- You're generalizing a failure at delivering one consumer solution and ignoring the successful infrastructure research and development that occurs behind the scenes.
Meta builds hardware from chip to cluster to datacenter scale, and drives research into simulation at every scale, all the way to CFD simulation of datacenter thermal management.
- performance degradation observed using the first approach at high concurrency recently discussed here: https://www.hackerneue.com/item?id=44490510
- This conception is simplistic, a straw man, and one that appears to be wholly ignorant of the Cosmopolitan tradition and reasonable criticisms thereof.
- What a fallacy to equate the status quo of lawyers running the country with the United States itself.
- No, on the balance it is lawyers who protect companies from the people they harm and lawyers who constitute the government officials who perpetually exceed and expand their mandates.
Most of the senate are lawyers and it’s the most frequent occupation of a legislator.
- That’s irrelevant to the fact that the expected PnL on a millisecond of latency improvement is a lot more than 1M in some markets. Obviously if you are getting what ever trade you are concerned with off in less than one millisecond, the question isn’t well posed.
There are many more games to play than delta one takeout and the solutions certainly don’t fit on one or a handful of FPGA’s.
- They don’t charge fees, because they’re not a brokerage or exchange.
They pay fees to exchanges.
As a market maker, some rebates are given back conditional on their activity.
They have no users.
You’re just constantly obliviously asserting falsehoods that betray an almost comical lack of understanding of the reality of these businesses.
- You have no concept of the infrastructure and organization necessary to operate these enterprises.
All your posts here are low-information anti-finance rants.
- There’s tons of latency sensitive code outside of the FPGA systems and it is not simple.
- No, there are absolutely electronic trading markets where a difference of milliseconds of latency to certain events is worth more than a M PnL. That’s a long time.
- You can fingerprint transmitters.
Antennas would be much more difficult and likely moot.
- Only a few percent of Down’s syndrome cases result from Robertson translocations and may be inherited.
If the mother carries the translocation, the rate of recurrence isn’t much more than 10%. If it’s the father, it’s significantly less.
- Fantastic post.
Please do one on your analysis and optimization workflow and tooling!
- I’m not sure why you’re comparing __new__ to constructors in other languages.
Ruby has the same thing but it’s called ‘new’.
Implementing the type of customization (idiomatically) that __new__ provides in Kotlin and JS isn’t any cleaner.
- That’s not “the normal way”, it’s just the case when the LHS is immutable.
This behavior is congruent to C++ and it’s documented and idiomatic.
It’d be weird if the in-place operator deeply copied the LHS, and problematic for a garbage-collected language in high throughput applications.
- No, he finally said something that cost Murdoch money instead of making him money.
- Of course every bit of it survived. U-235 has a half life of over 700m years.
Whether it’s yellowcake or UF6, it was merely dispersed within the facilities and has not been mobilized. Recovery would consist of chemical processing of debris using common chemicals.
Fissile materials like enriched uranium and plutonium can only be neutralized by secure disposal or (nuclear) burn up.
When the USSR fell, US nuclear weapons specialists assisted former Soviet republics with securing the tunnels in which many weapon cores were vaporized by subcritical weapon testing, where a weapon core is imploded in such a manner that reactivity is limited below the level needed for low or high order nuclear detonation.
- You can embed these requirements into conventions that systematically constrain the solutions you request from the LLM.
I’ve requested a solution from Sonnet that included multiple iterative reviews to validate the solution and it did successfully detect errors in the first round and fix them.
You really should try this stuff for yourself - today!
You are a highly experienced engineer and ideally positioned to benefit from the technology.
- Radiation at desirable operating temperatures is a relatively weak means of heat transport, hence spreaders.
Only a big power-efficient chip like M2 Ultra could survive if it could _only_ radiate from one side of the wafer into CMB.
The rest of the silicon will become molten at 100% TDP: H100, Xeon, Core, EPYC, Ryzen.
Most would be over 400C at 1% TDP.
Conduction and convection are linear or close and are effective at desirable temperatures, whereas thermal radiation is quartic and highly convex.
They're building it for themselves and employ world-class experts across the entire stack.
How can NVIDIA develop "more integrated" solutions when they are primarily building for these companies, as well as many others?
Examples of these companies doing things you mention as being somehow unique to or characteristic of NVIDIA:
Complex kernel drivers or modules:
- AWS: Nitro, ENA/EFA, Firecracker, NKI, bottlerocket
- Google: gasket/apex, gve, binder
- Meta: Katran, bpfilter, cgroup2, oomd, btrfs
Hardware simulators:
- AWS: Neuron, Annapurna builds simulations for nitro, graviton, inferentia and validates aws instances built for EDA services
- Google: Goldfish, Ranchu, Cuttlefish
- Meta: Arcadia, MTIA, CFD for thermal management
Optimizing Compilers:
- Amazon: NNVM, Neo-AI
- Google: MLIR, XLA, IREE
- Meta: Glow, Triton, LLM Compiler
Acceleration Libraries:
- Amazon: NeuronX, aws-ofi-nccl
- Google: Jax, TF
- Meta: FBGEMM, QNNPACK