azeirah

14 ms·

151.
by azeirah 2y ago
Or use Qt
152.
by azeirah 2y ago
I talk about this so much at my workplace and every time I mention slower phones I get laughed at by my boss. It's idiotic. "Just get an iPhone" Like... my dude, you rolled out of uni and started your own business and started
153.
by azeirah 3y ago
I think this strongly depends on the application and use-case. I've worked at two businesses so far, and they target small and medium-sized companies in a tenant-style manner. All data for one customer that is important enough to load
154.
by azeirah 3y ago
All of this makes me think we'll just be building meteor apps but with sqlite instead of mongo and minimongo whaha
155.
by azeirah 3y ago
You can simply check the Pull Request on llama.cpp on Github. JohanesGaessler (a core maintainer) has already ran the code and says it's an impressive speed-up. There isn't a thorough review by any of the core maintainers yet, but
156.
by azeirah 3y ago
Yeah, dottxt has a post on this, it's a technique starting with a c I believe, something like continuum or something? The op also has something about compressing finite state machines, it's basically the same thing with slightly d
157.
by azeirah 3y ago
I don't know if cross-platform applications are very popular anymore. Most of the good software I use is either single platform using native code, or it's Qt or it's electron. Only rare cases use something else. Sublime has t
158.
by azeirah 3y ago
I've been doing a lot of indie work with structured generation and llama.cpp, you can get extremely fast responses with caching and deterministic token skipping. When generating json, calling the llm when you already know that after
159.
by azeirah 3y ago
From how I understood it, it means they optimised the entire stack from CUDA to the networking interconnects specifically for data centers, meaning you get 30x more inference per dollar for a datacenter. This is probably not fluff, but it&#
160.
by azeirah 3y ago
He said he didn't rehearse well. I think it makes him come across very genuinely, not some dumb hyperpolished corporate blabla
161.
by azeirah 3y ago
Jensen revealed later that the LLM inference is 30x due to architectural improvements, it's massive. I don't know if it's latency or just 2-3x performance boost with 30x more customers served in the same chip. Either way, 30x
162.
by azeirah 3y ago
Yes.
163.
by azeirah 3y ago
I'm kind of surprised that for desktop applications more people don't use Qt. It's so good for these kinds of applications!
164.
by azeirah 3y ago
Jesus christ this is atrocious. I'm transgender and this is just the dumbest thing I've heard in a while. Yes we need some form of accomodation and it's nice if the system works _for_ us instead of against, but this is not it
165.
by azeirah 3y ago
The weight of that single word "tricky" in tricky test regression failures is very high! I can't imagine the obscure craziness you'll be encountering when working on a codebase that's easily over 25 years old, suppo
166.
by azeirah 3y ago
For myself I have several distinct cognitive weaknesses that I can often just unload onto an AI in a conversational style. For instance I know I'm good at breaking up large tasks into smaller ones, but because planning and executive fu
167.
by azeirah 3y ago
I agree with almost everything you said, although I do think this type of testing has a place. There are different types of testing, what you're describing sounds to me like testing the "core" of your code, part documentation
168.
by azeirah 3y ago
I recall a recent discussion about a technique to load the diff in weights between a lora and base model, zip it and transfer it on a per-needs basis.
169.
by azeirah 3y ago
Joe Armstrong can finally be happy :D
170.
by azeirah 3y ago
I revisit this article every so often. It is a truly amazing document.
171.
by azeirah 3y ago
Snapdragon is making big claims about the Snapdragon X Elite. It is an ARM chip instead of x86, meant for the laptop market. I haven't delved deep into it yet so I can't make any comparisons or give my own take. https:/&#x2F
172.
A Standard for the Transmission of IP Datagrams on Avian Carriers(rfc-editor.org)
2 points by azeirah 3y ago|0 comments
173.
by azeirah 3y ago
That's an impressive list of jargon whaha Love how deep the rabbithole has gone in just a year. I am unfortunately in the camp of understanding the post without needing a glossary. I should go outside more :|
174.
by azeirah 3y ago
This has a lot of overlap with the volunteer work I do. It's a charity for people with autism Once or twice a month me and a colleague gather and a couple of people are invited to talk about themselves, their lives and usually their pr
175.
by azeirah 3y ago
Eh, as I've learned, energy is limited. You can only do so much on one day. If you can improve something you've been working on all day very very mildly where it costs very little energy, it's fine. Switching to the next big
176.
by azeirah 3y ago
I was talking about m2 Macs. Just comparing that the best you can do with a gpu is 24GB, Macs go far beyond because of their integrated memory.
177.
by azeirah 3y ago
If you want to run decently heavy models, I'd recommend getting at a minimum getting 48GB. This allows you to run 34b llama models with ease, 70b models quantized, mixtral without problems. If you want to run most models, get 64GB. Thi
178.
by azeirah 3y ago
Not sure about this scenario precisely but I believe when it comes to filtering certain responses they have a separate model running in a similar manner to what meta recently released. I believe it's called llama purple? https:/&
179.
by azeirah 3y ago
Check out https://reddit.com/r/LocalLLama for better discussion on this topic. General rules of thumb * VRAM is king * Newer GPUs are better than older GPUs * Nvidia has better software support than AMD * Cloud is prob
180.
by azeirah 3y ago
ReMarkable recently got a new CEO, so who knows what's next? https://goodereader.com/blog/remarkable-news/remarkable-is-g...

More ›