I don’t think the previous system was using space heaters though, the PTAC they had before presumably had a heat pump in it if it was generating condensation.

I think their complaint was it was falling back to resistive heating for most of winter (so basically a space heater)

Yeah AVP/passthrough would make you very sick if you tried to wear it in a jet pulling high G maneuvers.

The temperatures we’re talking about (1000C) would be incredibly dangerous in residential applications, plus a small installation would lose too much energy to the environment due to the ratio of surface area to volume. More practical IMO is to use a daily cycle like what Harvest Thermal is doing: store energy in your water heater tank during the daytime and release it at night.

Apparently I’m the only one here who doesn’t know this but: What is RAG?

Yeah, even on HN that's not an acryonym you should use without explaining.

My first thought for RAG is "Red, Amber, Green" as used on risks etc. with project management.

Retrieval Augmented Generation; basically, find relevant bits from a lot of data, and load the relevant bits into context.

Great answer

It's the buzzword for "make an LLM work on custom data" such as a codebase.

Also radiosity for offline rendering got largely replaced by sampling based techniques, first things like photon mapping and more recently by path tracing. Sampling approaches are easier to parallelize and generalize better to other types of materials.

FYI this post was actually written by Maya the intern (Matt only writes the weekday posts).

Yeah it’s the trials that are really expensive, not the basic research. The government gets to spread its dollars around to lots of individual researchers but if they were instead using that money to fund phase III trials (1) we’d have a lot a lot less money for basic research (2) we’d fund a lot fewer phase III trials so (3) we’d see much less drug development overall. Also I’d worry about a potential conflict of interest if the same government is conducting and evaluating the results of drug trials.

I guess I would have assumed the opposite? A virus that just jumped from bats to humans is expected to evolve pretty quickly in the binding domain as it “learns” to better target human ACE2. I thought the reason they targeted spike was that it’s on the surface of the virus and so an obvious target for antibodies to attach, whereas if you target some structural proteins deep inside the viral structure you’re taking a risk that the antibodies won’t be able to reach them.

Yeah broadcast TV content has long consisted in large part of reruns, syndication, and older (cheap to license) movies, think of what was showing during non prime time hours when you were growing up. If anything there’s much more new content now thanks to the streaming wars.

You need to clean the filters regularly so they don’t become a fire hazard due to oil buildup; the better ones should have some kind of reminder system and will have filters that are dishwasher safe.