ponkpanda

There doesn't need to be an explicit conspiracy. It only needs all the US frontier labs to be facing the same economic pressure (logarithmic improvement / $). It's a pretty obvious strategy - it's not like the large labs can magic up huge volumes of extra compute as demand comes online; there are almost certainly tweaking model performance to occupy the compute available and margin/cash burn targets.

That was one of the main points of the movie 'A Beautiful Mind' - that actors can coordinate without any explicit communication.

Not original commenter but the reality with zig is a little in between being simply convention vs. a language requirement. Is it a language requirement? no.

However, there's no global allocator in zig. You simply cannot call the language's equivalent of malloc() because it doesn't exist - at least not as a global symbol. That leaves you with three choices: 1) Define a global allocator; is a valid choice and would make a zig program more like C, C++ or Rust in terms of not having to think about scope-level allocation patterns 2) Pass an allocator into that scope (this is the community convention) 3) Create/instantiate an allocator itself inside that scope

(1) would be valid, though may not be idiomatic; global allocator like malloc becomes an opt-in

(2) Expensive and inefficient for most scopes, though not all.

(3) cheap, idiomatic but potential for noise/boilerplate

Obviously not.

There is no tool that can do what this tool is claiming to do. At best, it can make reasonable inferences based on data it has available using whatever model it is using (assuming it's not a fraud) but key parts of the world are irreducibly complex. They cannot be predicted and that is without the problem of reflexivity which you correctly highlight as an issue - even for a notionally perfect prediction engine.

Given enough compute and data, you could model every human on the planet at subatomic granularity and you'd a) be chaotically sensitive to initial conditions; b) be subject to random quantum fluctuations.

Nice enough frontend though.

Very difficult to sense check that substack post without access to the TLB credit agreement.

There are likely management service agreements from xAI proper -> SPV to cover precisely what the author talks about. Clearly, xAI could play games but without seeing the docs (which are not public), it's very difficult.

This article's basic point is right though. On the other hand, the LTV of this deal was approx 50% debt-financed (not too high; very much depends on the "V"). At 12.5%, it's not as if its being priced as a high quality asset.

Overall, substack post was too bearish. The wider point is that there's a lot of froth tied to what has now become systemically opaque - namely the circular deal flow that every hyperscaler, nvidia, neoclouds and friends are now engaged in. When the proverbial hits the fan, that stuff will be difficult to price and find few willing buyers with the competence to underwrite.

The systemic issues are the bigger concern than one specific deal imo.

That's trailing and mighty impressive. However when you decided to quote Q1 earnings. $62bn of Q1 earnings ends up at only $10bn. The LTM numbers are less relevant.

AI-related capex is one hell of a drug.

Yes, earnings are essentially hype/BS.

Focus on cash/cashflow.

Perhaps requiring webauthn credential for any post/comment with a whitelist of permitted webauthn hardware devices which must have touch/interaction enabled.

I'd have to read the FIDO specs, however the only place I've seen webauthn hardware pinning in the wild is with Azure AD/Entra which is ostensibly based on token GUID. If this is the only enforcement mechanism available, it's spoofable.

Damn!

> And that if there is a clear and direct contradiction between the Supreme Court and the president, the former trumps (no pun intended) the latter.

The extent to which members of the executive branch adhere to their oaths is not written down. Ofc the oath is written and its power may partly derive from its written nature (clear; predictable; well publicised etc) but there is a lot more than its written nature that might cause a general to refuse to follow a Presidential order to arrest all people suspected of voting for their opponent.

> The lack of arbitrary rule... is emphatically not [a defining feature of both]

I guess it depends on whether you (or most reasonable people) would call countries like Russia a 'constitutional republic'. Of course there are plenty of dysfunctional and dictatorial countries which superficially describe themselves as XYZ but it lacks substance.

While there may be a textbook answer, I strongly suspect it is debatable within the field and comes down (like so many things) to how you define your terms. Do you define 'constitutional' as attaching more to the codified and written nature of any rules or whether it is more to do with predictable and enforceable rules limiting arbitrary government. My view is that it attaches more the latter.

If you go into the etymology of the term, I don't think codification is baked in - that you can find a large number of books discussing the English or UK constitution (using that term) is testament to the fact that it's not just some niche view. I do suspect the influence of US popular culture (e.g. Hollywood) has biased the term towards the US' arrangement vs. the alternatives.

Tighter (read better) integration with VSCode and Github than what you could get running claude code on the side.

Your question does raise a valid point - Github Copilot's value proposition is fairly limited in my opinion. Not to say worthless but limited and clearly varies depending on how Githubbey your dev workflows are.

