vogu66

The article has comments on pubpeer (below) and comments on the pre-print page. https://pubpeer.com/publications/973ABFB81F504E8CB1B50E941CF...

The gist of several comments is that the paper does not actually demonstrate an accelerated global warming, but instead an acceleration of anthropogenic global warming, when removing the influence of several natural factors. To be clear, they are not discussing the fact that there is global warming, just saying that currently, we cannot say that global warming has been getting faster after 2010 with statistical certainty.

My default is ecosia and below sponsored links there is only the github and pages talking about the thing, no official or unofficial page. I guess that's better?

It gives two sponsored links to openclaw things, so no fake either (presumably, I don't know what they are).

(For the record, that issue has now been fixed by the dev)

maybe it could

I've tried out hys recently (discussed at https://news.ycombinator.com/item?id=46293564). It's been pretty nice, though new, I hit a bug where the feed list disappeared. There are a couple others mentioned in the thread, https://newsboat.org/ seems mature.

Capable doesn't mean capable of office work though, I could see someone with a language disorder doing electronics and have trouble with words, not numbers. Or someone who has trouble with written words specifically doing most of their learning with classes and videos.

not software engineering, but https://practical.engineering/

Additional link with more details: https://www.nobelprize.org/uploads/2025/10/advanced-economic...

I left them exposed for 30 minutes. This is a ridiculous amount of time to leave agar plates open for, since usually you’d only open them for a few seconds.

It's actually really common to let them exposed that long. Not when using them, but to make them: once the agar medium is melted and poured in the petri dishes, closing the lid during cooling causes condensation. Having the petri dish full of water when using it is more difficult and annoying, so it's better to let them cool down with the lid open (a space-efficient way to do so is to have a pyramid of plates where each lid rests on two plates and can support one)

Also, two contaminations out of 4 plates sounds really really bad, but then the blog doesn't say how the agar plates were prepared (how many plates without exposures were contaminated?) and how long they were incubated (sometimes something starts growing after a week or two, if you're culturing a fast-growing bacteria then it's mostly irrelevant).

for what it's worth, my ophthalmologist recommended I use light mode only, with reduced blue and red (because red light actually activates the blue cones too apparently)

Then my screen time started affecting my sleep so I still use dark mode at night

But anyways all software should be configurable and follow the parent software (browser, OS) by default, css even allows for that now. There is even a "prefers-contrast" property in order to design for people who need high contrast stuff.

I do that too, I wonder how much of it is the LLM being helpful and how much of it is the RAG algorithm somehow providing better references to the LLM than a google search can?

that would be client side if it happens, though, I'm talking about server side generation

xsltproc was preinstalled on my machine actually, the fact I could just run it without installing anything is pretty cool

I've been thinking of how to make a blog simple recently, and I came across xslt. It looks really cool and seems pretty set in stone, so I thought I'd ask, what are the advantages/drawbacks of making your own tech stack versus xslt? At first glance, it seems perfectly able to handle rss and other simple linking patterns, and pretty much anything can easily be turned into an xml then xslt could be used to generate an html (server-side, or rather writer-side, not like the blog is gonna change) that you serve?

It isn't only a software license, though it does seem to be primarily aimed at that.

One of the compatible licenses is specifically: "Creative Commons Attribution-ShareAlike v. 3.0 Unported (CC BY-SA 3.0) for works other than software"

well, for now at least

https://news.ycombinator.com/item?id=45409794

You can do that with F-droid, it's a lot of fun seeing what open source app people developed

There's Palemoon's Goanna, also

Sorry, not exactly monopoly. Probably too big to fail, from what I gather, though. I don't know enough to know whether there is anyone else that could realistically compete with their scale.

My point was simply that their size is unrelated to this proposition, outside the fact they are big enough to be taken seriously.

Not mentioned: there would be a single gatekeeper for the internet, Cloudflare.

Cloudflare is already a monopoly though. From what I can tell, what they are saying here, besides proclaiming their continued existence, is that they and AI companies and content creators need the internet to exist.

And what they are building is on top of the 402 response, which anyone can use? So you could use that without using any CDN at all, without too much development cost?

I mean... on the one hand, more people getting in the field may (or may not, I didn't measure) mean less people whose vocation is software.

On the other hand, if that can cheer someone up... * https://ladybird.org/ * https://github.com/gorhill/uBlock * Zig rebuilt a C compiler from scratch https://ziggit.dev/t/zig-as-a-c-and-c-compiler/10963 * Rust rebuilt the core nix utilities from scratch https://github.com/uutils/coreutils I've heard a bunch of people making their own OS from scratch just to see how it works, heck there are guides online (https://github.com/cfenollosa/os-tutorial ... 30k stars...)

