alecco

42 ms·

841.
by alecco 2y ago
And obedient.
842.
by alecco 2y ago
As a foreigner, I think this is too naive. The Western elites are in cahoots, that's why they meet all the time. They negotiate things and then they make a DUAL narrative to be pushed by ad-controlled media and social networks so they
843.
by alecco 2y ago
> (2016) Albertson's had a dozen eggs for $0.99 I wonder what would be the budget for all this today.
844.
by alecco 2y ago
Yeah, it should have (2020) Previous discussion https://news.ycombinator.com/item?id=23132549
845.
by alecco 2y ago
https://kernelnewbies.org/Linux_6.0#io_uring_features but only mentions zero copy and https://lwn.net/Articles/879724/ also https://www.phoronix.com/news/Linux-6.0-IO-Block-IO
846.
by alecco 2y ago
I'd be surprised if the LLM gave no warnings with the suggestion of using spot instances. Did you fully read its advise? And it's the second sentence of the first result in a simple search. https://en.wikipedia.org&#x2F
847.
by alecco 2y ago
Share your "I used a tool and shot myself in the foot because I was lazy" stories.
848.
by alecco 2y ago
I'm sure one of the Big 4 accounting firms and Goldman could step up to help "fix" any problem given their experience with Greece.
849.
by alecco 2y ago
> People really underestimate the damage these tools are doing to society. Even when heads of state are being extorted. Morocco used it against France and Spain. It fizzled out of the news cycle and nothing happened. And those countries
850.
by alecco 2y ago
https://en.wikipedia.org/wiki/Whataboutism
851.
by alecco 2y ago
Why is this downvoted? The comment sounds plausible. The Chinese mob (wink wink) has the biggest money laundering operation in the world. They have brokers who pick up (source) and deliver (destination) the cash. They do all the currency co
852.
by alecco 2y ago
Aaaaand it's flagged out of the front page. @dang, so early in the day this is obviously some coordinated manipulation. 31. 206 points 9 hours ago US judge finds Israel's NSO Group liable for hacking journalists via WhatsApp (
853.
by alecco 2y ago
Gen Z was conditioned with algorithmic timelines and loot boxes (gambling).
854.
by alecco 2y ago
* acording to CCP's reported numbers
855.
by alecco 2y ago
https://www.google.com/search?q=apps+automate+activism+onlin... You an see there are a LOT for many different well funded groups.
856.
by alecco 2y ago
Bots are one problem. A worse problem is coordinated online activists for different causes flooding posts to steer the discussion. There are websites and apps and many thousands active for just one specific group.
857.
by alecco 2y ago
The majority look for leetcode monkeys more than experience. I gave up applying because of that.
858.
by alecco 2y ago
You can change your browser search engine to have the udm=14 parameter. https://www.google.com/search?q=%s&udm=14 https://arstechnica.com/gadgets/2024/05/google-searchs-udm14...
859.
by alecco 2y ago
I really like CUDA, but these are crazy expensive for the specs. It's sad there is no competition.
860.
by alecco 2y ago
> How is China filling a void in local energy production? China is investing in power generation almost as much as the rest of the world combined. That energy is used by their massive and growing manufacturing industry. Bye bye EU manufa
861.
by alecco 2y ago
I'm waiting for next gen hardware. All the companies are aiming for AI acceleration.
862.
by alecco 2y ago
The green coalition governments across the EU are leaving a huge mess behind. De-industrialization at a pace never seen before at peace times. China is happy to fill in the void.
863.
by alecco 2y ago
But the Swedish government is taking blame for local shutdowns. The previous day her boss (PM) said so. The 4th paragraph reads: "“I realise that nobody is happy when I say that ‘if we hadn't shut down half of nuclear power, we wo
864.
by alecco 2y ago
So Intel is doing the same artificial performance limitation. Nvidia 4060s are also PCIe 4.0 with half lanes (x8). Argh.
865.
by alecco 2y ago
And both 3090 and 4090 can do 32 GB/s host-device. Not far from CPU-RAM. You only load the matrix once. The bandwidth for the matmul is orders of magnitude larger and happens all in device and mostly in cache.
866.
by alecco 2y ago
4090 (2022) PCIe 4.0 x16 is quite decent. The major limit is memory, not bandwidth. And 3090 (2020) is also PCIe 4.0 x16, and used cards are a bargain. You can hook them up with Nvlink. Nvidia is withholding new releases but the current har
867.
by alecco 2y ago
Consumer cards are PCIe 4.0 x16. H100 PCIe version is PCIe 5.0 https://en.wikipedia.org/wiki/Hopper_(microarchitecture) . And it's been out 2 years already.
868.
by alecco 2y ago
> Each version is severely memory bandwidth bottlenecked, the CUDA version suffers the most with its practical 11.8 GB/s device-to-host bandwidth due to its PCI-Express 3.0 x16 interface. PCIe 3.0? What? https://cowfreedo
869.
by alecco 2y ago
I think the post is a bit disingenuous. But about bandwidth, matrix multiplications happen mostly in cache and that has a lot more bandwidth than RAM. Blocks of the matrix are loaded to cache (explicitly in CUDA) and used multiple times the
870.
by alecco 2y ago
> What is it that changed from AGPL in SSPL that was absolutely needed? https://opensource.stackexchange.com/a/8028 Big cloud players would put a thin layer on top and ignore the AGPL.

More ›