sohex

Do people just not even search their proposed name anymore?

It is interesting to see the degree to which Claude voice creeps into normal speech. It also leads to some annoying catch-22’s, like rejecting an em-dash where you might have genuinely written one to avoid the appearance of LLM text.

Thankfully I’m lazy and have always used hyphens everywhere so at least I’m safe on that point.

Oh very nice, bliss looks pretty slick. I see musefs as being kind of an unglamorous middle man layer that ultimately just has one job, separate the metadata handling from the backing files.

Since the DB is the source of truth you can absolutely version your metadata. The portable mirror use case is definitely one that people run into, but I see musefs more enabling that than handling it directly. It’d be really easy to handle transcoding and art handling just pointing at a musefs mount or by having a plugin for whatever you’re using handle it.

I definitely agree that continuous automation is the way to go, personally I set it up differently though. Whenever I have a new album in my music folder it gets detected and then beets imports it and museFS autoscans it. That handles 99% of cases silently in the background and the other 1% are the ones I’d want surfaced to make a judgement call on anyway.

I first read this in the jargon file when I must have been… 10? 12? And it’s been stuck in my head ever since.

Sonnet, GPT-5.2, Gemini Flash, in a set of 21 games, where conclusions are drawn from the LLMs self reported reasoning.

This is like writing a paper about kids in a literal sandbox fighting over ‘territory’.

The models employed don’t indicate the actual extents of machine reasoning even as we currently recognize them. They certainly don’t have the metacognition necessary to accurately understand their own reasoning. As we’ve seen with recent papers on how LLMs do math there’s a complete disconnect between actual and reported mechanism.

“Chilling” shouldn’t be the take away here.

Is it really though? Now admittedly I’ve never worked with LexisNexis or Westlaw, but would it really be that difficult to have a tool just check if the citations actually exist?

IIIRC in terms of clients mutt (&co) will actually handle “@“ in the local part correctly.

But the real reason I do that is just because I just like to sit in anger whenever this breaks the user experience because of programming errors or inconsistencies.

Genuinely delighted by the fact that I’m not alone in that.

Namespace collision

Well I have a new favorite website. I don’t know the last time that I read something that was so thoroughly and multidimensionally my shit.

Not actually measuring crispness when you self report having the perfect equipment to do so is a cruel, cruel tease though.

Shout out to the best named feminist group of all time, W.I.T.C.H. The Women’s International Terrorist Conspiracy from Hell.

This has a serious case of in-housing the edge cases that I don’t think many would want to pay the price for.

The problem with DNS per the haiku isn’t that it’s difficult to understand, or even that running your own DNS server is particularly difficult. It’s that coordinating information and exchange at scale is a tricky problem with a lot of non-obvious edge cases and foot guns.

So trying to reduce complexity by sidestepping DNS really doesn’t do that, it just leaves you holding the bag on all the problems that DNS was quietly solving for you in the background.

My interpretation would be that he feels his wife is incredibly loving in a quantity he isn’t able to match (degree) and in a unique way he’s not able to match (kind). General life experience plus the fact that he wrote that tells me he’s probably wrong and his wife would probably say the same about him, but that’s just speculation.

I’ve had one experience with Claude Code so far that genuinely frustrated me, but it did it to such a degree that it wrapped back around to hilarity. It tried to run a command got an error, realized it needed to cd to a different directory first, and then… didn’t do that.

It tried itself several times going “oh, I didn’t actually cd, let me add that and try again”. I tried correcting it several times “you MUST begin the command with `cd dir &&`”. There were a lot of variations back and forth to try to coax out the correct tool call. Including backing up the conversation and trying from earlier in the context.

It refused. Every time. It simply would not include the cd. Genuinely unhinged behavior.

Wildly true. Purely anecdotal on my part, but in my experience the cooking stack exchange was just awful to even try to ask something on.

My initial thought was that that would probably be taking the simulationism too far. My second thought was how funny it would be to swap calendars to get out of a truce early.

“My Lord, you can attack them until March of next year!”

“Then March of next year it shall be!”

I think that largely depends on how you engage with the internet. To continue the metaphor, the internet has an absurdly high noise floor, but it’s easily filtered noise. So if you do that filtering you’re actually left with a pretty high SNR.

I think this article hand waves and side steps what I see as two notable issues.

The main one is aquifer depletion. The consumptive vs withdrawal argument mostly holds water, but consumptive is a sliding scale. Evaporative cooling for data centers is solidly at the far ‘truly consumptive’ end of that scale because the consumed water will not reenter the local watershed. That’s problematic because aquifers are very slow to refill. So this is genuinely a concern in water stressed areas.

The other is the weak growth model. I suspect we’re only going to see faster and faster growth of data centers in coming years, making the consumption there more exponential than linear. Meanwhile the majority of the other consumptive consumers are strongly tied to demographics and population growth is slowing everywhere. For example agricultural water use in the US has held steady or even declined in recent years.

In fairness, part of that agricultural decline in use is from advancement in technology and methodology and we’ll likely see the same with data centers, but those numbers are unpredictable.

On the whole I agree that the concern over data centers in terms of water (and electricity) usage is overblown to an extent, but I think we do need to pay closer attention to the points that actually matter when looking at the situation.

Accelerando and The Quantum Thief by Hannu Rajaniemi (and that series as a whole) are the best examples of how weird the future is going to get I’ve read.

Other series like The Culture are amazing too, but the aforementioned feel possible in a way that others don’t. For me, I can see the causal chains leading from here to there vividly in a way that you don’t get with a lot of other sci-fi.

