jampekka

9 ms·

151.
by jampekka 6mo ago
Edit: This is incorrect, as pointed out below. Pulseaudio still does the device juggling etc on most systems even when there's a pipewire backend.
152.
by jampekka 6mo ago
> I bet they don't publish Linux numbers because it depends on which desktop you use etc. They ship with Ubuntu on it, which would be quite natural choice for such benchmark. Also they do do the standby test on Ubuntu for some reaso
153.
by jampekka 6mo ago
> Having mainline Linux on a system with 24h+ battery life in a 13" case is pretty damn impressive. Does it have such battery life on Linux? The benchmarks, apart from suspend battery life, are for Windows.
154.
by jampekka 6mo ago
You can have over 100°C. The amount of steam is a crucial factor. Even over 100°C doesn't feel that hot if it's dry, but at higher temperatures the "löyly profile" often becomes quite harsh.
155.
by jampekka 6mo ago
Finns don't do it that long. It's typically 5-15 minutes per sitting, with 2-4 sittings per session.
156.
by jampekka 6mo ago
95 is high for Finnish saunas in Finland at least. Public saunas are very rarely so hot here, and few like it that hot. Edit: to put it into some numbers, per one study[1] Finnish sauna sessions were on average at 75.9°C with SD 9.9°C. If w
157.
by jampekka 6mo ago
> But you can't really assume that the estimate of the mean represents the real value. For example, if the sensor is equally likely to show 80 or 81 BPM when the real heartrate is 80.7, the mean estimator will be biased. Bias is dif
158.
by jampekka 6mo ago
I don't understand what you mean by that. Precision (inverse of variance) of estimate of mean increases directly proportional to number of samples (given some assumptions that very likely hold here). If you have measurement standard de
159.
by jampekka 6mo ago
> What wearables were used? These devices don't usually have enough precision to reliably detect ~3bpm changes. For large sample averages this doesn't really matter.
160.
by jampekka 6mo ago
I'd like to see a bit more detailed methods. - How was the controlling for the other factors done? A linear model? - What were the sauna vs non-sauna baseline HRs in fig 1? Could you plot raw averages? - Was the min HR explicitly compu
161.
by jampekka 6mo ago
How about not writing the name at all? It's already in the email header.
162.
by jampekka 6mo ago
To me Gnome mostly feels like someone deliberately researched how to make desktop apps as intuitive as possible and implemented that as the policy. And I guess that's what they did, and they did a good job.
163.
by jampekka 6mo ago
That is part of it, but e.g. GDPR enforcement has been dismal since its inception, also against European companies. E.g. "Pay or Okay" is widespread, although it's clearly illegal. https://noyb.eu/en/noyb
164.
by jampekka 6mo ago
It's that EU has illusion of regulation that is nerfed with either loopholes or lack of enforcement.
165.
by jampekka 6mo ago
LLMs nowadays make aggressive use of web search. Thus they don't answer only on the base of what they were trained on. I don't think they are much more prone to using only the same popular frameworks, especially if you ask them to
166.
by jampekka 6mo ago
I've found that models have improved here significantly in past few months. They have the tendency to pile on ad-hoc solutions by default, but are capable of doing better architectural decisions too if asked. A better design can be mad
167.
by jampekka 6mo ago
What are you on about? I don't know of any software or services that would be banned at my university. People use all sorts of LLMs extensively. At least in Finland also civil servants are free to use what AI services they want, given
168.
by jampekka 6mo ago
> Code quality and IDEs aren't going anywhere, especially in complex enterprise systems. Was code quality ever there in complex enterprise systems?
169.
by jampekka 6mo ago
I think websites via DOM are gonna be quite easy for the models.
170.
by jampekka 6mo ago
Lots of scepticism here, but I think this may really take off. After 25 years of heavy CLI use, lately I've found myself using codex (in terminal) for terminal tasks I've previously done using CLI commands. If someone manages to m
171.
by jampekka 6mo ago
> Just the obvious one: the people who designed IPv6 didn't design for backwards compatibility. Nor for easy transition.
172.
by jampekka 6mo ago
I use Firefox on both Linux and Android for 99% of my web browsing needs. At least for me it's the best browser out there, and doesn't seem neglegted at all.
173.
by jampekka 6mo ago
Firefox on iOS isn't really a Firefox because Apple doesn't allow alternative browsers. It's a Safari skin.
174.
by jampekka 6mo ago
To me it seems like they burn so much money they can do lots of things in parallel. My guess would be that e.g. codex and sora are very independently developed. After all there's a quite a hard limit on how many bodies are beneficial t
175.
by jampekka 6mo ago
There are Android and iOS apps. It's advertised as a "white label solution", meaning productized apps should be built on it. https://ageverification.dev/Setup
176.
by jampekka 6mo ago
The app is available only on Android (and maybe iOS) and requires Google/Apple attestation. There's been extensive discussion about this, but the developers refuse to even fully acknowledge the problem.
177.
by jampekka 6mo ago
The slogan "Age Verification Solution for all Europeans" is a lie. The honest slogan is "Age Verification Solution for Europeans Who Use Mobile Devices that are Controlled by US Corporations and by Extension US Government Whi
178.
by jampekka 6mo ago
There's at least a strong case that Archive.is is Russian run: https://gyrovague.com/2023/08/05/archive-today-on-the-trail-...
179.
by jampekka 6mo ago
Z-library was/is very likely run by Russians. They were even arrested by FBI, but escaped. Archive.is is likely run by a Russian. LibGen was run by Russians.
180.
by jampekka 6mo ago
Oftentimes better understanding of the problem needs trying out solutions. Armchair architectures tend to blow up in contact with reality.

More ›