Why don’t you take a moment to explain to the class why you think web crawling means you can’t cache anything?

It seems to me that the very first thing I’d try to solve if I were writing a tool for an LLM to search the web, would be caching.

An LLM should have to go through a proxy to fetch any URL. That proxy should be caching results. The cache should be stored on the LLM’s company’s servers. It should not be independently hitting the same endpoint repeatedly any time it wants to fetch the same URL for its users.

Is it expensive to cache everything the LLM fetches? You betcha. Can they afford to spend of the billions they have for capex to buy some fucking hard drives? Absolutely. If archive.org can do it via funding from donations, a trillion dollar AI company should have no problem.

> But it's often disabled for the same reason as having router-level firewalls in the first place.

Yeah, anything that allows hosts to signal that they want to accept connections, is likely the first thing a typical admin would want to turn off.

It’s interesting because nowadays it’s egress that is the real worry. The first thing malware does is phone home to its CNC address and that connection is used to actually control nodes in a bot net. Ingress being disabled doesn’t really net you all that much nowadays when it comes to restricting malware.

In an ideal world we’d have IPv6 in the 90’s and it would have been “normal” for firewalls to be things you have on your local machine, and not at the router level, and allowing ports is something the OS can prompt the user to do (similar to how Windows does it today with “do you want to allow this application to listen for connections” prompt.) But even if that were the case I’m sure we would have still added “block all ingress” as a best practice for firewalls along the way regardless.

> Ingress being disabled doesn’t really net you all that much nowadays when it comes to restricting malware.

But how much of this is because ingress is typically disabled so ingress attacks are less valuable relative to exploiting humans in the loop to install something that ends up using egress as part of it's function.

Since we're talking about programs that are trying to set up a connection no matter what, I'm going to say "not much". It's not significantly shrinking the attack surface and forcing attackers onto a plan B that's meaningfully harder to do. It just adds this layer of awkwardness to everything, and attackers shrug and adapt.

You block inbound to block inbound. Of course it doesn’t do anything for outbound. Acting like you can just turn inbound filtering off because of that is disingenuous.

Nobody suggested "just turn inbound filtering off"?? We're talking about an alternate universe of program design.

And we're talking about malware in general, not inbound or outbound specifically.

This is what I keep thinking.

- Theres a 38 year old woman in the crew

- It’s a medical condition that likely wasn’t present when the mission started 4 months ago

- It’s serious enough to return the crew, but not serious enough that they must do so immediately

I guess we’ll find out in 9 months? (Or not…)

As I understand it, the studies done with mice suggest that microgravity prevents normal embryo development. The ISS should therefore be regarded as a teratogenic environment, and I'd be shocked if women of childbearing age weren't prescribed highly-effective contraceptives (ie. IUD/IUS or implant) before, during, and after spaceflight.

I’m sure they were prescribed, but it’s always possible for them to fail.

I’m curious at what point in the embryo’s development the zero-g becomes an issue, if its immediate vs long term thing. It’s very possible that if it was pregnancy, the embryo is already not viable but she still needs some procedures to ensure her own health (a DnC, etc) that are important but not enough for an emergency evac.

There are probably a hundred ailments or illnesses that can fit this description, maybe someone noticed a swollen lymph node or lump somewhere

Yeah it’s definitely just a thought. Getting pregnant in space the sort of sordid thing that’s fun to speculate on, but ultimately we just don’t have enough information. We’ll probably never know, either.

Are you implying that the pregnancy condition occurred onboard?

Yes. Sex isn’t allowed on the ISS due to complications with pregnancy, but it’s not crazy to imagine that maybe they just did it anyway. (Who wouldn’t want to? It’s sex in space and it sounds amazing.)

I don’t know much about the TPM but if it’s anything like Apple’s Secure Enclave, it should require exponentially longer time after each incorrect PIN past the first one, making it so you can’t reasonably brute force it without getting lucky.

I’m not sure how the typical “two factor” best practices would interpret one of the factors basically self destructing after 10 guesses, but IMO it’s a pretty decent system if done right.

That's not the issue. The TPM isn't blinded in the above description meaning that if someone cracks the TPM they can get your key. Ideally both factors are always required to access the secret.

