loudmax

41 ms·

301.
by loudmax 3y ago
Cute, but "le chat" literally means "the cat". I presume most young Francophones who are likely to actually use Mistral will pronounce it in Franglais as "le tchatte".
302.
by loudmax 3y ago
I think a lot of these are niche applications, intended for other startups working in this space. The services they offer are overkill for the majority of casual users trying to get a handle on AI by dabbling with llama.cpp or Stable Diffu
303.
by loudmax 3y ago
I certainly would not count out Google's engineering talent. But all the technical expertise in the world won't matter when the leadership is incompetent and dysfunctional. Rolling out a new product takes vision, and it means ta
304.
by loudmax 3y ago
At the end of the article, a single image of the bookshelf uploaded to Gemini is 258 tokens. Gemini then responds with a listing of book titles, coming to 152 tokens. Does anyone understand where the information for the response came from?
305.
by loudmax 3y ago
I get the impression that a lot of open source adjacent AI companies, including Stability AI, are in the "???" phase of execution, hoping the "Profit" phase comes next. Given how much VC money is chasing the AI space, th
306.
by loudmax 3y ago
The GPL was intended for computer code that gets compiled to a binary form. You can share the binary, but you also have to share the code that the binary is compiled from. Pre-trained model weights might be thought of as analogous to comp
307.
by loudmax 3y ago
What's really cold? There are heat pumps that are advertised as working at -15F. On the days when it gets that cold, your heat pump will need to switch to electric heat. So on those days, your heat pump is no better than an electr
308.
by loudmax 3y ago
Greg Egan's Mastodon profile is https://mathstodon.xyz/@gregeganSF He's fairly active there and likes to post math problems. I liked Permutation City, but Diaspora really blew me away. Every time you begin to ge
309.
by loudmax 3y ago
This is entirely due to complacent management. In the early 2000's Google's cleverness and engineering competence built a money cannon. The founders recognized the challenges that would be involved in running a huge corporation
310.
by loudmax 3y ago
I tend to agree. `.internal` is too easy to misspell or misread. Not that it's that hard, but operating at scale you really want to make things as simple and easy as possible.
311.
by loudmax 3y ago
You might be thinking of Fooocus: https://github.com/lllyasviel/Fooocus The Stable Diffusion web interface that got a lot of people's attention originally was Automatic1111: https://github.com/AUTO
312.
by loudmax 3y ago
Haaretz is Israel's most internationally respected Israeli paper. It's generally liberal and sympathetic to Palestinians, so the warning could also be a anti-Palestinian jab at a left-leaning news source.
313.
by loudmax 3y ago
Aircraft design and build has a very long lead time. It's plausible that Boeing success in the early 2000's was the result of sound decisions and long-term planning that were made early on. The emphasis on immediate shareholder
314.
by loudmax 3y ago
> Entreaties and ‘trying to make Fetch (firefox) happen’ won’t work, evidently. Pop culture references aside, `fetch` was also a browser! Sort of. `fetch` was a CLI predecessor to `wget` and `curl`.
315.
by loudmax 3y ago
Greg Egan is fairly active on Mastodon. You can follow along at https://mathstodon.xyz/@gregeganSF
316.
by loudmax 3y ago
I strongly second the recommendation for Diaspora. The story begins about 1000 years in the future, with characters that are trans-human, or post-human. Starting from that baseline of normalcy, then it gets weird. But it's not weird
317.
by loudmax 3y ago
I didn't hate reading Ulysses, but I didn't particularly enjoy it. I suspect my experience wasn't that different from yours. There were parts that made sense, and even some parts where I could follow along with the humor, b
318.
by loudmax 3y ago
That sounds about right. They provide a script to combine their "Predictor" weights to the original models, but I don't see anything obvious in the front page of the Github repo about how to create those weights. A 10x speed
319.
by loudmax 3y ago
I don't think that's a dumb question at all. It depends on your objectives and how much resources you're willing to spend. These open weights models can be retrained. Start with a foundational model like Llama2 or something
320.
by loudmax 3y ago
TheBloke's GGUF model card tells you how much RAM you'll want to run the various versions of the model: https://huggingface.co/TheBloke/Mixtral-8x7B-v0.1-GGUF#provi... In the article, Simon mentions the Q6_K.
321.
by loudmax 3y ago
That softness is critical. Not for the kids, but for the adults who have to clean up after them. Stepping barefoot onto a Lego brick hurts, but stepping barefoot onto a Duplo brick is much worse. Those things look innocent enough, but in
322.
by loudmax 3y ago
The types of processes that AI is shaping up to disrupt are processes that are currently performed by people. Take, for example, accounting. A $15,000/year subscription to an accounting AI might be slightly worse than a full time $150
323.
by loudmax 3y ago
I've had mostly the same success, but it's disconcerting that I even have to block those guys to begin with. Also, at this point the suggestion algorithm seems close to worthless. My Youtube suggestion feed is either channels tha
324.
by loudmax 3y ago
Interesting that they attach doors, then remove them, then put them back on again. It raises the question, why did they bother to attach them the first time around? My guess is that the answer is the paint process. There's some range
325.
by loudmax 3y ago
This is pure speculation, but maybe it's beneficial in some circumstances to make way for the younger generation? Natural selection doesn't necessarily favor genes that optimize the health of a particular animal. Natural selectio
326.
by loudmax 3y ago
Note that running these GGUF models currently requires a forked version of llama.cpp: https://github.com/ggerganov/llama.cpp/pull/4406 The GGUF handling for Mistral's mixture of experts hasn't been
327.
by loudmax 3y ago
Depends what you mean by "best", but Fooocus is very accessible for getting started with Stable Diffusion.
328.
by loudmax 3y ago
I applaud this project's focus on correctness and efficiency, but I'd also really like a version of `jq` that's easy to understand without having to learn a whole new syntax. `jq` is a really powerful tool and `jaq` promises
329.
by loudmax 3y ago
Apparently there are "200k" versions of the 6B and 34B models. I don't see an explanation in the Model card for what the "200k" means here. If that's context size, that is impressive. TheBloke has GPTQ, GGUF,
330.
by loudmax 3y ago
Most of your interaction with your computer is via the Desktop Environment. Which is Gnome in a default Ubuntu installation. If you're not particularly interested in exploring alternate operating system philosophies, I'd recomme

More ›