HNHacker News
TopNewBestAskShowJobs

xcv123

817 karma · joined May 10, 2023

submissionscomments
xcv123··on Elon Musk drops suit against OpenAI and Sam Altman
Elon offered funding. OpenAI declined the offer then signed a deal with Microsoft. Elon did not force them to choose Microsoft. It was simply a disagreement. After they signed up with Microsoft, Elon tried to get some of his money back. Didn't work. Too bad. Now he can move on with his life.
xcv123··on Elon Musk drops suit against OpenAI and Sam Altman
It was a public announcement. https://openai.com/index/openai-and-apple-announce-partnersh...

OpenAI models run on Azure infrastructure. Apple is just a client and will be paying Microsoft via OpenAI to run it.

Apple also has their own models running on their own infrastructure. I'm not saying that Microsoft or OpenAI has anything to do with that.

xcv123··on Demystifying NPUs: Questions and Answers
You're right that GPU is a misnomer these days. The top-end chips are for HPC and scientific computing in general. Good for training, inference, and non-AI applications such as molecular dynamics simulations. They can accelerate FFT and linear solvers. Their applications are broader than AI/ML. Useful for physics simulations.
xcv123··on Elon Musk drops suit against OpenAI and Sam Altman
Having Apple dependent on and paying for Microsoft infrastructure is a win for Microsoft. Of course the deal benefits Microsoft as much as Apple.

Microsoft needs Apple as a strong competitor. Otherwise they will be declared a monopoly and broken up by regulators. Microsoft was smart to use OpenAI as a proxy for this deal.

xcv123··on Elon Musk drops suit against OpenAI and Sam Altman
They declined to give Elon Musk control when he offered them a billion to save the company. Instead they turned around and gave up control to Microsoft for billions.

Do you honestly believe that if you were in Elon Musks shoes, that you would have given away a billion dollars without any conditions? How would you feel when you saw them turn around and sell out to Microsoft instead? Expecting $1B unconditionally seems like an asshole expectation. Delusional and childish entitlement.

Do you believe that Microsoft has donated billions to OpenAI unconditionally?

xcv123··on Elon Musk drops suit against OpenAI and Sam Altman
"it’s a dick move" implies that you are somehow entitled to that money. Otherwise it's a fair move.

It is extremely delusional to expect to receive $1 billion dollars unconditionally.

xcv123··on Elon Musk drops suit against OpenAI and Sam Altman
Fair enough. But I would do the same if I were in his shoes just out of spite for selling out to Microsoft. Sue them for everything they are worth.
xcv123··on Elon Musk drops suit against OpenAI and Sam Altman
"in the hopes of bankrupting them"

Where do you get that idea?

How exactly were they entitled to his $1B donation?

xcv123··on Demystifying NPUs: Questions and Answers
By the same reasoning, a CPU is no different to a GPU. They can both do matrix calculations.

A GPU is optimised for 3D rendering (and is useful for parallel computations in general). An NPU is optimised for neural network inferencing. These algorithms both involve matrix mathematics but they are not the same. The NPU hardware design matches the deep neural network inferencing algorithm. For example it has an "Activation Function" block dedicated to computing the activation function between neural network layers. It is optimised and specialised for one very specific algorithm: inferencing. A GPU would beat an NPU for training, and any other parallel computations besides inferencing.

xcv123··on Apple's AI Strategy in a Nutshell
That guy is just some random blogger and I wouldn't take that to literally mean they use the same chips as a Macbook. Apple only says they use "Apple Silicon". Common sense to assume it is designed for servers and AI inferencing.
xcv123··on Apple's AI Strategy in a Nutshell
No one said they are M chips. These would be custom made for servers and AI inferencing. Also note their "ARM based SoC" has a GPU with 800 GB/sec bandwidth in the M2 Ultra.
xcv123··on Apple Intelligence for iPhone, iPad, and Mac
I've been using Siri more often recently and surprised at how capable it is for something that runs entirely on a phone. The speech recognition is perfect and it can do basic tasks quite well. Send messages, lookup word definitions, set timers and alarms, check the weather, adjust timers/alarms, control Spotify, call people, adjust the brightness and sound level, control lighting in the lounge, create and read notes or reminders, etc. It all works.
xcv123··on Apple Intelligence for iPhone, iPad, and Mac
Try reading the thread before mindlessly replying.

"I fully expected them to announce a stupendous custom AI processor that would do state of the art LLMs entirely local."

State of the art LLM means GPT-4 or equivalent. Trillion+ parameters. You won't run that locally on an iPhone any time soon.

xcv123··on Apple Intelligence for iPhone, iPad, and Mac
Apple uses TSMC for fabrication. The roadmap for TSMC and Intel are planned years in advance.

Two orders of magnitude improvement in 6 months? Not possible. Have you heard of Moore's Law? Maybe in 20 years.

https://en.wikipedia.org/wiki/Moore%27s_law

xcv123··on Apple Intelligence for iPhone, iPad, and Mac
Why would you expect that? State of the art LLMs need a GPU with hundreds of GB of RAM costing tens of thousands of dollars. Apple doesn't have magical supernatural powers. They are still bound by the laws of physics. Siri is still running on device (mostly) and is quite powerful with this new update.
xcv123··on Detecting a PS2 Emulator: When 1*X does not equal X
You're welcome.