If you're wondering, yes this is a security issue in practice. There have been TPM vulnerabilities in the past that enabled exfiltration of secrets.

Strange that you assume a single solar panel? If you’re installing solar at home, wouldn’t you install more than one panel?

I have a 15KW setup at home (kinda large, but adjust the numbers to what’s reasonable for you), it should charge the Tesla in less than 4 hours.

> there are ABI breaks, but I only remember only one really bad one with std::basic_string and C++11

ABI breaks are everywhere in C++... consider a class

  class foo {
  public:
    foo();
    void do_something();
  private:
    int x;
  }

and an impl in the dynamic library:

  #include "foo.h"
  foo::foo() {
    this->x = 0;
  }
  void foo::do_something() { std::cout << this->x << std::endl; }

You build a libfoo.so and clients use it, calling `foo f; f.do_something();`, it calls your library and it's great.

But as soon as you ship a new version that adds a new field to foo (still source-code compatible):

  class foo {
  public:
    foo();
    void do_something();
  private:
    int x;
    int y;
  };

With a new function body in your shared object:

  void foo::do_something() { std::cout << this->x << this->y << std::endl; }

You're hosed. Clients have to recompile, or they get:

  $ ./main
  0, 0
  *** stack smashing detected ***: terminated
  [1]    2625391 IOT instruction (core dumped)  ./main

Because the size information for an instance of foo is only known at compile-time, so the clients aren't allocating enough space on the stack for it (ditto the heap if you're using `new foo();`)

The way around this is awkward and involves the pimpl pattern and moving all your constructors out-of-line... but you also need to freeze all virtual methods (even just adding a new one breaks ABI), and avoid using any template-heavy std:: types (not even std::string), since those are often fragile.

Most people just give up and offer an extern "C" API, because that has the added benefit that it's compatible across compilers.

"true" C++ shared libraries are crazy difficult to maintain. It's the reason microsoft invented COM.

Swift goes through crazy lengths to make this work, and it's impressive: https://faultlore.com/blah/swift-abi/#resilient-type-layout

Rust would have to do something like Swift is doing, and that's probably never going to happen.

Sure, you can break ABI in a library, but the compiler doesn't force you to, is what I mean. Same as breaking ABI in C when you change a struct...

It’s range anxiety and it’s real. And it’s not entirely unjustified. Where I live in the Midwest I have literally never seen a public charging station. I know they exist because I can search for them on maps app and I see dots, but it speaks to their general small numbers that I’ve never seen one in person that I can remember.

Now, prior to this, I lived in California for many years from 2011 until 2019, and I saw tons of EV charging stations there. I left with the impression of “wow, charging stations are everywhere”, and that was 7 years ago.

But now in my Midwest metro area, I can honestly say there are zero that I can think of within a 10 mile radius of my house. Not one. (They’re out there somewhere, but they gotta be tucked away because I never notice them enough to remember them.)

It’s no small wonder that all my friends from California drive electric cars, and all my friends from this area (near my childhood home, so I know lots of people) think EV owners are crazy. [0]

If EV charging stations were visibly everywhere and charged in 5 minutes I could say without a doubt that every one of them would be swayed. So I don’t think they’re being irrational at all.

- [0] It is common to go on long road trips here, since the weather sucks, and people really don’t want to rent a car to do it. Plus a ton of people tow shit. Half my friends have campers and the other half have boats.

Most current EV owners charge at home for daily needs, you really only need charging stations for long distance and owners w/o options where they park (i.e. renters or street parking only) Even with street parking, I see lots of people running cables across the sidewalk (with safety / step covers thankfully)

I drove across the country and accounted for midwest charging. Generally the rocky mountain states are minimal, but I was not without a charger ever 30-40 minutes of travel time through the midwest. Most of them are either in Big Store parking lots or at gas stations like Casey's. You need far fewer of them than gas stations, so we should expect to see fewer of these vehicle refill stations in the future anyway

Yep. Mine is always charged at home, and so I've never needed to use a charging station locally. They're around just by virtue of the area being part of a major metro, but I haven't needed them.

