More to the point, they're *guaranteed* bucks.

People don't understand the significance of risk to profitability. By underwriting 80 billion dollars of risk for banks, it's essentially guaranteeing them profits. When it's politically infeasible to spend money on something, the government guarantees loans. That's politically popular across the board because it's spending *later* money and it puts money in bankers' pockets.

Note that if you have a public IPv4 address, you likely don't have use for IPv6rs at all. It's only useful for people hopelessly NATed or firewalled. If you have a routable, reachable address, you don't need ipv6 per se, as every client will be able to reach you. I've never seen a network client that purports "internet access" that can't access an IPv4-only site, including some 'pure ipv6' networks I've been on.

So if Tunnelbroker will work for you, you don't need it for self hosting in the first place.

Think the point is that the pricing isn't really cheaper than the cheapest cloud instances. In fact, you can still find free tiers that are able to serve a lot of self hosting needs. If your needs are light and have nothing to do with your actual house, then the free tier is a much better deal, as your service is likely to be more available than your home innternet connection.

Now once you have need of even a relatively low end desktop grade system, the tunnel becomes *much* cheaper than a cloud instance.

Also wonder how many of the would-be customers think they need it when their ISP does or could provide at least a /64 as part of their existing service. In that scenario, the only thing the service uniquely provides for such a user is an IPv4 reverse proxy, which means you lose control of TLS termination.

So it's going to be a good deal for people with ISPs that won't provide or will filter out external access and also have significant enough needs to push them into pricier instance pricing. It's going to be a bad deal for people with light enough needs to be cheaper to host, or for people that have it naturally with their ISP service.

I've never come across a scenario where my home hosted server was blocked by anything web wise. I don't think blind blacklisting of ip ranges is a significant thing in web client scenarios.

For SMTP, absolutely this is a thing, where SMTP has basically become a relatively small cohort of servers and everyone else is blacklisted by default.

Well, if your 'self hosted' is 'cloud managed', then I could see the complaint. I know a lot of companies are going for that "even on premise is bricked without cloud" business model, and if this is that, then I'd be wary. Especially since I don't think deploying most self-hosted software is actually that hard and don't need a 'cloud seed' to help.

But yes, the tunnel aspect of it seems unavoidable for users without routable addresses. Though at least everyone I know with a vague interest in self hosting at *least* has a /64 natively provided by ISP that is firewalled in a way that can be customized (done on the in-house equipment), so they don't even need a tunnel. Guess there's likely some ancient ISPs that never did the IPv6 thing, but generally ISPs do it because NAT becomes a bigger pain the more traffic tries to traverse it.

It looks like they are truly describing hosting yourself, with optional ipv6 tunnel provider for those stuck behind NAT. Admittedly more external dependency than is ideal, but unavoidable if the isp grants no sort of external address, or filters traffic to make it infeasible. A tunnel at least means you can be sure the "meat" of the service is under your control at least.

To be clear, I think nuclear can and should play a key role in our response to anthropogenic global warming. I just think we shouldn't (a) talk about it like it is *the* answer in and of itself and (b) misunderstand the full breadths of risks and challenges, the most difficult of which are likely to be economic rather than political objections by environmentalists.

Ss you point out, climate change is in effect an economic externality that fossil fuels get a free ride on. This is a key reason for nuclear power's economic non-competitiveness -- in effect fossil fuel use is subsidized by future generations. If you made fossil fuel users pay the true cost of their energy use, nuclear would *instantly* become competitive. But politically that's not going to happen. The only politically possible way around that is to subsidize other energy sources as well.

If you haven't seen any nuclear advocates claim that we should stop investing in renewables, you haven't been paying attention. Usually they come out in response to some article on climate change or perhaps renewables and they will trot out the bogus argument that environmentalists killed nuclear, which is (they say) the only solution to climate change.

