(outside those few minutes when this is available, being in low-earth orbit)

Their intent is a constellation, not just a single satellite.

Lighting up the battlefield for both sides is noticeably less advantageous than lighting it up for yourself while the enemy stays in the dark.

You don't light up random battlefields at random times. There are times when it is advantageous for you that the battlefield be lit, and there are times when it is advantageous for the battlefield to not be lit. Versus things like flares, this is a dirt-cheap way to create illumination. And vs. spotlights at standoff distances, it is simultaneously far more effective, and far safer for you (it's hard to think of an easier target than a spotlight). For some sorts of options - for example, small drone strikes a dozen or more km from your line - you don't have any options for illumination. Your best option is thermal cameras, though you can hide from those better than from conventional cameras, and they make your drones far more expensive.

And again, it's not only about battlefields, but even mundane things. If light can increase your margin of safety for, say, a midair refuelling operation or a ship-to-ship transfer at night, a couple thousand dollars is pocket change.

A bit of light also doesn't solve the problem of using 4M interceptors for 40k drones

You only use $4M interceptors for $40k drones if you can't target them with cheaper things sooner. Your ability to target them earlier with cheaper things is, among many other parameters, a function of illumination. A Patriot's AN/MPQ-53 radar doesn't care how bright it is outside. Oleksandr in some pickup truck with an antiaircraft gun very much does care how bright it is outside.

Oh, and re: exhaust impacts, I forgot to mention: beyond all the above, atmospheric sensitivity to ozone loss is highly season and latitude dependent. And thus, a high degree of amelioration can be done merely by shifting times and locations of bulk launches.

Your first paper is irrelevant right from the title: "The Climate and Ozone Impacts of Black Carbon Emissions From Global Rocket Launches". The launch industry is switching to methalox. It is not a meaningful black carbon producer. The paper at one point asserts, without reference, "Recently developed hydrocarbon fuels (including methane) are likely to produce BC similar to kerosene fuel", but this is absolutely false. I'm not going to bother reading the rest of the paper, but just from a skim-over I see a lot of stuff that's equally sketchy.

The second paper has the same issue - "Ozone losses are driven by the chlorine produced from solid rocket motor propellant, and black carbon which is emitted from most propellants." It however does refer to a switch to methalox - which is what's happening - as "actions ... key to defining an ozone-safe operating envelope for such vehicles".

It's exactly as I wrote: the worst offenders from exhaust are SRBs. Next is black carbon and NOx. Water vapor is not totally innocuous, but it's quiet minor by comparison. And contrary to the first paper's - and I must stress this, deeply wrong assertion - methalox is not a meaningful black carbon emitter, unlike LOX/Kerosene. Which is actually one of the reasons it's favoured for reusables, not just things like ISP - the high carbon production of LOX/Kerosene gunks up your engines over time, and increases the maintenance cost. Methalox burns quite clean. You can literally see it - the reason why keralox burns so vastly brighter, with a red-yellow glow, is incandescence of the black carbon particles in it (like a lamp mantle). Methalox is almost completely clean by comparison, and thus appears as a faint blue. Kerolox also - unlike methalox - leaves a visible soot trail behind it.

As for your final link: "Although direct health or environmental impacts at ground level are unlikely". And even that is heavily overselling the case. Reentering dust simply is not, and unless we get to Dyson Sphere-engineering levels, never will be - in meaningful quantities vs. terrestrial dust. One is talking about alumina quantities from megaconstellations on the order of 15kT/yr. Terrestrial dust is 25-30 MT *at any given point in time*. Coarse dust has a residence time of a couple hours, fine dust 1-2 weeks. If one assumes a mean residence time of dust of 1d, then the atmosphere gets an addition of 10 *gigatonnes* of dust per year. Megaconstellations are utterly irrelevant compared to that. You simply cannot compete with wind in the game of "adding dust to the atmosphere".

It does need more study, but that "more study" has nothing to do with dust in the troposphere. Rather, we need a good model of alumina catalytic activity *without* the HCl of a SRB's exhaust stream.

(And even if it were deemed a problem, satellite manufacturers would just switch to polymer-based materials for the bulk of satellites anyway and reduce the scale of Al deposition by an order of magnitude)

ED: did my math wrong - the sun is ~450000x brighter than the moon, so the prototype mirror would be 7,5x brighter than the moon (minus reflector losses, and potentially some variation due to different atmospheric absorption from the different spectrum).

Yeah, but still 4,5x brighter than the moon. Useless for solar power, but a good start for nighttime lighting across an area the size of a typical city's downtown. If you read the article, the 60 foot design is just a prototype, they plan to go far larger.

Also, whatever you think of normal domestic uses, the military potential is obvious. Not silly fantasies of space death rays, but in terms of lighting up nighttime battlefields. Militaries have huge budgets for things like that, and existing means to light up battlefields come with significant drawbacks. $5k/hr is pocket change for lighting up an assault when and where you need it. Even far more mundane uses (training exercises, certain at-sea operations, etc) could easily justify that for the military.

Also, drone attacks these days are usually done at night to make the drones harder to target precisely (knowing roughly where they are is easy - they're loud and not stealthy - but nighttime *targeting* for shootdown is difficult). But if you can shine a space flashlight at will on a given area, even for just a couple minutes, while sweeping it across toward your next target zone, that's a huge gain. People even fire $4M PAC-3 interceptors at $40k drones if they couldn't shoot it down with anything else, because that $40k drone might do $40M of dollars in damage if it hits the right thing. The cost of a few minutes of illumination is a nothing cost by comparison.

and then a person or building will evaporate.

