And the correct policy regarding “AI”, too. Viva Zig

I knew nothing about, and had no opinion on, daycare facilities in Minnesota run by Somali immigrants, before Trump-supporting media entities decided to make the topic an astroturf issue. On the other hand, I had plenty of experience with people whose worldviews had been severely warped by such coordinated media campaigns. Mr. McKenzie should take some time to reflect on this.

Btw, been appreciating your posts lately!

from bsky photos looks like entire gathering was 30 people. t h i r t y p e o p l e i might have counted some reporter or someone passing by randomly by accident

annieposting

i think that business logic goes against your first point. spatially: if you have source of cheap energy and want to make money out of it, instead of making little money (by making fuel) why not make more money? (by setting there energy intensive manufacture) this seems to be current meta, with places like iceland and norway making aluminum and nitrogen fertilizers respectively. this can continue in other places and maybe extended to some other industries.

temporally (because there are also sunny and windy days when regular people won’t consume all energy): this scheme requires cheap electricity, which is needed for cheap hydrogen. this requires massive renewables buildout, which means electricity is cheap for regular people, which means that every gas stove/heater and car will get replaced with electric ones, both residential and maybe perhaps faster for industrial users (more available loans). this means you have to reinforce transmission grid anyway. this also means cheap hydrogen, and because main input to its production is electricity, it makes more sense to use electricity when it’s cheap. this means it’s naturally suited to suck up all excess generation (both daily and seasonal), and also if electricity production is seasonal then so should be price of hydrogen. if price of electricity or hydrogen varies, then some industries can suck it up at greater rates when it’s cheap. i’m thinking here of aluminum smelting (electricity input, daily variation, already done), or ammonia synthesis, or direct reduced iron smelting. i bet there’s more. the point is, maybe you get to avoid storing hydrogen to some degree, because you can effectively store energy in finished or semi-finished goods. you can, for example, make some direct reduced iron and just store it when hydrogen is available, and then smelt it into steel in arc furnace when it’s not. fertilizers are already sold in annual cycle and stored long term, and anyway ammonia is much easier to store than hydrogen. how it plays out will depend on energy/hydrogen costs vs storage costs vs capex for overcapacity costs. all together, i think this means that because of large amount of generation needed, you don’t actually need to store energy this way at all, because when generation is low then electrolyzers turn off, and something will work at all times, probably. when you’re able to do that, you won’t need to

in terms of scale, first your lunch is eaten by EVs of various shapes, then by use of hydrogen for transportation (rocketry fits there), then you have to compete with biofuels (jet engine will take anything that burns without ash and can be pumped). then some of methanol will be used for fuel first, because it just works in engines and fuel cells, and it’s a step before hydrocarbon synthesis. only then synthetic petroleum makes sense, this basically leaves some aviation (that won’t use methanol) and military uses

i don’t know if they started it. what i suspect as their contribution is bold claim that electrofuels might be cheaper than regular petrol in the glorious future, while currently they’re much more expensive. (30x?) strict prerequisite for their competitiveness is cheap electricity, but at this point they’re not needed. there was also Porsche owned wind power to methanol plant, and while methanol works as petrol replacement, all the plastics in contact with it must be resistant which is not a given. i guess the main value of it for them is propaganda, they’re not ready for EV manufacture

the point is, as always, to continue doing business as usual (in this case, by inhibiting BEV adoption). that fuel is carbon-neutral but also extraordinarily wasteful. trump’s deal is something called “clean coal”, which isn’t (it suggests carbon capture, but it’s not a thing, they marketed normal emissions control like we have in europe as some unusually green innovation). i think he was also captured by gulf monarchies for the one hour when their representative talked to him

e: wait it still makes smog so checks out

It was basically the only “empirical” (scare quotes well earned) data they actually used in their “model”, even then, they decided exponential improvement wasn’t good enough, they plugged it into a hyper-exponential model that went to infinity at just a few years regardless of the inputs.

They absolutely are. I am just giving them a tiny bit of credit for at least attempting academic research on LLM performance. But only a tiny bit, as they blog post I link discusses, their methodology is really sloppy and not to the level of most academic research and wouldn’t get through peer review of most decent journals.

AAaas: Arthur Anderson as a Service.

I knew the Anthropic blog post was bullshit but every single time the reality is 10x worse that I anticipated.

@sailor_sega_saturn @dovel the nice thing about rogue chatbots is that there’s no difference between that and a non-rogue chatbot

The first issue filed is called “Hello world does not compile” so you can tell it’s off to a good start.

The comments are a hoot, at least.

did you have a point you wanted to make?

This could be regarded as a neat fun hack, if it wasn’t built by appropriating the entire world of open source software

This shouldn’t be left merely implied, the autoplag trained on GCC, clang, and every single poor undergrad who had to slap together a working C compiler for their compilers course and uploaded it to github, and “learnt” fuckall

I wonder what actual experts in compilers think of this.

Anthropic doesn’t pay me and I’m not going to look over their pile of garbage for free, but just looking at the structure and READMEs it looks like a reasonable submission for an advanced student in a compiler’s course: lowering to IR, SSA representation, dominators, phi elimination, some passes like strength reduction. The register allocator is very bad though, I’d expect at least something based on colouring.

The READMEs are also really annoying to read. They are overlong and they don’t really explain what is going on in the module. There’s no high-level overview of the architecture of the compiler. A lot of it is just redundant. Like, what is this:

Ye dude, of course it doesn’t depend on the IR, because this is before IR is constructed. Are you just pretending to know how a compiler works? Wait, right, you are, you’re a bot. The last sentence is also hilarious, my brother in christ, what, why is this in the README.

