That's because Clang has a mountain of sugar to sweeten the developer experience. Between invoking the program from shell and printing the version, it's executing thousands of lines that set up error handlers, initialize all of the hundreds/thousands of possible args, do basic driver setup, probe different parts of the filesystem, and so on.
If Google's switch promises to disable the use of all "AI" techniques down to the implementation level, that would imply it also disables spam filtering in gmail. I haven't seen that reported and it isn't in the original announcement post, so it probably isn't how they're thinking.
I would be very surprised if LLMs were adequate for doing archaeological lit review. My experience is that you often find sites documented halfway through a stack of undigitized records in a physical archive somewhere, and it's usually a verbal description that's difficult to localize to a specific location. Heck, I've even contributed to the problem by finding new sites as we're packing up camp, when I didn't have time to properly document them outside field notes.
my brother in law did a lot of research in paleontology as a grad student. He said a lot of fossil discoveries take place in basements or storage rooms. Unknown things that were collected but never documented.
Datalog (i.e. prolog without function symbols) still has a lot of practical applications for as a relational language and for logic programming. The polonius borrowchecker for rust originated as a datalog program, for example.
They're not different in any fundamental way, except for function symbols. Cut is obviously irrelevant, so let's talk about the semantics and execution. There's only two important aspects there: infinite recursion and top down/bottom up. Infinite recursion is pretty easy to deal with. We just use tabling (and tabled prolog is still a prolog). Once we've done that, top down and bottom up are equivalent in the absence of function symbols. Then all we have to sort out is the negation semantics. Datalogs typically require restricted negation, so they get 2-valued negation. Prologs don't usually restrict negation, so they usually assume 3-valued WFS. Neither one of these is required though. WFS datalog exists, as do prologs with restricted negation.
Unification is a bit different, but not that important here. It's more important to note that a datalog program will compute the same results in prolog and thus, my previous statement that datalog is essentially prolog without function symbols.
Looters will absolutely destroy a site to find valuables that may or may not even be there. Ai-khanoum is a classic example, with the vast majority of the site having been filled with pits on the off chance there were valuables. Nowadays they use bulldozers and excavators to loot at an industrial scale. It's pretty difficult to do archaeology on a pile of sifted dirt.
This feels a decade late. I don't really see the value prop for customers in 2026 (or realistically, 2030+). Let's say I'm willing to rewrite all my software to a language that compiles down into interaction nets. Why don't I use HVM2 instead and buy GPUs from Nvidia? If I just want to use an exotic architecture, AMD is shipping dataflow processors at scale today, and they'll license related IP. And if I just want a massively parallel architecture to run my existing ML code on, I could just rent google cloud TPUs instead.
Who's the customer that wants an exotic architecture on an old process node running a custom programming language half a decade from now?
because rewriting your software is easy now, there are plenty of examples of Language Models looking at the architecture in front of them and recompiling themselves, their inference engine and harnes for faster inference on it or whatever their need may be
others have made entire OS's or gone back in time reinventing of removal of the OS and just running a single application with its own I/O stack per application
this is a guess that exotic architecture on old nodes means that the hardware can come out now at a lower price - or at least fulfill market needs at the price the market can bare without the delay - all because using exotic architecture is easy now, unstable opaque drivers are no longer an issue, and there is just a need to fulfill compute demands efficiently
Language models are efficient because they are trained with millions of examples that don't exist for newer architectures, and they are still mediocre to bad on stuff where they hadn't been trained
If software stacks and compatibility didn't matter, AMD GPUs would be in nearly the same demand as Nvidia's. But we know they aren't and that's despite the fact that porting between GPUs is much easier than what we're talking about here.
this technology to make it easy is only a couple quarters old
the antiquated understanding of “CUDA being sooo stable and AMD drivers not” just takes a while to be revised
hyperscalers are buying AMD GPUs
AMD’s stock has tripled this year based on investment into their driver stack getting better and others realizing they don’t need to wait for AMD to fix their driver stack in order to leverage the compute
A point that stopped being relevant as soon as it was proven possible, I don't know what kind of validation you need to see about exotic hardware being easier to develop for, but being slower to pick that up doesn't change the viability of the exotic hardware built on more available nodes now that there are examples of it being easier to run things on
I don't know where you could work that this policy isn't at least nominally in force, outside very small companies. Common certifications like ISO:27001 and SOC2 basically require systems to be running only "authorized" software. It usually isn't actively enforced on developers, but IME virtually all companies have it as an official policy alongside a rarely used process to get approval for unauthorized applications.
Don't know what you're doing, but it's good enough for the reasonably sized tasks I give LLMs. It's also a much better price. My browser stopped rendering the UI frame recently because I messed up permissions on /usr/share/icons. Took DS 4.1 flash about 3 minutes and a few cents to diagnose it. Same harness with opus 5 took twice as long and several dollars.
Moreover, the months I've used anthropic/OAI exclusively my API token costs are in the hundreds (or sometimes thousands). With DS flash as the first agent I reach for, I rarely exceed $40/mo.
I figured it was assumed I was including that? It's been fine as long as I break tasks up into reasonably sized chunks, which is fine because I think you should only be generating as much as you can comfortably review at a time.
I'm a heavy DS v4.1 user and been satisfied with its performance. I use it just like you described -- "... break tasks up into reasonably sized chunks...".
Every once in a while I come across comments that just hates DS. I wonder if it's a matter of how people use it or the prompt. I admit pre-v4.1 was pretty bad, hallucinating too much and not getting the abstractions right so I had to correct it often.
I also have $200 Claude subscription (now called "Ultra"). Fable gets abstractions right on highly complex problems (e.g. defining protocol that spans multiple services) that DS simply cannot do. However, DS is comparable to Opus 5.5, although less "intelligent", but DS excels at speed which is what I prefer than waiting minutes on Opus 5.5.
Painting actually introduces some scaling issues in manufacturing. Any painting step prevents you from haphazardly mixing all the parts produced on a line. Most vehicles will have multiple lines that eventually converge across multiple suppliers. Painting forces you to either coordinate the production of each individual component so that they're all painted the correct color during final assembly (which becomes difficult if any parts fail QA), or buffer painted parts in a giant warehouse. There's other advantages to the warehouse, but I wouldn't call it cheap.
Very common to use a mix of both strategies in practice.
reply