The argument that a particular technology is a panacaea isn't confined to nuclear advocates; I think renewable advocates oversell what's possible in the near future, just as anti-renewable people -- and yes, they exist if you're paying attention -- exaggerate renewables' limitations. Really any all-eggs-in-one-basket approach is unnecessarily risky and likely more costly than having several approachs that can work together and compete economically. Key to making that happen will be improvements in grid infrastructure, which will increase the size and therefore the efficiency of the energy market, allowing multiple sources of power to compete.

As for thorium, that's something we'll have to turn to if fission remains a long-term part of our energy supply, but it's not really a help in the time frame we have to respond to climate change. I think the most promising developments are in the development of fail safe reactor technologies and small modular reactors. There are such things as both economies of scale and *dis*-economies of scale, and SMRs are a different way of scaling production than the traditional and every expensive nuclear power plant.

It was never the case that the public being scared caused nuclear to be outlawed, or even *discouraged*. The problem is that investors are scared by the high capital costs, long construction times, and uncertainties about future electricity prices.

This is why nuclear requires government subsidies, either in straight grants, loan guarantees or price guarantees. It's no coincidence that the only country in the world that did a serious nuclear crash program was France, where the electric system was *nationalized*. They didn't go in big for nuclear to make a profit, for them it was a national security issue in result of the OPEC oil embargos. As soon as France privatized its electric system, nuclear construction stalled, just like it did in every other privatized system.

In any case, even if we *were* to underwrite a crash nuclear program, it's neither necessary nor desirable to put *all* our eggs in the nuclear basket. One place we can put investment in is a modernized grid. This will not only help renewable sources like wind and solar, it will be a huge boon to nuclear plants, eliminating questionable siting choices that were driven by the need to locate the plant within 50 miles of customers.

No, I don't think of security as a 'product'. Also you may want to be more specific. The low hanging fruit is the people that put some sort of service on a network with the password 'admin', because it's "trusted". There's no world in which that is a cost saving behavior, it's just supreme laziness. Now if you get into embedded space, particularly with a lot of legacy design components, ok, that 'network' is going to be trusted. I'd eye things *very* skeptically if someone claims it must be a trusted network and at the same time it is an *IP* network. It's not impossible, but it's *highly* likely that the "must be a blindly trusted network" is lack of understanding, or laziness rather than a statement of feasibility or cost.

No, security is not *necessarily* the opposite of usability, the problem is that in a world where "security" guy is over here and the "get stuff done" person is over there, the security guy tends to mandate things without understanding getting stuff done and makes bad recommendations. Further, you have various vendors convincing people their "security" solution improves security, while it merely adds complexity, frustration, and in many times, vulnerability. Ironically enough, a lot of security "solutions" decrease overall security by any reasonable measure (e.g. a web proxy solution that forces a local certificate authority into all browsers to let it man in the middle, a security monitoring system that demands root/admin level access to *everything* and becomes a point of infiltration, etc). If a "security" product doesn't look like a pain in the ass, then unfortunately the business decision callers don't believe it's security.

However, in new deployments and if you demand it of new products, we can have credible security without a huge cost or pain in the ass. The specifics vary situation to situation, but nearly all IP connected strategies I've encountered that previously had "blind trust is just needed" had a credibly secure hardening strategy possible that was also transparent or nearly transparent to operators and users.

Again, some older networks it's not worth the cost to rip and replace, but everyone should be striving for "I don't automatically trust someone just because they could access this specified network" as they integrate new things.

Yeah, that was shocking that they had the thought to put the battery in a remote location, but thought laptop grade electronics and attendant heat management belonged right there with the lenses. Also, to use lots of glass and metal, which while more 'premium' than plastic, weighs a relatively large amount. If you were committed to a tethered pack *anyway*, why not make all the 'not needed on face' stuff go with it?

I might have understood if the battery were right up there, but they were already in 'awkward' territory with tethered battery pack, might as well have went all in.

