pixelesque

Mission control is not supposed to print the name of the control/window visible to the user though, surely, which is what the article is talking about.

Follow up bit adds more context:

"Carter Leffer only directed GPT–6 Astra to see if it could break any of the unbroken Enigma messages published on the Crypto Cellar Research web page."

> Rust had like 3 allocating types total.

Seriously? Categories of types maybe, but literal types it's more than that.

Even closures allocate if they need to capture their environment.

Some of this surely stems from Rust not originally supporting allocators at the standard API level until later. You can manually create your own, and create/modify your own containers, but until GlobalAlloc came around it sometimes involved doing things like passing through Arc<Allocator> through to things if you wanted multiple things to use a single one.

Sometimes using a global one isn't the best thing (it's often a good idea to specialise base on allocation size, reuse and lifetimes), but I've used them quite a bit in C++ over the past 16 years doing HPC for graphics, rendering and simulation, so calling them only useful for "toy problems" likely shows you just haven't found a need for them in what you've been doing.

> and is barely even supported at all in the standard library.

What does that mean?

But that was has nothing to do with the incident in the Middle East, it's a more general one.

Back in 2023 when the previous issue happened, it didn't actually "crash", it detected what it perceived as an inconsistency (which was due to invalid waypoint logic for waypoint codes in multiple countries) and put itself in "maintenance mode".

It was blatantly obvious Iran would do that if attacked. This has been known for decades.

The only thing previously stopping them holding the strait "hostage" was the threat from the US of doing an "Operation Praying Mantis v2" and destroying their navy and attacking them.

And when the US did that anyway, they had nothing to lose.

They've got the IDL files as well for the APIs, but you can't use them without their libs:

> The data model is defined using a custom format .rmodel based on the ddkit > framework (rmodel-api).

> Please note that the .rmodel data model is for documentation purposes only and > does not reflect the actual interface used on the communication layer. Network > communication is only possible through the provided libraries.

DDS sort of took over from CORBA in the military domain, and the statically-typed IDL was still quite similar to CORBA last time I looked at it, so that takes you back to around 1998 when CORBA was being used a lot in the military in my experience.

What other protocols and systems were you thinking of that can provide (depending on broker) hard-realtime delivery that can be used in safety-critical systems?

Its giraffe / grandfather clock one is pretty bad... (two necks? wearing a suit?)

Weird as well, it's clearly pulled out some 1884 patent on clock designs, and a quick ddg/google doesn't show it as anything to do with grandfather clocks.

Sounds like it might not have been fully fixed after 2023 then.

Other than the fact it's mathematicians signing it, why is mathematics special in this regard: surely this generally applies to a lot of different industries and sectors of research / academia?

Mathematics is being used as a benchmark because there are some high-profile awards in this area I guess, and possibly because 2/3 years ago LLMs were pretty atrocious at it so the level of improvement has been significant.

It's wider though, and surely going to weigh more...

I wish they'd alternate each year or something between iPhone / iPhone mini if they're struggling for sales.

The body shapes and cases seem to change each year anyway now, so it's not like people get to re-use cases that much anyway...

Same, got battery replaced 5 months ago.

Wish I'd got the biggest storage size when I bought it though :(

Where are you seeing them having an issue with the larger (Pro) model?

The announcement specifically says 4.1 Pro will be released in the future.

I've had this for a few months trying to log into https://console.vultr.com/...

It does it for me if I use a VPN (Mullvad) - if I don't use a VPN then I haven't noticed I get them.

But yeah, very annoying.

Block 4's still an absolute shit-show in terms of management and progress...

It's not even that though if we are going to be nit-picky: it's a coordinate (not quite sure what the pixel extent would subtend to solid-angle-wise, but...) based off the coordinate frame of where the space telescope will be when the image is taken in spherical coordinates.

That's not "exclusive". The "pixel" is the exclusive bit.

Title doesn't match submission: it's not really a "coordinate" in 3D space, you "adopt" a pixel from the early imagery taken by the Nancy Grace Roman Space Telescope.

Maybe that can be reverse-projected back to something like a coordinate (if you know the distance), but it's not the same thing.

Hermes' Github has 12K open issues and 25K pull requests.

I assume they're using agents to sort through all the slop?

Looks okay to me.

The first column has both the Hy4 and Hy3 scores overlaid on one another (Hy4 is darker blue and the taller one), with both scores written below the top of the respective bar - maybe you're seeing that?

Too sloppy, didn't read.

Yes, but come on...

That's due to the need for ultra-fast memory and inference compute, it's not the same thing.

Is that the same as storing state in a DB and on a filesystem though?

Like, I get there are other features they have, and things like history, but I don't think it's quite the same.

Isn't most (not all) of the LLM's resources stateless compute to a large degree?

(Not defending MS at all though, I currently can't create a PR on the required branch for my work (won't load the branch list) and their availability is really bad recently).

As someone typing this on an M1 Mac with 32 GB of RAM who tried using 3.8 27B (Q4_K_M) yesterday in both LM Studio and llama.cpp, I wouldn't call it particularly usable in terms of token speed. (and that was with `--spec-type draft-mtp` for llama.cpp).

If you want to leave it running with the fans going crazy for 40 mins or overnight or something, fair enough, but otherwise it doesn't seem worth it to me. It's certainly not "interactive", even taking into account the over-thinking it does by default.

The 3.6 (maybe they'll release a 3.8?) MoE model is much more usable (but obviously not as good) on this machine spec.

> really crazy optimizations that they are doing

Any examples of this?

I heard on the "China Talk" podcast talking about China and exports of open weight models they often do this because under Chinese law, they have to submit each new public model release to a government department for "validation", and if they keep the version the same, they don't need to re-submit it.

I've found Pro to be a lot better per "task" than the recently released Flash for code reviews and things (via OpenRouter running in pi.dev).

Flash makes a lot more initial mistakes, and then has to re-check stuff, and produces much more output compared to Pro. It often gets to the correct result eventually, but the output volume is often 5x more than for Pro, and the initial outputs are often wrong, with the first few saying something wrong (like there's a bug, or the code won't compile when it does), and then saying things like "Wait, let me re-check:", or "Actually, looking at it more carefully:" and then it thinks a bit more and eventually gets to the right answer.