I haven't yet done a cross-country drive but would like to and have plotted out routes with ABRP, and yes, there's more in the midwest states than you'd think. Enough that just about any EV with EPA 250mi range or better can manage a long haul trip without too much trouble (just with a few more bathroom/snack/coffee stops).

yup ABRP was awesome for the trip, nothing out there comes close

Be warned, as you approach and cross the Rockies, there is a lot of uphill and wind. Didn't mention this, but I also did the trip in early March, temps were just above freezing most of the trip, range was terrible. There's a spot in NE, which I'll never forget. Only stop in the middle of a long gap in stations, steepest incline in NE, blew through most a charge in 90 miles. Then the charger is old and very low charge rate (like 5+ year old speeds). The saving graces are (1) public restroom (2) Awesome awesome coffee shop owner who made me a nice brew after hours when I asked him where I might find one as we passed on the sidewalk.

WY was worse on the charger infra, most unreliable and sketchy part of the trip (mostly because of snow in the mountains as temps dropped below freezing, but also the worst charging infra at any point)

Yeah, if/when the trip happens it'll probably be during the warmer months just to keep things simple, and the west → east half of the loop will probably be on I-40 which shouldn't pose too many problems.

Yes, I think people mistakenly believe that if chargers aren't as ubiquitous as gas stations, there must not be enough of them. Range anxiety can still happen on a cross country trip, but not for the vast majority of daily driving. Now, if you live in an apartment building where it's difficult to charge at home, it's a different story. Not so much because of range anxiety, but because the cost of public fast charging rivals and sometimes exceeds the cost of gasoline.

I suspect that the causation runs the other way. They think EV owners, Californians, and anyone who doesn't smell like petroleum is crazy. Therefore they won't buy electric cars and so nobody builds charging stations.

The biggest issue is that people think EV's are gas cars that have electricity instead of gasoline, with the main difference being you have to sit at the "gas pump" for 45 minutes instead of 2.

But the way you daily an EV is totally different than a gas car, and even the way you travel is totally different. People have no concept of EV ownership, so they just go with the gas model that they know. But it is totally incorrect.

> The biggest issue is that people think EV's are gas cars that have electricity instead of gasoline, with the main difference being you have to sit at the "gas pump" for 45 minutes instead of 2.

If you don’t live in a conventional house with access to overnight charging, this is exactly what EV’s are. But we keep talking down to people like this, as if every non-EV owner must just be stupid or something.

We're not speaking down to people like that, we are telling them they are not part of the conversation. You really shouldn't get an EV if you cannot plug in overnight.

"You are completely wrong" and "it's totally different" is not how discussion works. Please, prove your point.

> I can confidently say that every one of them would love EVs if [...]

And I can confidently say I see rural pickup owners rolling coal about weekly, so no, they will not be loving EVs, because it hampers them destroying their surroundings.

As I said, I am only talking about people I know personally. Not everyone is like that, please stop lumping everyone together.

Not everyone who dislikes EV’s is doing so irrationally. Not every one of them is a moronic anti-environmentalist. Most people are just trying to get by and they’re looking at what they think is best for them. Thinking everyone who disagrees with you must be a backwards coal-rolling moron is… not a great approach. You can do better.

> Thinking everyone who disagrees with you must be a backwards coal-rolling moron

Certainly uncharitable, but you should hear what those people say about everyone left of the far right.

I cannot understand “rolling coal” at all. I’d love to know more about the psychology of this and what makes it so attractive that you actually spend money and time to do so.

It's uncomplicated. I have coal rolling enthusiasts in my extended family. They're flag waving 'patriots' who have legitimately drank all the Kool-aid and believe that everyone to their left hates the country and is trying to destroy it in any way possible. And since the left-leaning folks often support green energy and efforts to reduce damage from the impending climate disaster, that hatred manifests as doing whatever they think is the polar opposite of what their left-leaning friends and family would like. It is precisely the same motivation that underlies embracing the 'deplorables' moniker (I think none of them actually read the whole remark) by intentionally acting like an asshole.

It's not issues based at all, they really are playing hard core identity politics and they consider anyone who disagrees with them to be morally contemptible and inhuman.