Now this evaluation only makes sense if the compiler actually works - which it doesn’t. Looking at the filed issues there are glaring disqualifying problems (#177, #172, #171, #167, etc. etc. etc.). Like, those are not “oops, forgot something”, those are “the code responsible for this is broken”. Some of them look truly baffling, like how do you manage to get so many issues of the type “silently does something unexpected on error” when the code is IN RUST, which is explicitly designed to make those errors as hard as possible? Like I’m sorry, but the ones below? These are just “you did not even attempt to fulfill the assignment”.

It’s also not tested, it has no integration tests (even though the README says it does), which is plain unacceptable. And the unit tests that are there fail so lol, lmao.

It’s worse than existing industry compilers and it doesn’t offer anything interesting in terms of the implementation. If you’re introducing your own IR and passes you have to have a good enough reason to not just target LLVM. Cranelift is… not great, but they at least have interesting design choices and offer quick unoptimized compilation. This? The only reason you’d write this is you were indeed a student learning compilers, in which case it’d be a very good experience. You’d probably learn why testing is important for the rest of your life at least.

I wonder why this blog post was brazen enough to talk about these problems. Perhaps by throwing in a little humility, they can make the hype pill that much easier to swallow.

I feel this is an artefact of the near complete collapse of mainstream journalism, combined with modern tech business practises that are about securing investment and cashing out, and every other concern is secondary or even entirely absent. It’s all just selling vibes.

People only ever report the hype, the investors see everyone else following the hype and panic that they might be left out and bury you in cash. When it all turns sour and people ask pointed questions about the exact nature of the magic beans you were promising to grow, you can just point at the blog post that no-one read (or at least, only poor people read, and they’re barely people if you think about it) and point out that you never hid anything.

I only sampled some of the docs and interesting-sounding modules. I did not carefully read anything.

First, the user-facing structure. The compiler is far too configurable; it has lots of options that surely haven’t been tested in combination. The idea of a pipeline is enticing but it’s not actually user-programmable. File headers are guessed using a combination of magic numbers and file extensions. The dog is wagged in the design decisions, which might be fair; anybody writing a new C compiler has to contend with old C code.

Next, I cannot state enough how generated the internals are. Every hunk of code tastes bland; even when it does things correctly and in a way which resembles a healthy style, the intent seems to be lacking. At best, I might say that the intent is cargo-culted from existing code without a deeper theory; more on that in a moment. Consider these two hunks. The first is generated code from my fork of META II:

while i < len(self.s) and self.clsWhitespace(ord(self.s[i])): i += 1

And the second is generated code from their C compiler:

while self.pos < self.input.len() && self.input[self.pos].is_ascii_whitespace() {
    self.pos += 1;
}

In general, the lexer looks generated, but in all seriousness, lexers might be too simple to fuck up relative to our collective understanding of what they do. There’s also a lot of code which is block-copied from one place to another within a single file, in lists of options or lists of identifiers or lists of operators, and Transformers are known to be good at that sort of copying.

The backend’s layering is really bad. There’s too much optimization during lowering and assembly. Additionally, there’s not enough optimization in the high-level IR. The result is enormous amounts of spaghetti. There’s a standard algorithm for new backends, NOLTIS, which is based on building mosaics from a collection of low-level tiles; there’s no indication that the assembler uses it.

The biggest issue is that the codebase is big. The second-biggest issue is that it doesn’t have a Naur-style theory underlying it. A Naur theory is how humans conceptualize the codebase. We care about not only what it does but why it does. The docs are reasonably-accurate descriptions of what’s in each Rust module, as if they were documents to summarize, but struggle to show why certain algorithms were chosen.

Choice sneer, credit to the late Jessica Walter for the intended reading: It’s one topological sort, implemented here. What could it cost? Ten lines?

I do not believe that this demonstrates anything other than they kept making the AI brute force random shit until it happened to pass all the test cases.

That’s the secret: any generative tool which adapts to feedback can do that. Previously, on Lobsters, I linked to a 2006/2007 paper which I’ve used for generating code; it directly uses a random number generator to make programs and also disassembles programs into gene-like snippets which can be recombined with a genetic algorithm. The LLM is a distraction and people only prefer it for the ELIZA Effect; they want that explanation and Naur-style theorizing.

I wonder if this is going to hold out long enough to get some obnoxious AI-first language created that is designed to have as obnoxiously picky of a compiler as it can in order to try and turn runtime errors that the model can’t cope with into compile failures which it can silently retry until they’re ‘fixed’

Newspeak never became a real problem

never heard of wokeness, i see

The thing is, the world in 1984 is feminist! as imagined by a bloke who hates feminism. Sex for pleasure is outlawed, makeup and dresses are banned, women look and act like men (and indeed are worse than men) instead of following their womanly nature. It’s a feminist dystopia!

I mean, if god damned Asimov thinks your book is misogynistic, you know you’ve fucked up!

Orwell had Julia working in the novel factory - where machines spliced together romance trash pablum for light entertainment. So he accurately prophesied LLMs.

I don’t know if there has been any on-the-ground journalism about our dear friends except possibly the Zizian murders and some of the investigations into misogyny and sexual abuse. What reporter without a bankroll from EA has money to spend a few months in the Bay Area making friends with introverted bloggers?

One of the mentors at Inkhaven will be Jesse Singal who has said good things about pedophiles and KiwiFarms and is worried about so many young people identifying as trans. Does he have prior connections to eugenics or race pseudoscience? Pinkerite says he is Bluesky buddies with Razib Khan (one of the names RationalWiki can no longer mention).