 12:00 AM
I think about this a lot. When he first said this to me, I felt it was a little overly rosy. After all, most decisions in OSS tend to be made by a pretty small cabal of well connected maintainers. But at the end of the day, even maintainers like Linus or Guido would mean nothing without the […]