Rolling coal is cumulatively destructive to the local environment and to the health of the people in the area in the more general sense, and in cases where truck drivers do it to cyclists and hybrid/EV drivers directly and immediately harmful to their health.

If nothing else, rolling coal with the intent of placing somebody within the plume should be considered assault. Diesel fumes/soot is some nasty stuff.

Yeah surely rolling coal and SULEV are exactly the same. Never mind the fact that the other one is literally specced to produce less than 10% of the average emissions, surely they are the same.

Intent matters. Some older diesel vehicles smoke a bit. That is normal. Intentionally detuning the engine to inject massive amounts of fuel in order to induce billowing clouds of black smoke is indefensible. It is especially bad because they are not out there using this capability in innocent fun, it is specifically aimed at passersby who happen to be driving a fuel efficient vehicle or riding a bicycle.

Some of us do enjoy large, overpowered trucks. Like me -- with my Lightning. Faster than a Hellcat (off the line, at least ;-)) and more efficient than a Prius. And waaaaaaay faster than nearly all of the coal-rolling morons. Best part is that I'm not intentionally polluting the air everyone around me is obligated to breathe. Go out, have fun, be civil about it.

It's always a lifted chromed-up truck with oversize exhaust. The one's I'm talking about are 100% intentional, and put money into it. Work pickups look distinctly different.

Work trucks are usually more along the lines of a beat up old 2006 Tacoma or a 2014 F-150 with a stripped down trim and fleet white paint. Totally different species.

> My point being that they’re not irrational, and it’s not EV hatred that is driving it.

That's just your particular bubble. I have met very few anti-EV folk who were not deeply political about it. They don't oppose EVs on rational grounds, they only have the talking points. Matters not at all to them that the talking points were proven false years ago.

With my plugin hybrid I have currently the best of both worlds. The 2x25km commuting is electric, and longer weekend drives are gas. And being in Europe, I am not worried about charging stations for whenever I'll switch to full electric - I enjoy taking gas station breaks. I know it's only one data point, but it's my data point :)

I'm in California, and cost per mile of electricity vs gasoline is pretty similar in a Gen 1 Chevy Volt. I get 35 miles per gallon. That's also how for I can go on 10kwh in the good conditions. If gasoline is below $4.50 a gallon, it's cheaper to just run the volt on gasoline than it is to charge it at home.

That's particular to PG&E, as I recall, not all the utilities in California are so horribly mismanaged or got sued for burning an entire city to the ground. IMO the state should burn PG&E to the ground and replace the entire management structure with people who don't suck.

It’s also the case for me in the Midwest. I drive a Chrysler Pacifica PHEV and it gets about 32 miles per gallon and 32 miles on a charge, and a 16KWH battery. Electricity is $0.22 during peak hours, which equates to a breakeven point of $3.52 per gallon. Non-peak is $0.18, which is a breakeven of $2.88 per gallon. Gas is usually around $3 per gallon so I have to remember to not charge it until after 7pm (this is a setting you can set in the car to make it automatic.)

I’ve since bought solar panels for my house so it makes charging a lot more obvious, but I think it’s actually quite common for people to be paying more for charging than for gas.

Where I live (Ōtepoti Aotearoa) charging at a charging station works out to be about the same price as petrol

I have a plug in hybrid, and close to zero expertise, but I only charge at home, now

I am doubtful that an EV is remotely economical if you cannot charge it at home

That is basically correct in the US as well. Fast charging is about the same cost per mile as gasoline. If your only way to charge the car is with a fast charger, I recommend considering carefully whether it's the right choice for you. The improved driving dynamics may still be worth it, for sure, so it's a very personal choice -- rarely do big purchases like this come down entirely to the bottom line cost.

It matters what the tongue and voice box are doing in the surrounding sounds. The next letter (t) is voiced, and the prior sound is a vowel, so in practice many English speakers will continue to “voice” the c sound between e and d, the “g” is just a voiced “c”, which makes them homonyms in most speakers.

