Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
jampekka
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
26 ms
·
541.
▲
by
jampekka
2y ago
Page height/vertical scrolling is also broken in both Chrome and Firefox on Android. On FF the comment box gets hidden under the menu bar. Probably the page height gets forced with percentage or vh, which should be usually avoided. If
542.
▲
by
jampekka
2y ago
> They have to pay the bills somehow. The alternative to "consent or pay" is "pay". There was plenty of ad funded media before tracking.
543.
▲
by
jampekka
2y ago
A propaganda machine? It's an openly center-left newspaper.
544.
▲
by
jampekka
2y ago
Majority of Americans support stricter gun laws (radical anti-gun stance?). The gun nuts are a loud minority, and dems are unlikely to win elections where they have sway anyway. https://news.gallup.com/poll/1645/gu
545.
▲
by
jampekka
2y ago
MAGA is not popular. Only about 50% of republicans are MAGA and about 30% of the population are republicans. Being generous, MAGA has something like 20% support. Trump's win was hardly a destruction. It was quite slim.
546.
▲
by
jampekka
2y ago
Trump is unpopular. He's never gotten majority approval in a Gallup poll for example. https://news.gallup.com/interactives/507569/presidential-job...
547.
▲
by
jampekka
2y ago
> 2. Drop the radical anti-gun stance The Democrats have a radical pro-gun stance compared to about any other country in the world.
548.
▲
by
jampekka
2y ago
The post mentioned offline usage for maps and notes. Neither are significantly limited by service workers' capabilities. Platform differences are annoying indeed, especially due to the deliberate sabotage by Apple. Sure there are limit
549.
▲
by
jampekka
2y ago
> The browsers are controlled and manipulated by the likes of Apple and Google. Who do you think controls Android and iOS native APIs? Web standards at least have public forums and specs, with multiple parties involved. And all the major
550.
▲
by
jampekka
2y ago
PWAs can be fully offline. Are you sure you understand what you criticize?
551.
▲
by
jampekka
2y ago
Sorry, half a century.
552.
▲
by
jampekka
2y ago
It actually predates computing machines by about half a decade.
553.
▲
by
jampekka
2y ago
The material system in Blender 1.0 was vastly simpler than it is currently. If you want to use a panel instead of nodes, the material panel is not that dissimilar from the 1.0 screenshot. That said, I find the node interface a lot more powe
554.
▲
by
jampekka
2y ago
Specific features are hard to find from tutorials. LLMs don't replace tutorials, they augment documentation and stack exchange and such.
555.
▲
by
jampekka
2y ago
Do you mean that the exact prompt for tool use shouldn't matter? Has this been tested? Is the tool use trained with a variety of prompt styles? I would imagine training with a specific, perhaps structured, prompt could make the functio
556.
▲
by
jampekka
2y ago
This sounds a bit like the attitude that held back Blender UI for a long time. There were people seriously objecting introduction of the undo feature FFS. The UI stockholm syndrom is real. The 2.8 overhaul was not "slapping on a dark t
557.
▲
by
jampekka
2y ago
PCA was introduced in 1901. Who knew Karl Pearson was actually doing AI!
558.
▲
by
jampekka
2y ago
Blender's UI has improved tremendously over the years, but I wouldn't call it one of the cleanest UI in software. It still has, for better or worse, quite idiosyncratic UI patterns (e.g. the panel juggling) and the underlying logi
559.
▲
by
jampekka
2y ago
The major overhaul was Blender 2.8. You can see a brief overviews of changes in the release notes: https://www.blender.org/download/releases/2-80/ It may be a bit hard to appreciate how much it changed (got b
560.
▲
by
jampekka
2y ago
The statistics in the Pras & Guastavino paper seem a bit suspect. They do a lot of significance tests with no multiple comparisons compensation, and find few statistically significant results. The problem is that with enough testing, yo
561.
▲
by
jampekka
2y ago
20 ms on top of other latencies (buffers) is not good for e.g. live monitoring.
562.
▲
by
jampekka
2y ago
Leaves me wondering how much "ultra-low" is in milliseconds. Is it essentially a USB audio device?
563.
▲
by
jampekka
2y ago
Is the format used in the examples the same that's used in the function calling instruction training, i.e. should it be the optimal prompt for function calling? I find it a bit frustrating when details of the training is not known and
564.
▲
by
jampekka
2y ago
How so? Businesses and middlemen make a killing with non-web-apps.
565.
▲
by
jampekka
2y ago
Then use it as a web page?
566.
▲
by
jampekka
2y ago
I've been wondering how the training process of the huge models works in practice. If an optimization run costs millions, they probably don't just run a grid of hyperparameters.
567.
▲
by
jampekka
2y ago
Azure is the biggest revenue source, but my guess is (didn't find data) that Office is more profitable. Although Microsoft is especially good/evil at getting Azure lock-ins and overpriced exclusive deals from e.g. governments.
568.
▲
by
jampekka
2y ago
The monopoly is live and well with Office, which is their real money maker.
569.
▲
by
jampekka
2y ago
Laws are usually passed with longer texts that explain why the law was passed. These are consulted by courts if there are issues in interpreting a law.
570.
▲
by
jampekka
2y ago
Calling code across languages at least is very widespread. And typically requires custom, often clunky, solutions for all language pairs.
More ›