The PoC code should work. You just need to install Kass as a dependency. If you have done that, are there any other issues you are facing?
As far as risks are concerned: any app with the ability to get a send right to NetAuthAgent (pretty much any un-sandboxed app) can just silently as NetAuthAgent for any saved credentials for file drives (FTP, WebDAV, Samba, etc.), as well as chaining into a leak of all iCloud Contacts and Calendars (plus other stuff from iCloud). Sandboxing makes it difficult, but not impossible.
The risks are zero if you're up to date (and the patch was in October of last year, so you honestly should be up to date already). If you are not up to date for whatever reason and choose not to be, the risks are far more (unless you diligently check every single process that ever runs on your device).
Aren’t viruses much harder to detect than bacteria? Viruses are generally smaller and are completely inert without a host cell. Bacteria, besides be larger, also have their own metabolic processes and distinct structures you can do things like grow them in a laboratory culture until the colony is much obvious.
Your comment makes it sound like bacteria are harder to detect but if we’re already identifying viruses, locating bacteria seems easier.
(Though some viruses are bigger than the smallest bacteria, like Mycoplasma at 200nm, viruses are generally smaller)
What makes a hypothetical brain microbiome so hard to find? I would think that once you’re doing microscopy on brain slices that a biome would show you quite fast. But if you’re still optimistic after a negative search I assume there must be many reasons why a brain microbiome could exist but still be hard to detect.
> He foresees great platforms all around the ocean, harvesting energy from magma holes and using it to produce low-carbon synthetic fuels that could be shipped to shore.
This sounds like an energy storage solution as well? Where would the low carbon mass come for such a fuel production scheme? Haven’t heard this idea for energy storage/transport before.
It is weird that proposes floating platforms where there are plenty of volcanoes on land that could be tapped. Much easier than going through miles of ocean before drilling. Also, land can be connected to the grid; they could run cables from Iceland to Europe.
Making low carbon synthetic fuels is talked about a lot here and elsewhere. The main problem is that it is more expensive than hydrogen. But lets you use existing vehicles but with much larger fuel costs.
We know that the oceans are a sink for atmospheric carbon dioxide (oceanic acidification is a direct consequence of this), and that it's energetically less costly (~30%--60% of atmospheric capture) and probably technically simpler to recover CO2 from seawater than the atmosphere.
The concept's been studied for decades at M.I.T. and the US Naval Research Laboratory, the latter of which released a flurry of research articles in the 2010s on the concept. I'd written of those a ways back at Reddit:
If you take the CO2 out of sea water to create synthetic fuel, wouldn't that contribute to the green house effect? (Unless the CO2 is captured when that synthetic fuel is burned, but in that case why not use non-synthetic fuel, for a more economical solution with the same carbon footprint)
1. Both atmospheric and oceanic CO2 are biospheric reserves, that is, already present within the biologically-active portion of the carbon cycle, as opposed to fossil carbon, which had been sequestered over hundreds of millions of years and represents a vast store which humans have been re-introducing to the biosphere at rates millions of times greater than which it was originally sequestered. That is, it's the reintroduction of this vast bolus of carbon to the atmosphere and oceans which is problematic.
2. The atmosphere and oceans are in rough equilibrium, with as I'd said above the oceans acting as a sink for atmospheric carbon. Removal of carbon from the oceans means that further excess atmospheric carbon can enter into solution. A key current concern is how much carbon can be absorbed into the oceans, and any net removal should increase the rate at which atmospheric carbon is dissolved. Given my first point this is something of a net flush, but it means that the net effect still remains carbon neutral.
3. Ocean acidification by way of CO2 absorption is already problematic, so any incidental reduction is advantageous of itself, though I suspect net effects would be small. What impacts localised reduction might be (as in the immediate neighbourhood of Iceland) I don't know.
There are futher considerations, notably that extracted carbon might itself be sequestered or stored (we know that petroleum analogues are stable over exstremely long terms --- tens to hundreds of millions of years), and that there are current applications for which there are few reasonable alternatives to petroleum analogues, notably in marine, air, and rocket transportation.
There is already too much CO2 in seawater. Taking it from air will also reduce levels in water. If we stop adding new carbon from underground, in several thousand years levels should return to current normal, but this may be too long for our civilization.
I feel quite strongly that LLMs are not a fad. Something like LLMs will be useful tools to people decades from now. Especially given that a whole generation of students is now using/relying on them for course work.
Even if we want to call them sophisticated autocomplete, why would I ever switch back to a dumb autocomplete?
> I build "brushes" for refactoring my code every day.
Anyone know what the author means by brushes?
Also any recommendations for non-Copilot IDE LLM integrations? I've tried a couple but they felt far behind Copilot in terms of quality and smooth IDE integration.
You select some code and apply a brush and it'll use a preset prompt to modify it: "Make Readable", "Add Types", "Document" are some of the options.
Copilot has a pretty poor implementation, and mostly they just butcher your code - but they seem very powerful, especially with custom brushes.
Things like "replace this PHP5 class with a new PHP8 readonly dataclass", "Convert this react components to have imageLinks as prop srcsets". Nothing fantastic, but damn this stuff is tedious to do manually. Now I can just batch process my files et voila.
Realistically what are the risks?