ratorx

Can you give some examples of where Imperial works better from first principles and why you think so?

Yep, but the certs.pem file only contains the root and does not know anything about intermediates.

So the only way to fix apps that don’t handle revocation of intermediates (eg the ones that only look at certs.pem) is to revoke the root.

The issue might be Root CA vs intermediates. The root CA that eMudhra rolls upto may be used for other intermediates which were not compromised or revoked. Removing the root CA will mean that these valid certificates no longer work.

This is technically a problem in the CLI - it needs to support revocations properly, not rely on root CA verification only. I’m not sure though whether there’s a standardised mechanism (that Apple supports) for CLIs.

> These steps affect every tool that uses /etc/ssl/cert.pem. Afterward, tools that use only this file might not be able to verify some websites whose certificates chain to these root CA certificates through intermediate CAs that weren't affected.

Being pedantic, the risk/benefit of remote perspectives are theoretical and dependent on the destination network also having multiple independent links. Otherwise any “uplink” ISPs are all SPoF regardless of anything Let’s Encrypt does because the traffic must traverse them to get to the destination.

But remote perspectives are still worth doing, Let’s Encrypt should not be the “less diverse” than the infrastructure of their typical certificate requestor.

I don’t think looseness is always more comfortable. You need looseness in the right areas which is what tailoring provides.

Maybe it’s cold and the extra airiness is working against you.

Maybe the clothes rustle and rub more when you move.

How do (non-elastic) pants stay up, without some minimal measurement for the waist?

I own some more baggy T-shirts and layering with a sweater feels more uncomfortable because the clothes will fold in places you don’t want them to with a sweater on, which is not a problem for more trim fit clothes.

I didn’t like the article because all of the missing defense-in-depth necessary to prevent the intrusion is firmly in the category of traditional cyber security. I could sympathise with an argument about organisational pressures and lack of time/people leading to an insecure sandbox, but that doesn’t seem to be the argument they are making. People have been securing systems that can run untrusted code for a very long time, and whilst mistakes happen, there are not really many novel ones.

The problem of “how to stop the model from attempting to be a hacker” is probably a more interesting and novel one, but it does not need to be solved to prevent the incident.

I think (for now), it is still relevant. A language is an abstraction, and a good abstraction, like a good LLM harness, can be quite valuable.

Let’s say I’m writing some concurrent code with an LLM. I’d probably feel much safer having it write Rust, rather than C. So even in a post-LLM world, languages will continue to evolve as long as abstractions can be improved.

Oops should have RTFA before commenting, I thought the CNAME was only to your servers.

Out of curiosity, how is MPIC relevant? Not that familiar with it, but CNAME would resolve to your server regardless no?

Yeah, I don’t think it assumes zero knowledge, which is what I’d expect from a “first programming language” book.

But it does say it’s targeted more for an “algorithms” course in the introduction.

As a side note, I think a REPL is the best environment for learning to program for absolute beginners. You get instant feedback and can focus entirely on just short snippets of real code, without having to worry about compilers, build systems, and “the OS” as a whole. Especially if it comes with a GUI wrapper, you can skip the terminal entirely.

I think it’s more of a time vs code tradeoff, if done properly.

For example in the Vec case, you could theoretically build an alternative which encodes the “three sections” property internally, and ensures correctness at construction time for the pointers. Not as completely safe as a Vec, but you can still get similar benefits for the “business logic”.

But I agree, just having a custom structure that does not provide a safe wrapper around this would be sacrificing standard guarantees.

The existing, human standards are based on human capabilities. Humans may not have read all the papers that could be cited. However, all these papers are almost certainly in the AI’s training set and it potentially has the ability to retrieve them over the internet into its context. Why shouldn’t we expect the AI to cite more comprehensively?

It seems more like a goal-setting problem, than an actually novel one. But a paper with lots of citations is not as “cutting edge”, so maybe doesn’t make sense for marketing reasons.

Do you have source for IWF funding being by big tech?

Haven’t found anything that breaks their funding down by source and the majority on the UK govt site is from “charitable activities” (https://register-of-charities.charitycommission.gov.uk/en/ch...)

Anycast DNS doesn’t mean what you think it means.

Your DNS traffic to Cloudflare is routed via anycast. If Cloudflare is sending this DNS query (eg to an authoritative DNS server), the IP address it uses for this is not going to be the anycast one. These IPs are geolocatable and Cloudflare even publishes feeds of their approximate location. The response you get will be geolocated based on the IP that Cloudflare is using to send traffic to the authoritative.

Cloudflare explicitly does not use ECS (the edns extension to provide client subnets to authoritatives): https://developers.cloudflare.com/1.1.1.1/faq/#does-1111-sen...

Putting aside the WASM sandboxing (I’m not familiar enough with it to understand how sandboxing works) there’s a DoS vector at least. Even regexes have had many DoS issues, and I can’t imagine WASM being easier to sandbox for DoS risk.

There’s a big difference in the expected use of a file. If the file is attacker provided, and the fallback path is being used, the attacker can embed whatever WASM payload they want into the file since the file will be “opened” by “execute this offset into the file”.

