It's whatever their integration is with a vendor that does weight and balance before the flight. Likely not the vendor, as that vendor serves many airlines.
We need to get beyond the arguments of 364 vs 365 and develop a new equilibrium in terms of availability that accounts for humans being equipped with these new found breaks as we spend time off of work.
Who the NVDA board wants answering questions in some inevitable hearings when the other side of all the money behind AI is trying to push for controls on Chinese models.
Edit: A little more context. Because it doesn't matter if Alice Huang is a Chinese spy. She's also held titles at Megaspeed, a Singapore based company that legitimately rents GPU time to Chinese model providers. The big money AI people would be pushing for more controls on that legit path also.
Reminds me of the Coldplay concert couple, which happened to be a CEO and head of HR. Head of HR! And what's more interesting is that they only got removed when the affair became public. At the same time the company was small potatoes so presumably less interesting to state actors.
If you look at how airlines make money, you'll see that an agent skipping over the airline website's internal ads for the loyalty program, credit card partners, pushy upsells of seating/refundability, etc...destroys their revenue model. Not endorsing this approach of just banning agents, but it's easy to see why they don't know what to do yet.
Pass rates from the release's scores.json (commit 7753983), joined with the eval suite's manifest.jsonl. Tests = behavioral checks touching the service. All = overall pass rate. Happy/Reject = valid requests vs. requests AWS should refuse. Ctrl/Data = control plane vs. data plane. Solo/Integ = single-service vs. multi-service checks. Wire# / Wire% = count and pass rate of recorded request/response replay tests (a test passes only if every call matches).
That's one of the more visible innovations -- or should I say "deviations from common practice" -- but I don't think it's the most important one.
I think the most important good one is that it takes "register" renaming seriously. Calls start with a belt that only contains the arguments, for example, and that's done through the magic of renaming. Loops do something similar (they are a lot like calls on the Mill).
I believe there were lots of important bad ones: explicit speculation (super interesting but unlikely to work well), separating memory mapping and memory protection, having "dirty" flags in the cache at the byte level (to minimize writebacks), two PCs moving in opposite directions, very "interesting" encoding, two encoding levels: a general encoding shared between all implementations and many implementation-specific encodings via an installer/converter/loader/JIT, novel protection hierarchy mechanism, backing the renaming mechanism with a more compact representation of older values that aren't likely to be reused soon, automatic spill/restore of that to/from memory, ...
They came up with an interesting compromise between exposed pipelines and guaranteed repeatable execution for loads ("drop the value from this load onto the belt in n cycles").
If you desperately want a high performance implementation that doesn't use out of order execution then the Mill has a lot of good ideas... but an out of order execution implementation is much safer and is not nearly as scary as people think (in terms of complexity, gate count, frequency loss). Mill does not seem to be easy to implement as OOO which means it would have to work as an architecture with only in-order implementations, which sounds like a guaranteed dead end for the future of the architecture.
Maybe you can have a decoder that runs ahead and converts everything to something that isn't so full of relative value references (like the loads) -- and then bolt that onto a fairly standard OOO engine?
They spent way too long procrastinating, with zero silicon after decades of work.
For comparison, RISC-V started years later and had a tapeout within a year and working silicon available as a commercial product in less than seven years.
2003: Mill computing started
2010: RISC-V project started
2011: RISC-V first tapeout
2017: RISC-V production chips available in the market
RISC-V intentionally contained almost no innovation of any kind. The whole point was to have the most boring, standard RISC design imaginable, without anything clever in it. That's what you want for introducing students to their first RISC architecture, and also as a basis for doing computer architecture or CPU fabrication research: on zero-overhead loops, compressed instruction sets, scalable vector instruction sets, IoT machine learning, indium gallium zinc oxide thin-film transistors printed on kapton, zero-knowledge virtual machines, or whatever.
They specifically wanted to avoid the kinds of "quirks" that vendors like Intel and ARM used to make people subject to their patents.
They did try to avoid what they saw as the blunders of past designs, like not leaving enough opcode space for compressed instructions, making instructions conditional, leaving out byte and halfword memory access instructions, having a flags register, having a complicated page table structure, having register windows, having delay slots, etc. None of that was innovative. The actual design of the compressed instruction set and the vector instruction set were innovative, but that came later.
Even in 02016, I don't think US$2M was enough to tape out a silicon prototype of a working high-performance CPU. MPW ("shuttle") programs can give you working prototypes for as low as US$3000, but only in process nodes a decade or more behind the state of the art. (This is fine for analog, but for high-performance digital designs it's generally fatal.) US$30M is a much more plausible budget for the multiple rapid-fire tapeouts you need for anything close to a state-of-the-art process.
The Alpha left out byte and halfword memory access instructions, and ARM left out halfword access instructions; both had to correct the omission later. I said RISC-V was determined not to repeat "what they saw as the blunders of past designs", and you are quoting from that list of blunders, so it had byte and halfword memory access instructions from the beginning.
Yeah, but RISC-V came out of RISC-1,2,3,4; it didn't magically appear our of nowhere. "RISC-V was developed in 2010 at the University of California, Berkeley as the fifth generation of RISC processors created at the university since 1981".
> For “Vanilla,” note that the agent was told to use “vanilla JavaScript or TypeScript with Vite, without a UI framework.” Surprisingly, this underperforms several frameworks, which seems to demonstrate that having some framework can be better than nothing
I'd be curious how it does if you allow it some helpers or "micro frameworks" that are WAY short of a normal framework. Say vanilla js paired with nanostores as an example.
Not convinced those two pages help you in any substantial way to estimate your token burn rate across input/output/caching, etc.
Edit: It's some info, but not much around caching, or what context does vs just what you type in the chat, how reasoning or tool use affects cost/token burn, what control you might have on output tokens, etc.
https://www.cbsnews.com/news/united-airlines-system-outage-d...
reply