tyfon

14 ms·

151.
by tyfon 3y ago
They can steal a lot more data from you using the app, that's why they insist on you installing it. I avoid native mobile apps for anything but the most essential for this reason. If it wasn't for the cars and public transportatio
152.
by tyfon 3y ago
This sounds like a massive security hole to me, especially if this runs in the cloud. Is this thing sending _all_ context of the programs on your windows computer to MS? I'm also a bit curious how they will actually pay to have million
153.
by tyfon 3y ago
To me it seems like turbo pascal is no longer a valid trademark[1], or am I missing something? I like it at least. I used turbo pascal, turbo c/c++ and tasm quite a lot in my youth for this sort of stuff :) [1] https://trade
154.
by tyfon 3y ago
That's really weird, but I remember in the past there were some older usb-a ports that did not send enough power to charge some stuff, perhaps your headset requires a higher current than the controller or something. Here I can charge m
155.
by tyfon 3y ago
Llama 65B is actually quite decent in other languages. I can just barely fit it in memory though with my 128 gb ram. Usually I run the 8 bit quantized version that use 80, but even the 4 and 3 but are ok compared to the fp16 30B version.
156.
by tyfon 3y ago
Until one team need a feature in another service that makes their development grind to a halt and the other team is not prioritizing. I have only seen this from the business side (I'm not a developer), but I have seen teams start codin
157.
by tyfon 3y ago
I see, the 2.9 GB requirements seems to imply a 3 bit weights? In any case I am happy to see these projects taking form. Perhaps one can eventually make the level of quantization dynamic based on the available vram etc :) I will definitivel
158.
by tyfon 3y ago
llama.cpp is not using the GPU, it runs fine on the CPU (if fast enough) I've scoured the web page for ram requirements for the various models but I can't see anything, will it be able to run let's say the 30B open assistant
159.
by tyfon 3y ago
Not just a raincoat, I've had snow, sunny summer, storm gusts, hail and rain on the same day just this week, I'm int he SW though not entirely west. I remember one time a few years back when I lived in Stavanger, I went to the pos
160.
by tyfon 3y ago
Because they want your data, it is a dark pattern. Also when you disable chat history / data collection they display a big box on the left side about "do you really want to do this" and a big green button to enable it again.
161.
by tyfon 3y ago
I stopped watching and blocked the linus channel (and may others) due to the faces. But I guess I am only one and the many likes these faces for some reason. It is also probably a failure of optimization where one started doing it and made
162.
by tyfon 3y ago
Most people are not capable seeing longer than the tip of their nose as we say in Norway. It is also very unpopular to tell people that they need to change.
163.
by tyfon 3y ago
Here in Norway, if they "let you go" immediately they still have to pay you for your contract notice time which is usually three months. There are some with six (my last job) or some with one. In my last job I was able to negotiat
164.
by tyfon 3y ago
I've found that GPT4 does very well for problems that have examples on the web, but I have yet to make it display a 3d cube using opengl (any version) in linux using c++ or python. So it depends heavily on the task and how much it has
165.
by tyfon 3y ago
The 272 Wh/kg is at cell level not pack level which the Tesla weight refers to. In my X has about 92 kWh usable energy when new and it uses around 225 Wh/km at 120 km/h and 170 at 90. The 3 and Y is even more efficient, mostl
166.
by tyfon 3y ago
I think the biggest issue with VRAM right now is that it's very expensive compared to regular system memory, perhaps with the exception of the DDR5 for the latest CPUs. But there should also be absolutely no issue in making a commercia
167.
by tyfon 3y ago
I am really hoping we'll get something like AMD infinity fabric up and running with GPUs accessing system memory or something like that soon. Then there is also moore's law :)
168.
by tyfon 3y ago
My local 65B llama prompt gave me this answer: tyfon:What is heavier, one pound of feathers or two pounds of lead? Omnius: Two pounds of lead are heavier than one pound of feather. Not bad :)
169.
by tyfon 3y ago
No they are not exclusive. But we do need alternatives to the big corp cloud models that are fully open source :)
170.
by tyfon 3y ago
This should be illegal, my 2010 mac pro only got support for like 5 years. Especially when it's just a matter of flipping a bit in the bios.. I have a PC from 1998 that still works using linux. That mac pro also runs happily with linux
171.
by tyfon 3y ago
And this is only the beginning, I don't believe for a second that "AI" will belong to the big corporations. These LLM models have no benefit from running "in the cloud" except for processing power. Lots of disadvant
172.
by tyfon 4y ago
I have a 5950x with 64 gb ram and they are quantized to 4 bit yes :) The weights are stored on a samsung 980 pro so the load time is very fast too. I get about 2 tokens/second with this setup. edit: forgot to confirm, it is llama.cpp e
173.
by tyfon 4y ago
> Why do these takes around open-source AI remain so popular? I can only speak for myself, but I have a great desire to run these things locally, without network and without anyone being able to shut me out of it and without a running co
174.
by tyfon 4y ago
If I remember, the reason they were not welcome was that they didn't want to censor stuff and hand everything over to the CCP. But I see the sibling post, I think the management probably wanted to get in here, but apparently the intern
175.
by tyfon 4y ago
Yes, they might not be 100% ethical in this decision, but the fact is that they are not operating in China. It would potentially be a huge market for them.
176.
by tyfon 4y ago
I am really surprised that the western governments still talk about how the free trade will somehow magically turn the CCP into a democratic government with a serious face. The reality is that they are extremely afraid that the flow of chea
177.
by tyfon 4y ago
Apple is the worst offender here, they really bend over backwards. It is proper to single them out too as they are one of the biggest providers of consumer goods at a high margin. There is absolutely no reason they should be forced to use s
178.
by tyfon 4y ago
I just bought another 64 for my computer that will arrive in the mail after easter. That will allow me to run the full 65B FP16 model, however it will probably be much slower than the 4 bit quantized version as it has to do more math. My bi
179.
by tyfon 4y ago
See my sibling comment for a bash script to launch it :) Edit: I have not tried to use it for extracting keywords / summarizing text. Perhaps I shall experiment a bit with it.
180.
by tyfon 4y ago
So I didn't see this before today, I will respond anyway and the siblings can also see. There is quite a big difference between the 65B model and especially the 13B and 7B models. But here is the bash script[0] I launch my "go to&

More ›