That combination of plausible weirdness is unique and I’d highly recommend The Quantum Thief to anyone who enjoyed Accelerando or Stross’ other writing.

Hm, with regards to the difficulty of scaling these systems specifically and really the whole topic at hand more generally I don't know that there is sufficient domain specific data to justify a viewpoint either way. I mean how many examples of scaling a national healthcare system (public, private, or otherwise) to support 300+ million people do we have? Literally three, right? The US, China, and India. Each of which has such significantly different circumstances that comparing them is less apples to oranges than it is apples to giraffes to glaciers. That being the case I feel like it's only reasonable to draw upon non-domain specific data wherein so far as I'm aware and in my experience we see a pattern of scaling existing systems being simpler than originating new ones.

Pretty much this. For further consideration it may be worth looking into Daniel Dennett's writing on qualia (notably that they're ineffable and private).

I feel like it's apparent why comparing healthcare in the US with healthcare in, well, any of the aforementioned countries, is problematic. If you look at Western Europe (a bit less populous than the US) or the EU as a whole (a bit more population than the US) they have functional healthcare at much more reasonable rates than the US. Scale matters, but scale is also a matter of division. If a single system can't serve 300+ million people then it can be broken down into regional systems or state systems. That being said medicare already serves something like 60 million people. I'd argue that scaling a system to support 5x the number it currently serves is significantly more doable than scaling anything from zero.

The file uploading functionality of ChatGPT is just awful, it has nothing to do with the file name. You can test it yourself with any arbitrary file, the number of failures to upload you experience will be significantly higher than you would experience with, I'd hazard to guess, any other upload function around the internet. Now whether that's something with their processing pipeline or just their servers being perpetually overwhelmed I have no idea, but it's almost certainly a case of ineptitude, not malice.

Using fossil fuels for plastics has to be better environmentally speaking than using them for energy. Simply by virtue of the fact that the carbon is trapped as a constituent of the plastic, not binding off as CO2 and floating away.

That aside though, even if it was ‘just’ a 10x reduction in usage for energy that would still be a huge win because that becomes massively easier to mitigate and offset in effective ways.

Every time you increase or decrease the scale of a problem by an order of magnitude it changes the dynamics of approaching it significantly.

Well the question at that point is, protected from whom? Security is never a binary state, it can only be assessed relative to a given threat model. I'd say that being protected from all the sites that don't use fingerprinting scripts is certainly better than not. If fingerprinting is a concern to you though grab a factory default iPhone, should be pretty a pretty common fingerprint.

I think you might be trying to form a model based on false premises here. The universe doesn't know anything. Or to put it another way, the universe doesn't know anything because it is everything. You seem to be assuming a sort of director and actor based model wherein the universe (the director) needs to be telling the planets and stars (the actors) what to do, but that's not the case. The planets and stars interact with each other via the various forces that they're subject to and their state may or may not change as a result.

I guess, to generalize, my point is that things are as they are as a consequence of the physics that they're subject to. A rock has its particular color, weight, and texture due to its elemental composition. A molecule doesn't need to know what it is in order to reflect or absorb certain wavelengths of light, that's what naturally occurs when those wavelengths interact with particles that have a certain composition and state. There's no conscious direction happening at any point, nor is there any data being computed.

The universe is essentially a medium that exists with certain properties and everything in it is stuff that also exists with certain properties. We call the manner in which these properties interact physics. So in summation I'd say that there are intrinsic properties and emergent phenomena based on those properties, that's where a rock gets its color, weight, and texture.

I mean, the combination of not being laminated and disintegrating with ease is the point so far as I understand it. The idea being that if you ever drop your wallet with it in it somewhere then hopefully environmental conditions will degrade it beyond recognition before someone can find it and make use of it. Not saying that's a good or bad idea, just that it is in fact intentional.

That strikes me as being a much higher unit cost than I would have anticipated. Are you way overprovisioned on the arduino and the steppers or is there something else fancy going on that's driving costs up?

Are we talking about IPv4 addresses or are we talking about the internet as a whole? Because that quote seems to be about the latter and I was referring to the former above. I read the Licklider paper that they link and it's much more on the abstract side rather than discussing actual implementation details.

Licklider and Taylor did have some very interesting predictions about how the internet would shape up though. Probably my favorite quote from the article:

"Unemployment would disappear from the face of the earth forever, for consider the magnitude of the task of adapting the network’s software to all the new generations of computer, coming closer and closer upon the heels of their predecessors until the entire population of the world is caught up in an infinite crescendo of on-line interactive debugging."

The final few pages on about what internet communities will look like is very interesting. They discuss the idea of an OLIVER, that is an "“on-line interactive vicarious expediter and responder”, which seems not terribly far from ChatGPT + your google data.

So far as I'm aware IP addresses exist out of a functional need, not a moral imperative, but if you have a source with regards to that I'd love to read it. You can hardly call the number of IPv4 addresses that AWS has a monopoly either, iirc they have something like 2% of the IPv4 space? Not charging for them was comparatively strange given that other cloud providers have been charging for them for a good long while. Managing their address space isn't a zero cost endeavor for AWS, so why wouldn't they charge for it? At the least to cover their costs, but given the way the world works these days making a profit isn't weird or unexpected either.

If you're really upset about it you can go through the trouble of registering your own ASN and get on the list to get your own allocation from ARIN. No one is stopping you from doing so.

I think there's an enormous number of things that it's very worth criticizing AWS (and other cloud providers) for, but this really isn't one of them in my book.