They might not have gotten to 120G (note the beyond doesn't have 'passthrough' grade cameras nor the lidar sensor array that Apple likely considered a must have, but they probably could have been under 200 or so.

If the setup code is truly random sequence of 12 or so ascii characters only available through inspecting the labeling, then I think that could live on a while.

Problem is that scheme is highly automation hostile, and if applied to something like a fleet of hundreds or thousands of devices, wouldn't be workable. So a device targeting fleet style deployment needs some other strategy. Not sure if the UK law provides some sort of provision that would work for a mass headless deployment.

Well, for serial number, you have some issues:
-Some devices announce their serial number via some unauthenticated mechanism. There are utilities that list devices from a certain vendor and they'll list details like serial number so you can decide if you should/should not try to set it up.
-Even if not announced, the serial number makes for a nice dictionary. If I know that the password is XXXXYYYY where X is some alphabetic prefix number and Y is a numeric incrementing by 1, then the password may well be pretty guessable.

And they specify that it can't be an algorithm with something like serial or mac or a counter as input.

The norm thirty years ago for a hardware store battery was zinc-carbon, with premium batteries being alkaline. The norm today is alkaline, with fancy batteries having a lithium chemistry. So it's absolutely true that the "regular AA" battery you put in your flashlight back then had something like an 800 mah capacity; there is nothing on the market today that is that weak.

In any case that's primary cells, which have zero relevance to this topic. We're mainly interested in secondary cells, and there the improvements in the common rechargeable battery has been dramatic and continual. Thirty years ago the standard hardware store rechargeable was Ni-Cad; a AA probably had about 700 mah capacity. A modern alkaline AA has a capacity of 2000 mah or so roughly 3x as much. This understates the case because modern rechargeable alkalines can typically be recharged easily twice as much as a 1990s NiCad. And *rechargeable* alkalines are getting significantly better almost year to year.

Of course the hardware store battery only has minimal relevance to what we're talking about. What we really care about is Li-ion, and capacity, lifespan and cost for *those* are improving faster than any other battery technology ever has.

Depends on the product.

Let's say you have a router, that router implements enough to work by default, and the 'setup code' is indeed needed to set it up, but the default operational state was good enough that you didn't actually "set it up". There's a lot of devices that are likely to have a "probably good enough" default operational state that may leave the password/setup code alone so that at some point in the undetermined future an attacker gets to be the one to "claim" the device rather than the reasonably considered authorized user.

Now a law trying to filter that out might be nice to make some concessions, as autonomous setup or setup from miles away may want something better than 'obscure code on a sticker'. For example a well known default that is only viable if the device is 'off' and inoperable. Or a time limit from power on after which the setup code becomes inactive (like an hour or something), after which you have to unplug/replug device to get the setup code active again.

You're not going to Gish Gallop your way out of this one. You're the one who brought up your personal experience with the price of batteries at the *hardware store* as proof that batteries have not gotten cheaper. I'm actually being charitable in assuming you're talking about shopping for primary cells; if we're talking *rechargeable* cells the argument is even stronger because they are recharged over and over again which means the steady increase of capacity and lifespan in secondary cells over the decades dramatically lowers your lifetime costs.

As documented in my links above, the cost per energy stored of secondary storage has gone done dramatically in the past twenty years, over 90% since 2000. As for why the Tesla Powerwall isn't dirt cheap yet, customers report waiting months from order to delivery; Tesla already has more customers for this product than it can handle at the current price, why would they drop it? This is Tesla milking the early adopter market segment for a product that they can't produce in high enough volume to sell to the pragmatist market segment.

In any case we're not talking about home storage, we're talking *utility* scale grid storatge with is three orders of magnitude larger. There have been economically successful grid storage projects for years now. Hornsdale in Australia earned back its construction costs in just two years [source]. That's probably close to an ideal econmic situation for grid storage, but as costs continue to drop more and more projects that wouldn't quite clear the normal profit bar will become economically feasible.