So... Cheer up ! If people are building all that, I'm sure innovation and creativity are at least not dead everywhere.

Incidentally, I found out Haiku OS went in beta in September 2018, and that coincides with the bump in unknown in the counter. Not sure how much that's a cause and effect for 2018 or for later, though.

I looked up the numbers in the original paper a little (https://www.nature.com/articles/nature25476)

I don't know much about materials science, but I had a few classes about it.

Seems like their wood gets ~550 MPa in ultimate strength in tension. Seems like their material is brittle (so it behaves like a spring until it breaks), therefore you probably want a safety margin, because at 550 MPa it breaks. Note the unit is a Force/Area, you can compare materials with the same cross-section. In compression they say it's about 160 MPa in axial load, it can be more or less in the other directions (due to wood having fiber it's not the same in all directions, and there they compress it perpendicular to the fiber so they get one direction stronger than the axial load and one weaker, but I guess for a beam you mostly care about axial strength). Torsion and flexion are directly dependent on compression, shear and tension, didn't find shear. Although I'm not entirely sure how it works for materials that aren't the same in all three directions like steel.

For steel, depends on the steel but a quick search (https://www.steelconstruction.info/Steel_material_properties and https://eurocodeapplied.com/design/en1993/steel-design-prope...) says ~200 to 400 MPa in tension for yield, at which point it starts changing shape instead of behaving like a spring, then 350 to 550 MPa for strength, at which point it breaks. I believe in multiple applications they do go apply forces where the metal bends a bit and adapts to its application, but I'm not sure. Regardless, that would mean the wood in tension is equivalent to very strong (presumably very expensive) steel.

In compression, steel is from 170 to 370 MPa apparently(https://blog.redguard.com/compressive-strength-of-steel, didn't find much else easily because numbers were strange on other sources), so I guess steel would win on that one.

But this is comparing the raw strength. In reinforced concrete, you add the metal for tension resistance, concrete is there to sustain compression, so it wouldn't matter much. For beams, the shape of beams is optimised to resist in the direction it needs (e.g. the H cross-section resists to bending in one direction). But you probably can't do that with their wood (they say for now they are limited in shapes), so you'd need more material, and probably it would be stronger overall since you have more material. Question then is how much material (in weight, compared to steel) do you need (they say 10 times less but it probably doesn't take into account the shape), and how much does it cost?

I'm guessing they could also make composite beams at some points too, with not only wood in them.

Then for mechanical applications, there might be also other things that enter the game. In their paper they needed to coat the wood so it wouldn't swell with humidity. For any application with friction, not great. Also, I wouldn't be surprised if it's more sensitive to friction than metals.

Note that the numbers are from 2018, they may have improved the process.

Might be a bit naive, but by the time 2029 comes around, and AI companies have started 'monetising free users', won't a lot of people/companies have open-source models tuned and tailored to their needs running locally, no provider required?

Would working with unums side-step the issue for math problems?

It seems to be mostly from instructables.com, though there are other sources here and there. A search for 'multimeter lead' lead to much better results than instructables though.

Just discovered the existence of large scale armed political extremist groups in the States. Wow. And both left and right wing with any kind of mix of ideologies, too. I wonder how many of them are able to use a computer to find the myriad holes in the world's infrastructure.

Here is a really nice book to give some perspective: End times, by Peter Turchin Best most terrifying birthday present I ever got, I think.

Actually I think I'll put it as a post in and of itself, folks here may like it.

For additional insight, we are not actually able to fully model a single bacterial culture, gene expression is random and we are barely getting to know one gene at a time when considering dynamics and not just steady state (we also have some idea how to represent an average bacterium's gene expression at steady state, if they were all doing the same thing, but they aren't)

As far as I know, we are actually closer to (less far from) in vitro models where you culture human cells into organs very close to real ones and apply drugs to those organs. I think they already do that with skin for cosmetics but everything else is far away (key word: I think, not sure).

I mean, it doesn't need to be centralised on a single entity

There was https://tea.xyz/ that does something like this, kind of, to a point, using a blockchain approach to see which packages are used. An extension of the same concept to gits would be interesting?

I've actually come across and started using Pyjion recently (https://github.com/tonybaloney/pyjion); how does Pypy compare, both in terms of performance and purpose? There seems to be a lot of overlap...