I learned about the basics here http://www.emulator101.com

Then I read the docs at NESdev Wiki and eventually wrote my own NES emulator.

https://www.nesdev.org/wiki/Nesdev_Wiki

If you have good documentation then you're just following the specs and turning that into code. Next level difficulty is reverse engineering the hardware, figuring out how it works when there isn't any documentation. These guys did that for NES back in the 90s and early 2000s, writing little test ROMs to see what the hardware does, in addition to whatever official or unofficial developer docs they could find.

xcv123··on Detecting a PS2 Emulator: When 1*X does not equal X
You start by reading hardware documentation and you don't need to understand the machine at an electronic level. It's not a simulation of digital circuits. It doesn't need to be that complicated. An 8-bit CPU is a simple state machine with only a few bytes of state (registers). You can read the program one byte at a time and simulate whatever operation the CPU would do after reading that byte. They are very simple operations like adding and subtracting numbers, loading and storing bytes.

http://www.6502.org/users/obelisk/6502/registers.html

http://www.6502.org/users/obelisk/6502/instructions.html

Your 6502 CPU emulator would read the next few bytes of the program, interpret the bytes as an instruction, execute the instruction, which involves updating a couple of registers/counters in the CPU, doing some arithmetic or bitwise operations, and possibly loading or storing a byte of data from one location to another. This process is then repeated in an endless loop.

It is a simulation of the fetch-decode-execute cycle.

https://en.wikipedia.org/wiki/Instruction_cycle

xcv123··on Detecting a PS2 Emulator: When 1*X does not equal X
Mainframes typically execute batch processes on a CPU. Much simpler than a game console with a GPU. Cycle-accurate emulation is less relevant for mainframes.
xcv123··on Super Heavy has splashed down in The Gulf of Mexico
The good guys dropped nuclear bombs on 200k+ civilian men, women, and children. There was literally no other option. USA #1.
xcv123··on Detecting a PS2 Emulator: When 1*X does not equal X
You would probably need a terahertz CPU for cycle-accurate PS2 emulation.

Cycle-accurate PS2 emulation means emulating the state of the CPU, GPU, other interacting processors, and their various interconnecting busses at clock cycle granularity and possibly at sub-cycle granularity if the processors are running asynchronously.

xcv123··on Detecting a PS2 Emulator: When 1*X does not equal X
That is not perfectly cycle-accurate, but it is accurate enough to run almost anything without issues.
xcv123··on Super Heavy has splashed down in The Gulf of Mexico
6. Accept Japan's request for peace negotiations. Stop fighting and begin negotiating their conditional surrender.
xcv123··on Super Heavy has splashed down in The Gulf of Mexico
> Conditional surrender would have been completely unacceptable

And some would argue that nuking entire cities is completely unacceptable.

xcv123··on Super Heavy has splashed down in The Gulf of Mexico
I understood what you wrote. The interesting part of this quote is apparently General Patton also realised they should not have been fighting against Germany in the first place. There are other quotes from him that indicate some regret of fighting against Germany instead of the Soviets, by the end of the war.
xcv123··on Super Heavy has splashed down in The Gulf of Mexico
General George S. Patton:

"We may have been fighting the wrong enemy (Germany) all along. But while we're here (on the Soviet border), we should go after the bastards now, 'cause we're gonna have to fight 'em eventually."

https://books.google.com/books?redir_esc=y&id=32DxAAAAMAAJ&d...

xcv123··on Super Heavy has splashed down in The Gulf of Mexico
Japan was already trying to negotiate conditional surrender when Hiroshima and Nagasaki were nuked. The US demanded unconditional surrender, and they nuked hundreds of thousands of civilians to force that demand.

The bombing of Dresden is arguably a war crime by todays standards. It was unnecessary.

Remember that we were allied with the Soviet Union (Joseph Stalin was the "good guy" on our side). After WW2 he was given half of Europe as a reward, forcing that half of Europe to become communist, and the Soviets got to write the history books about Germany and WW2. Not the most transparent and unbiased source of information.

The Soviets and other allied soldiers (the good guys) also had a free-for-all with the German ladies after winning the war.

"The majority of the assaults were committed in the Soviet occupation zone; estimates of the numbers of German women raped by Soviet soldiers have ranged up to 2 million"

https://en.wikipedia.org/wiki/Rape_during_the_occupation_of_...

xcv123··on Vulkan1.3 on the M1 in one month
Wow.

  triangle.cpp
  1181 lines (1048 loc) · 56.2 KB

  z80.cpp
  493 lines (440 loc) · 19.3 KB
  
  z80.h
  37 lines (30 loc) · 720 Bytes
https://github.com/SaschaWillems/Vulkan/blob/master/examples...

https://github.com/rasky/cz80/blob/master/z80.cpp

xcv123··on Microsoft blocks Windows 11 workaround that enabled local accounts
Most of the hardware is supported now. My old HP laptop disintegrated within a few years and became landfill, but my 9 year old Mac is indestructible and runs anything I want. Linux, Windows, macOS. Also repaired it myself (new battery, new left speaker). The latest Macs are built like tanks and will easily last 10+ years.
xcv123··on I am sick of LeetCode-style interviews
Memorization of the hashmap package name is the stupidest Java programming test I have ever heard of.
xcv123··on Microsoft blocks Windows 11 workaround that enabled local accounts
https://brew.sh

https://asahilinux.org

← PreviousPage 4 of 28Next →