linuxftw

6 ms·

91.
by linuxftw 1y ago
Journalists get fired for making up breaking news, not lifestyle pieces. 100% this article is fake.
92.
by linuxftw 1y ago
Well, at some point you won't have a choice. The government is going to ban ICE vehicles, tax the existing ones, and all the electrics will be everything by wire.
93.
by linuxftw 1y ago
Nasdaq lost 3.47% and SP lost 2.71%. That's more than 1%, obviously. Very large downwards move.
94.
by linuxftw 1y ago
NVIDIA's forward P/E ratio is approximately 29.94, indicating the price investors are willing to pay for each dollar of estimated future earnings. This ratio is lower than its trailing P/E of around ~53. How are they overval
95.
by linuxftw 1y ago
One would think this question would have been answered directly in the FAA report, but it got zero mentions. We got zero data on how often MCAS made adjustments and how often in the report. Basically, we only have Boeing's word for it
96.
by linuxftw 1y ago
The ability for the masses to create any video just by typing, among the other features, is not novel technology? Or is it just the lack of emotional response?
97.
by linuxftw 1y ago
This is at best a Boeing talking-point. We don't have any data that states it is safe to operate the MAX without MCAS. It's quite possible that probable scenarios would result in stall faster than a human can react.
98.
by linuxftw 1y ago
During that era, SPARC servers were the absolute premium units inside the datacenter. That aligned better with Oracle selling servers than Google, who doesn't sell servers, IMO.
99.
by linuxftw 1y ago
There's a lot of hate for Meta in here. They are actually innovating unique new products here. I don't know if they will be commercially successful, but they at least look interesting.
100.
by linuxftw 1y ago
To me, this just validates what AMD has been doing for over a decade. Integrated GPUs for personal computing are the way forward.
101.
by linuxftw 1y ago
Changes across shared resources should be few and far between. Terraform really should only be used to setup long-lived resources, such as your VPC, initial IAM, and a bootstrapping system/management plane (eg, your kubernetes cluster
102.
by linuxftw 1y ago
If you're at the point of managing thousands of resources inside a single statefile, and that makes the most sense to your setup: you've outgrown terraform.
103.
by linuxftw 1y ago
Perhaps being thinner allows for people to put their phone into an external case, the combined size which is approximately the bulk of the current generation.
104.
by linuxftw 1y ago
Sure, but that doesn't seem to apply here whatosever. 475 people isn't 'knowledge transfer.'
105.
by linuxftw 1y ago
The whole idea of building manufacturing in the US is to employ Americans. And if they can't find enough qualified Americans, it's because WAGES ARE TOO LOW.
106.
by linuxftw 1y ago
All my personal machines run linux. At work my choices are Mac or Windows. If Macs were still x86_64 I might choose that and run a VM, but I have no interest in learning the pitfalls of cross arch emulation or dealing with arm64 linux dis
107.
by linuxftw 1y ago
This. I run docker in WSL. I also do 100% of my development in WSL (for work, anyway). Windows is basically just my web browser.
108.
by linuxftw 1y ago
The average windows box is full of so much bloatware, it likely evens out.
109.
by linuxftw 1y ago
Customarily duties are paid by the importer, often through a broker. How will this be handled for small items? Most likely a business will import products in bulk and pay the duties as normal. How will it work for small, one off purchases
110.
by linuxftw 1y ago
Yeah, that sounds decent for some one-shots. The unified memory systems can have longer back-and-forth context chats, but at slower speed (at least on AMD). I find Qwen 3 Coder to be quite usable, I get around 20TPS on my AMD AI 350 system
111.
by linuxftw 1y ago
What nobody seems to ever share is the context and TTFT (time to first token). You can get a very good TPS by using small prompts, even if the output tokens are very large. If you try to do any kind of agentic coding locally, where contex
112.
by linuxftw 1y ago
Yeah, I've kinda quickly thrown in the towel trying to figure out what's 'best' for smaller memory systems. As things are just moving so quickly, whatever time I invest into that is likely to be for nil.
113.
by linuxftw 1y ago
> AIUI, the UK is a democracy The House of Lords disagrees and the Monarch disagree. Sometimes they cosplay as a democracy.
114.
by linuxftw 1y ago
That's true only in a vacuum. For example, should I run gpt-oss-20b unquantized or gpt-oss-120b quantaized? Some models have a 70b/30b spread, and that's only across a single base model, where many different models exist at
115.
by linuxftw 1y ago
Languages evolve over time. Someone's rules from 200 years ago need not apply, just as they likely didn't apply 200 years prior to that.
116.
by linuxftw 1y ago
Less than half-full planes were very frequent prior to the big airline mergers in the 2010's. The loss of US Air, Contentintal, and Air Tran has completely ruined air travel. On most domestic flights, the middle seat was for children
117.
by linuxftw 1y ago
It's always just a few stations. Driving up the property values of the landlords in the cities. It's a never ending cycle.
118.
by linuxftw 1y ago
It's an easy problem to solve if you want to use government levers: Raise the minimum wage inside London. Keep raising it until businesses start moving to other cities.
119.
by linuxftw 1y ago
> Well, first up, they'd have spotted that our major cities need more frequent and faster rail connections from suburbs to centres and that these are prevented at the moment by insufficient platform capacity in stations like Leeds,
120.
by linuxftw 1y ago
I worked a big health care company recently. We were using Azure's private instances of the GPT models. Fully industry compliant.

More ›