> It's a fundamentally substantive difference in the two structures

Yes, it is a substantive difference but it does not follow that this difference provides the 'constitution' property.

> one of these has an indisputable source of truth... the foundations are not the same

They are so similar as to be almost the same and if an 'indisputable source of truth' exists anywhere, it is not in the written documents or their structure but unwritten norms and rituals sit beneath both.

What stops a President from simply choosing to ignore a Supreme Court ruling and what prevents the King from returning to personal rule?

The lack of arbitrary rule is a defining feature of both and relies on something that emerged rather than something imposed from without by written words.

>It doesn't look like a duck

Some special amendment procedure is not the only or even defining feature of constitutional law. There is non-constitutional law that has this property and there is constitutional law that does not.

Repo hosting is the kind of thing that ought to be distributed/federated.

The underlying protocol (git) already has the cryptographic primitives that decouples trust in the commit tree (GPG or SSH signing) with trust in the storage service (i.e. github/codeberg/whatever).

All you need to house centrally is some SSH and/or gpg key server and some means of managing namespaces which would benefit from federation as well.

You'd get the benefits of de-centralisation - no over-reliance on actors like MS or cloudflare. I suppose if enough people fan out to gitlab, bitbucket, self hosting, codeberg, you end up with something that organically approximates a formally decentralised git repo system.

fixed rates embed best current (market) estimate of path of future floating rates.

The fixed rate payer/receiver is protected but they typically pay for that protection.

Except FPGA chips/boards aren't free from malware either: https://www.iacr.org/archive/ches2012/74280019/74280019.pdf

Nor will you be immune from AMD Vitis/Vivado sideloading crap into the bitstream.

Sadly, you have to fab your own chips using sovereign facilities if you want security. Individuals simply cannot access genuinely high assurance product and there's no major government in the world with the slightest interest in changing their stance on this policy. There are simply too many governments long on SIGINT to go down such a route.

You're not adding the +nightly switch and the required flags to compiled std. Just setting panic=abort in cargo.toml isn't enough.

One can get binaries pretty damn small (low-mid tens of kilobytes for a basic cli program doing something like hashing of a file).

Problem I've found with manually compiling std (which has ancillary benefits of being able to compile to a specific uarch) is it can break the compilation process when bringing in third-party deps. The config.toml (stored in $PROJECT_ROOT/.cargo) overrides cargo's behaviour for all dependencies as well - which may break those compilations.

Tbh, it's one reason I don't particularly rate the rustc+cargo toolchain - but for most people writing regular applications: just being able to do ```cargo build -r``` and not care about binary size, uarch optimization or custom llvm/rustc optimizations (PGO etc), most won't care.

I habe a few lying about. My fun project is to implement a usb CCID (smartcard) device to create a pkcs11comiant HSM from scratch (FYI the ESA dis this on MAX10s for one of their satellites or probes). Assuming you dev environment is not compromised. USB 1.1 and USB 2.0 have free/OSS IP cores.

Personally, I do not see point in implementing a softcore Cpu unless design absolutely requires it.

Separately, FPGA makers has shifted focus a bit from facilitating hardware design to providing coprocessors/accel cards.

I can't speak for you personally but I think its simpler than that - Java is unfashionable because it's ubiquitous. It's the Ford of programming languages. Not every problem needs a Ferrari - indeed its bad for the organization to be building everything with Ferraris.

I do wonder whether someone has looked at the game theory of all this - how devs (and employers) pick langs and frameworks as a strategy to maximize their respective positions under competition.

It's also used by some pretty huge and non-middling corporations. I don't know why I'm defending it - I find Java a too verbose and don't use it personally- however if I were employing and needed to build anything that wasn't a game, an OS, embedded etc. I'd go with Java for large chunks of it because I can find good people to build, maintain and improve it - for years to come. I'd probably throw in some Go and perhaps C++ for compute bottlenecks.

Also - don’t discount all the new features JVM and Java have added or are adding: virtual threads, structured concurrency, FFI, class primitives, SIMD. JVM is highly compute performant given its GCd.

I’m not even a Java dev but I respect its place in the ecosystem and find evangelism quite irritating and normally, has some corporate backer(s).

Firstly, as long as there is still a broad need for software engineers - there is a future for Java. Together with dotnet - it has huge market share of enterprise middleware, lots of web backends.

Its future in some segments is extremely safe while in others it has more competition. Whether it has a future for you depends on the kind of fields you want to work in.

In any case - once you learn a couple of languages (especially from a couple of different paradigms or layers of abstraction) - it becomes easier to pick up new languages.