Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
mysteria
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
11 ms
·
121.
▲
by
mysteria
2y ago
Services like Github Actions, Google Collab, and web-based IDEs likely meet this definition though as it lets users execute their own custom code on their cloud. So basically all developer stuff may require an ID check.
122.
▲
by
mysteria
3y ago
I'm pretty sure the drivers already gave consent to have their data disseminated and sold freely to third parties in the T&Cs they clicked through when setting up the navigation system. Of course this doesn't make it acceptabl
123.
▲
by
mysteria
3y ago
If you have the local infrastructure for it (or do it on cloud) VDI is an option. It's easy to snapshot, backup, and reimage, and if you store their files on a network drive a reimage won't affect it. You can also easily remotely
124.
▲
by
mysteria
3y ago
I don't use Discord anymore but the phone number thing seems new, in the past I was able to visit as a guest and be able to read messages but not chat. Then again Twitter and Reddit are doing the same thing now and forcing people to lo
125.
▲
by
mysteria
3y ago
If you want to make money you're often creating for your audience rather than for yourself too. For instance I know professional artists who don't enjoy their work that much since they end up having to please the client rather tha
126.
▲
by
mysteria
3y ago
I clicked on the Discord link in the spam message for fun, entered a random username, and immediately got asked to verify with my phone number before I could even read the messages in the chatroom. HN doesn't need to do that and I don&
127.
▲
by
mysteria
3y ago
I don't want them to require a email/phone for privacy's sake, but they should definitely have a captcha of sorts to limit bot accounts.
128.
▲
by
mysteria
3y ago
What a mess, this is literally the first time I saw something like this on HN. They've even started posting on this thread! HN has been running slow since the flood started and I wonder if it's causing a mini-DDoS effect. The user
129.
▲
by
mysteria
3y ago
You really have to write it for yourself if you want to keep going - unless you're lucky or have a large following already it's hard to get a lot of readers. I've written stories and serials and posted them on social media, m
130.
▲
by
mysteria
3y ago
If you interpret this very strictly it could even imply that this editor couldn't be used to say write open source code that's distributed to the public. Basically all you could do with it is write code at home for your own use.
131.
▲
by
mysteria
3y ago
Or at the very least release a short gameplay video so people know what they're getting into before paying.
132.
▲
by
mysteria
3y ago
Typically it's to the end of the school year OR until a parent comes to get it.
133.
▲
by
mysteria
3y ago
Yeah this is exactly what happens when you ask a base model a question. It'll just attempt to continue what you already wrote based off its training set, so if you say have it continue a story you've written it may wrap up the sto
134.
▲
by
mysteria
3y ago
They're on Hugging Face. Interestingly enough they don't compare it against Mistral 7B. https://huggingface.co/allenai/OLMo-7B
135.
▲
by
mysteria
3y ago
Is this one of the first LLMs of note that was successfully trained on AMD GPUs? I wonder how seamless the process was and if they faced any issues there.
136.
▲
by
mysteria
3y ago
> Pretty sensible IMHO - I live in a country with a reliable electricity grid, and outages due to UPS malfunction are about as common as power outages. Hence why modern servers and their associated networking equipment have dual power su
137.
▲
by
mysteria
3y ago
The problem is that this now becomes a long term investment, which doesn't work out when we have CEOs chasing quarterly profits and all that. Meanwhile Nvidia stuck with CUDA all those years back (while ensuring that it worked well on
138.
▲
by
mysteria
3y ago
I don't think virtio-gpu works on SPICE. I only use it for VMs on my desktop as I can display the output on the local machine using virt-viewer.
139.
▲
by
mysteria
3y ago
Are those WiFi phones plugged into the wall or are they cordless? I believe the advantage of DECT is that the phones consume much less power compared to WiFi which makes sense if they're on battery. Many of the IP phone vendors use DEC
140.
▲
by
mysteria
3y ago
The SPICE backend has decent OpenGL 3D support with software rendering, it's slow but it works for simple graphics. It's intended for 2D so the desktop's pretty fast IMO. That only works for Linux and Windows guests though, n
141.
▲
by
mysteria
3y ago
A lot of places with hardware authentication will provide a recovery key you can store somewhere you can access (e.g. as an encrypted file on an cloud storage service with no 2FA/geolockouts). Obviously the passwords for the file and t
142.
▲
by
mysteria
3y ago
On a 6.7" display, assuming 16:9, the screen is 5.8" wide. With 2.8k pixels across each pixel is 0.002 inches wide. I'm pretty sure you aren't viewing it at 100% resolution on the phone as each glyph would render as a n
143.
▲
by
mysteria
3y ago
Same for me! I could definitely understand the upper case text at 100% resolution on a 24" 1080p screen. The lower case text was terrible and nearly unreadable.
144.
▲
by
mysteria
3y ago
The only way out of this is an early retirement in a LCOL area or a job with a very good WLB (which is likely pretty rare for most HNers in the tech industry). Even ignoring overtime I'm typically tired when I get home from work and
145.
▲
by
mysteria
3y ago
This is the classic "money is time, time is money" conundrum. A teenager doesn't have the money to buy a fancy car or computer but they have the time to tweak and experiment to get the most out of it. Meanwhile an adult has t
146.
▲
by
mysteria
3y ago
How much do these CPUs cost? There'll be international buyers if they can beat Intel and AMD in terms of price to performance ratio, even if the chips are slower and consume more power. For general home and business computing Zen 1 per
147.
▲
by
mysteria
3y ago
The Ollama backend llama.cpp definitely supports those older cards with the OpenCL and Vulkan backends, though performance is worse than ROCm or CUDA. In their Vulkan thread for instance I see people getting it working with Polaris and even
148.
▲
by
mysteria
3y ago
The commercial internet services with 4 or 5 9 SLAs have multiple redundant lines to the building to guarantee this and they charge handsomely for it. Otherwise a downed telephone pole or something will automatically bring the service down
149.
▲
by
mysteria
3y ago
I've heard of people randomly signing up their coworkers to create plausible deniability.
150.
▲
by
mysteria
3y ago
> While making a three-point turn, she had put the car in reverse instead of drive, she said. It is a mistake she had made before with the Tesla gearshift. The car had zipped backward, tipping over an embankment and into a pond. I'v
More ›