xcv123

35 ms·

451.
by xcv123 3y ago
Try it with the most illegal electric vehicle. Won't compete with Top Fuel dragsters. Fundamental physical limitations of batteries versus combustion. But the Cybertruck has 10x the torque of the equivalent ICE vehicle, so it can beat
452.
by xcv123 3y ago
By "loser" I only meant by comparison with the best ICE in that category of pure drag racing. It's definitely not a loser, but it is a "loser" versus ICE. I'm not saying the Cybertruck is shit. It's a torq
453.
by xcv123 3y ago
Motorsports enthusiasts look at Nürburgring lap times. The Porsche 911 is at the highest level of performance on any track where you actually have to use the brakes and steering wheel. If the Cybertruck can do a good lap on the Nürburgring
454.
by xcv123 3y ago
> Language Model that puts one token after the other, The interesting thing about the one-token-at-a-time process in OpenAI transformer LLMs is how the Attention Mechanism is executing ~1600 processes in parallel over the entire context
455.
by xcv123 3y ago
If you were using the free ChatGPT then you were playing with an obsolete LLM. GPT-4 has an estimated 10x parameters of GPT-3.5 (1.8 trillion vs 175 billion), and other improvements.
456.
by xcv123 3y ago
> And it'll be under trade embargo no doubt due to the IP issues so they will have a very limited market to scale into. This was created for their own domestic market. China's goal is achieve independence of foreign technology.
457.
by xcv123 3y ago
China's goal with Loongson is to reduce reliance on foreign technology. Their government can easily build a "road" with a billion workers.
458.
by xcv123 3y ago
TSMC depends on ASML lithography equipment for manufacturing. Europeans have the leading technology.
459.
by xcv123 3y ago
Not viable. "An aviation-size, worldwide hydrogen supply and airliners capable of using it are decades and trillions of dollars away. In terms of a timely green return on investment, the money would be much better spent on SAJF, both f
460.
by xcv123 3y ago
To put it in perspective, the A380 max landing weight is 391 tonnes, versus the 7000 tonnes of lithium batteries required to power a long haul flight.
461.
by xcv123 3y ago
> you are asserting fundamental physics prevents it from happening? The energy density of batteries vs fuels is the current limitation. We don't know if this problem can ever be solved. A business cannot gamble on some magical hopes
462.
by xcv123 3y ago
Try reading the article before commenting Here's a link from the article https://www.rolls-royce.com/media/press-releases/2023/13-11-...
463.
by xcv123 3y ago
Yes, hence my quick edit of my comment above just before you replied
464.
by xcv123 3y ago
The scenario is a customer opens a chat box on your website and asks some questions for the LLM. You wouldn't expect your customers to search your internal Confluence pages. The LLM would be trained on all of your internal documentatio
465.
by xcv123 3y ago
But it only works if you have already written all of the documentation manually, and kept that up to date. It's basically a chat bot that knows all of your documentation.
466.
by xcv123 3y ago
It was started by him but he is no longer involved. Even if he was still running the company I would trust him infinitely more than Google or Microsoft. Also I like that they are based in New Zealand. Their software is well engineered and i
467.
by xcv123 3y ago
https://mega.io The quality of their software is vastly superior to Google Drive. They have mobile, desktop, and CLI apps on Windows/Apple/Linux. Been using them for a year without issues.
468.
by xcv123 3y ago
The failures being reported in the comments here are unacceptable. Better to have redundancy across services that are more reliable with better customer support. Choose companies that specialize in file storage, not a shitty advertising com
469.
by xcv123 3y ago
> based on a table of contents? The entire pdf is available as a free download on that page. First link at the top. https://github.com/udlbook/udlbook/releases/download/v1.16/U...
470.
by xcv123 3y ago
A poorly trained LLM would do that poorly, by definition. What I'm saying is you review its findings and verify it. Use your own intelligence to determine if its bullshit or not.
471.
by xcv123 3y ago
FYI, LSTMs are not obsolete. They are still the best option in many cases and are being deployed today.
472.
by xcv123 3y ago
So what? The same fundamental machine learning concepts are still relevant to deep learning. It's almost like arguing that everything you learned as a Java developer is completely useless when a new programming language replaces it.
473.
by xcv123 3y ago
Agreed. Look at the table of contents of this book. Whatever fundamental machine learning concepts you learned with SVM or other obsolete algorithms is still useful and applicable today.
474.
by xcv123 3y ago
Agreed that this is a different situation. If they drop Rosetta then Crossover is gone and Game Porting Toolkit is finished. Also Docker will be limited without x86. I'm using that to run SQL Server on Linux on MacOS for work.
475.
by xcv123 3y ago
Based on decades of historical precedent. Apple has migrated from 68k to PowerPC to Intel to ARM. You can no longer run PowerPC applications on current MacOS, as Rosetta v1 was dropped five years after migrating to Intel. Likewise 68k emula
476.
by xcv123 3y ago
Yes it requires an extremely large diverse training set for the first unsupervised stage (pre-training). Then fine tune it on the smaller data set. But we may need to wait for the next generation of LLMs that incorporate planning algorithms
477.
by xcv123 3y ago
Given the cost of the hardware and power and everything else required to run it and support it, how can you say that 20 cents an hour is unreasonable? There's very little profit margin there. At this price it would take roughly a year
478.
by xcv123 3y ago
That would be a question for the LLM. It could explain where exactly our understanding is incorrect and provide examples.
479.
by xcv123 3y ago
That only contains a few million tokens. Useless for pre-training an LLM from scratch. You would need to find billions of tokens.
480.
by xcv123 3y ago
https://www.ebooks.com/en-nz/book/210497580/efficient-linux-...

More ›