Compare that to JSON. The parser NEVER needs to execute arbitrary instructions. Parser might have bugs, but it avoids a whole class of issues.

Providing access to the data is easy. It is just an MCP or equivalent, and coding such CRUD is cheap now.

Applying the actions is unsolved. Unless you YOLO the LLMs, taking stateful actions automatically requires a lot of protective infrastructure, solid testing infra etc.

It’s all just more code, but a “create me a shopping website” LLM is likely not going to be doing the infrastructure level thinking required to handle it for now.

If your goals were fame, then yes. But you can still pursue excellence even if there is an alternative “easy” path.

The equivalent is something like hand tool woodworking - it’s still a thing despite the advent of machines, but more of a niche. You can still aim to become excellent, but maybe you won’t be famous.

UK has never had net neutrality, there are many limited data phone plans that include unlimited music/video etc

I think I read somewhere (but could be wrong) that it was because they didn’t want to own any “authentication” services. Their infrastructure was zero trust (as in they don’t hold any passwords or private keys), just a discovery server for different devices.

Making it explicit wouldn’t be particularly problematic no? Option<&Spouse> in Rust terms. Or for this specific case, a GADT (Single | Married<&Spouse>)?

It could even use a special “NULL” address. Just don’t pollute every type with it.

Pure CS is not necessarily equivalent to pure maths. For the “science” bit of CS, you do need to do the equivalent of experiments (for more applied topics).

For example, a physics degree is expected to have experiments. You are not required, expected (and possibly do not want) to know the tools required to professionally build a bridge because you did courses on mechanics. But you might do an experiment on the structural integrity and properties of small structures.

Whether this is a good split is an entirely different question.

Changing the links and doing nothing else would be a pretty dumb MITM. You could do a more complex variant which is not so easy to spot (targeting specific networks, injecting malware whilst modifying the checksum)

The key property of SSL that is useful for tamper resistance is that it’s hard to do silently. A random ASN doing a hijack will cause an observable BGP event and theoretically preventable via RPKI. If your ISP or similar does it, you can still detect it with CT logs.

Even the issuance is a little better, because LE will test from multiple vantage points. This doesn’t protect against an ISP interception, but it’s better than no protection.

I find it interesting that they bought Astro (https://blog.cloudflare.com/astro-joins-cloudflare/), which from my definitely-not-a-frontend-person perspective seems to tackle a similar problem to Next. A month ago.

If it is so cheap to make something that they recommend using (rather than a proof of concept), why buy Astro (presumably it was more expensive than the token cost of this clone?).

One conclusion is that, at the organisational level, it still makes sense to hire the “vision” behind the framework, rather than just clone it. Alternatively, maybe AI has improved that much in 1 month!

You can make it so employees don’t have ambient access to data, and require multi-party approval for all actions that require user data. Giving away a user password should be treated as a routine risk.

I’m not saying that’s how it actually works, and this process doesn’t have warts, but the ideal of individual employees not having direct access is not novel.

It doesn’t have to be a compile time constant. An alternative is to prove that when you are calling the function the index is always less than the size of the vector (a dynamic constraint). You may be able to assert this by having a separate function on the vector that returns a constrained value (eg. n < v.len()).

A global train outage would be quite a spectacle, is that even possible?

Management not having to listen to engineers is the structural problem. How do managers know which concerns that engineers bring up are actually relevant? How do engineers know which concerns have real world consequences (without having a incredibly high burden of proof)?

Having regulation, or standardisation is a step toward producing a common language to express these problems and have them be taken seriously.

Leadership gets a strong signal - ignoring engineers surfacing regulated issues has large costs. Company might be sued and executives are criminally liable (if discovered to have known about the violation).

Engineering gets the authority and liability to sign off on things - the equivalent of “chartership” in regular fields with the same penalties. This gives them a strong personal reason to surface things.

It’s possible that this is harder for software engineering in its entirety, but there is definitely low hanging fruit (password storage and security etc).

I think it depends on the scope and level of solution I accept as “good”. I agree that often the thinking for the “next step” is too easy architecturally. But I still enjoy thinking about the global optimum or a “perfect system”, even it’s not immediately feasible, and can spend large amounts of time on this.

And then also there’s all the non-systems stuff - what is actually feasible, what’s most valuable etc. Less “fun”, but still lots of potential for thinking.

I guess my main point is there is still lots to think about even post-LLM, but the real challenge is making it as “fun” or as easily useful as it was pre-LLM.

I think local code architecture was a very easy domain for “optimality” that is actually tractable and the joy that comes with it, and LLMs are harmful to that, but I don’t think there’s nothing to replace it with.

I guess this is also a matter of organisational policy and how much power individual teams/organisational units have.

I would imagine mature organisations without serious short/medium term existential risk due to product features may build some push back mechanisms to defend against the inherent cost of maintaining existing business (ie prioritising tech debt to avoid outages etc).

In general, it is a probably a mix of the two - even if there is a mandate from up high, things are typically arranged so that it can only occupy X% of a team’s capacity in normal operation etc, with at least some amount “protected” for things the team thinks are important. Of course, this is not the case everywhere and a specific demand might require “all hands on deck”, but to me that seems like a short-sighted decision without an extremely good reason.