(This post brought to you by YouTube, who keep putting Dr Geoff Lindsey in my recommendation queue, and now I’ve become a part time linguistics enthusiast. Other interesting facts: “chr” and “tr” are also almost entirely homonyms in most speakers. Try saying “trooper” and “chrooper” and see what I mean. In fact my 4 year old, who is recently learning to write, drew a picture of a truck and wrote “chruck” on the paper.)

Plus all of the differences between native speakers.

Canuck here. Color and colour are pronounced differently(mildly), and ant and aunt wildly different. Suite and suit are different pronunciations too.

Yet to some US speakers, those words are the same.

> By giving all of your hosts dns names you don’t have to care about the individual addresses much. If they change just update the dns zone

"just" update the zone? Yikes. I prefer to not take that downtime in the first place. (And I know from experience, I've written hooks for dhcpcd that automatically reconfigure my zone file, firewall rules, rad.conf, etc, if I get a new network prefix! But I don't pretend that this is a workable approach for everyone.)

> The second is to configure a Unique Local Address for each host using SLAAC

Yes, this is the way. Where you used to use RFC1918 addresses, just use ULA. It's simple and fits the mental model you used to have with IPv4. You don't even need NAT, just give both the GUA and ULA addresses to each host, and use the ULA everywhere you want LAN-like semantics.

> as long as [...] (S|D)NAT are not first class citizen in IPV6 Standards and Implementation

Yeah, I mostly agree... IMO, a ULA (equivalent to RFC1918, so 192.168.x.x and so forth) is the only sane way to set up your IPv6 network at home, unless you're one of the wizards who owns their own prefix. Dynamic prefix delegation just breaks too many things when the prefix changes, and I really wish NPTv6 was more supported and ubiquitous, because it solves the problem in the most elegant way IMO.

> there's no mapping of the IPv4 Adresspace into the v6 space

Uh, what? What do you think ::ffff:1.2.3.4 is?

https://datatracker.ietf.org/doc/html/rfc4291#section-2.5.5....

https://datatracker.ietf.org/doc/html/rfc4038#section-4.2

You don't need NPTv6 to use ULA. Just use both ULA and the dynamic prefix from your ISP. The latter is handled automatically by DHCPv6-PD, and if you're only using it for outbound connections then it changing isn't going to break anything.

I'd say this is actually elegant, compared to NPTv6 which is a kludge and will break things (and isn't well-supported anyway).

I definitely do both ULA and GUA at home, but this only really works well to the degree that the OS will prefer the ULA when connecting to things. Like if I want to put hostnames in netgroups, I need reverse DNS to work (which only works if the client is using the ULA address I expect.) In fact the whole idea of reverse lookups working and having expected hostnames show up where you want them to (logs, etc) really depends on not only using ULA for connections, but using the stable address and not the privacy address, which can also cause issues.

For the most part it works today, if I stick to using ULA’s only in my zone file, and configure hosts to prefer the DHCPv6-provided ULA for connections in the ULA subnet, it’s fine. But suddenly if you connect to somehost.local instead of somehost.fqdn, the machine picks a GUA source address and you’re back to being unpredictable.

So although I say I want to use NPTv6 and be ULA-only, I don’t actually do that today, so I’m not super familiar with the downsides to the approach. But it does sound a lot cleaner to me in theory.

> if you're only using it for outbound connections then it changing isn't going to break anything.

A prefix change absolutely does break things in a lot of setups though. It happens something like:

- Your router reboots unexpectedly (no time to rescind the RA)

- Router comes up and gets a new prefix, starts advertising it

- Clients are brain dead and continue using the old prefix when making outbound connections.

I’ve had this happen and both Apple devices and Linux devices (I had no Windows machines) kept using the old prefix until I went around and rebooted them. So connecting to any IPv6 WAN address would fail, and only IPv4 was saving me from my internet being down until I went and manually rebooted everything.

There have since been RFC’s that come up with recommendations for routers to keep a stateful log of old prefixes, so that they can rescind them (advertise a zero TTL) when a new prefix arrives… but afaict none of them actually do this.

huh, i was NOT aware of that. NICE!

now applications (including DNS/NAT) have to support it

i also forgot something (but not against your comment):

* there needs to be guidelines how applications should differentiate between used ipadresses (link, site, global and so on)