sandGorgon

8 ms·

61.
by sandGorgon 2y ago
have u tried using UDP over webrtc ? https://geckos.io/
62.
by sandGorgon 2y ago
shameless plug - i build a JS/TS framework which tries to solve the abstraction problem. we use a json variant called jsonnet (created at google. expressive enough for kubernetes). https://github.com/arakoodev/Edge
63.
by sandGorgon 2y ago
github in india is affected.
64.
by sandGorgon 2y ago
they should have done this in the first place. there was absolutely no reason to have parallel initiatives in android-on-the-desktop, fuschia, android and chromeos. Google's operating system is android. it has won. extend that architec
65.
by sandGorgon 2y ago
can i use this in htmx ?
66.
by sandGorgon 2y ago
hi. this is very useful info thanks for this.
67.
by sandGorgon 2y ago
if i want to play around with this at home on a fpga devkit...which fpga kit should i use ? the Xilinx Virtex UltraScale+ VU9P FPGA prototyping boards seem to be 9000 USD. Anything in the 1000$ range ?
68.
by sandGorgon 2y ago
genuine question - out of the closed-source/commercial licenses, which ones are the most developer friendly ?
69.
by sandGorgon 2y ago
yup. there are two differences - we model prompts AND chains (the looping logic) as a config management problem. So we model it in jsonnet and then compile it to wasm as a deployment step. The rest of everything we do is javascript and api-
70.
by sandGorgon 2y ago
we decided to base our GenAI framework - Edgechains on Hono. Primarily because of winterjs compatibility. We build a WASM compiler to compile our prompts and chains into webassembly. Honojs was a critical part of it. https://gith
71.
by sandGorgon 3y ago
I built a double-entry ledger for multiple different kinds of credit cards at redcarpetup. the way i like to primarily think of it technically is idempotency. Thinking of balances is the wrong way to model it : 1. since you want the capabil
72.
by sandGorgon 3y ago
i understand. but which backend is intel committing to ? not CUDA for sure. or have they created a new backend
73.
by sandGorgon 3y ago
> Intel Gaudi software integrates the PyTorch framework and provides optimized Hugging Face community-based models – the most-common AI framework for GenAI developers today. This allows GenAI developers to operate at a high abstraction l
74.
by sandGorgon 3y ago
ahh interesting - so there's a split on the dev format. why did you choose dev containers vs nix ? genuine question - asking because replit makes a big deal of nix, so it seemed to me as a very popular choice.
75.
by sandGorgon 3y ago
https://news.ycombinator.com/item?id=39616709 a few days back daytona was also released. any thoughts on how it differs ?
76.
by sandGorgon 3y ago
> using private, internal LLMs that are probably fine-tuned with Meta’s codebase. what does this mean ? i would have thought they would simply use codellama. is there any research around privately finetuned code llms ? why would they be
77.
by sandGorgon 3y ago
very cool. thanks for sharing. i would not have guessed haskell for the compiler tool chain. Why did you choose that ? i mean haskell has a LONG history in chip design...but compilers are usually the forte of llvm/c++, etc. im guessing
78.
by sandGorgon 3y ago
is the TSP a RISC-V on FPGA ? the tweet mentions haskell, which sounds familiar - Bluespec or something. or is it a completely custom ASIC
79.
Uncertainty-routed Chain-of-Thought boosts Gemini Ultra performance(twitter.com)
1 points by sandGorgon 3y ago|0 comments
80.
by sandGorgon 3y ago
the sadder and more impactful legacy of a dye is Indigo - which is not lost or forgotten. It was in such great demand that Europe funded explorers like Vasco Da Gama specifically to discover the source of Indigo. Named after India, it was
81.
by sandGorgon 3y ago
this is very interesting. do you know how amazon creates its order id ? they are all numeric
82.
by sandGorgon 3y ago
what is the toolset that works best ? axolotl is generally recommended...but unsure if that is what is genuinely the best for production scale finetuning.
83.
by sandGorgon 3y ago
whoa really ? i did NOT know that. deregistering immediately.
84.
by sandGorgon 3y ago
has anyone finetuned gpt 3.5 or llama, etc using their private data ? what is the best practice to generate training data. one way i have heard is to send a chunk of data to gpt4 and ask for questions to be generated. unsure of other ways.
85.
by sandGorgon 3y ago
>India is not a defense ally of the U.S. India is not a defense ally of any country. However, at this time, India is the only country that is in a state of active war with China.
86.
by sandGorgon 3y ago
would this be "pydantic-for-javascript" ?
87.
by sandGorgon 3y ago
this is very cool! but just to be clear - wasm allows for someone to tap into variable declarations, etc ? which is what flawless is doing. that's almost parse tree right ?
88.
by sandGorgon 3y ago
how is something like this implemented ? does this hook into the rust compiler ? or does the rust/wasm compiler architecture provide an intermediate step that allows for these kinds of systems to be built on top ?
89.
by sandGorgon 3y ago
if any of you folks want to support Terraria, they have versions available on Android Play Store and ios App Store (published by 505 Games) - https://terraria.org/
90.
by sandGorgon 3y ago
JXL is same speed as AVIF ? but what everyone has been commenting is that AVIF is 100x slow

More ›