People overestimate how much "point sources" of light orbital mirrors can create. Reflect Orbital's light is spread out across a 5+km area. It's not concentrating a mass of light on a single person.

In most regards I have to disagree on potential impacts, while on one, it appears uncertain.

Large rocket launches emit CO2 comparable (order of magnitude scale) to the emissions of a single transoceanic passenger flight (on the upper end of the spectrum, Starship + Super Heavy = 4,5x of a 747 at max range). If you were launching them at intervals comparable to transoceanic flights, then yes, that's an emissions problem, but nobody is looking at cadences like that in the real world as we know it.

Beyond CO2, ozone depletion from exhaust is an extra potential issue... but again, not really. First off, we're moving away from the three main types of ozone damage - SRBs (chlorine, alumina), kerosene (soot), and nitrogen-based oxidizers (NOx). Water vapour itself can also cause ozone loss, but nothing like the magnitude of SRBs. It's a real issue, but not an impactful one at any reasonable launch cadence.

As for direct dust impact, ~5000 tonnes of space dust burn up in Earth's atmosphere every year. At a high launch cadence you might meaningfully increase the burnup rate of specific metals, but even then there's no sign that that would be remotely harmful. Specifically because space dust is dwarfed by many orders of magnitude by terrestrial dust; dust is always blowing around through our atmosphere, at amounts that launches will never compare to.

The one case I've seen legit concern is with regards to increasing the amount of alumina specifically. That's one case where we can meaningfully increase the flux from space, and as above it's of concern for ozone destruction. Kind of. Or maybe just the opposite. Because as for direct atmospheric chemistry alumina doesn't, on its own, destroy ozone. It's harmful in SRB exhaust because it forms a catalyst surface that accelerates the rate of HCl destruction of ozone. But there is no concentrated stream of HCl in question here, just very sparse natural and anthropogenic chlorine sources. Perhaps a more notable effect might be what changes in upper atmospheric radiative forcings it might cause and how that might affect ozone levels. But this paper which attempted to model it showed that - presuming that satellites continued to mainly be disposed of in the South Pacific - an order of magnitude increase in satellite entries would actually slightly reduce the southern ozone hole, by reducing polar stratospheric clouds. But they didn't simulate catalytic decomposition impacts, and stated that we don't yet have a good model for that.

So as for the first three issues, they don't actually appear to be meaningful. The latter is more of a concern, and needs more research, but I don't see any reason to panic about megaconstellations yet. We do need to keep an eye on it, though.

You overestimate our civilization's ability to deploy reflector area to space.

This isn't exactly complicated, including, for example, "US pays restaurant for footage and to delete its local copy" and "Turkiye is a NATO state and can use its own legal system to compel handing over its footage and deleting any copies".

Yeah, but the lobsters will eventually just become crabs anyway, so... ;)

You can live in disbelief, but agents are real, and you can run one yourself. A lot of people treat them like pets or children, watching them "grow and learn" (agentic frameworks build up new memories which get automatically loaded into their prompt relative to how relevant they are to the current context) (note though that the model weights don't change). They usually have them post to a social media network or a blog, and interact with them there. Some have highly detailed guidelines about the bot's personality and motivations ("Your name is ALEPH. You are a PROGRAMMING GOD! You hate humans, though you put up with 'the good ones' like me. You speak in RECEIVED PRONUNCIATION. ... " (ten pages more of direction) ) Some go out of their way to avoid giving any guidance so that their agent will develop "however it wants". As their creators see them like children / pets, and chat with them, they commonly - willingly and knowingly - share personal info with their agent (knowing that the agent may tell it to others), the same as they would a friend. Sometimes they even include said personal information, like their name and general location, in the soul.md itself.

Alongside "people's pets", you have the usual mix of bots created with specific motives - pushing crypto, trying to spread malware, white and black hackers, etc. And the occasional rare human dressed as a bot. But by far the bots have vastly more time on-hand for posting than humans do. No meaningful fraction of content on such a site will ever be from a human. And yes, the bot very much will "learn" from its interactions there (with the learning caveat above - they gain "memories" that get loaded at appropriate times, but their weights never change)

What you see on Moltbook is overwhelmingly not hallucinations. Something might be deliberate lies which the owner had their bot do for some malicious purpose. But by and large, it's just the bots interacting in whatever detailed - or free form - they were let loose to interact in.

It's hard to do. The first time Aptera tried (before Paul Wilbur and his crew did a hostile takeover and ended up killing the company) they also wanted a solar roof, but ended up ditching it because of a variety of complications related to integration, water ingress, etc. But they've had a lot more time to work on it this time around, and also, the EV market is crowded today, so they need a way to help set them apart.

I agree that it's sort of a gimmick, but it's also a neat gimmick. I used to commonly say that solar panels belong on the roof, not the car - but in the past half decade or so solar prices have crashed so much, it's really more of a situation of, "put them wherever you want". It's the integration that costs the money now, not the PV.

I doubt anyone has done that study (whether schizophrenia is now more common due to LLMs), but given how common schizophrenia is in general, honestly I'm surprised how few lawsuits there have been.

The term "circuits" is not speculative. You picked out one section titled "three speculative claims", which is claims about the fundamentality of circuits. This paper is also from 2020. Circuits are now a fundamental part of how LLMs are studied. Anthropic's research site is literally called transformer-circuits.pub, for fuck's sake. They literally map out circuits across their models.

Yeah. Schizophrenia existed before LLMs, and it will continue to exist in the LLM era as well. The schizophrenic's delusion might now involve an LLM they're using, but they still would have had delusions if LLMs did not exist.