hobofan

41 ms·

811.
by hobofan 3y ago
Amethyst[0] has options for screen padding. [0]: https://ianyh.com/amethyst/
812.
by hobofan 3y ago
The presence of the license strongly suggests that it's a fork. "Fork" and "API compatible wrapper" are also not two mutually exclusive things. Often, one of the easiest ways to build an API compatible wrapper is to
813.
by hobofan 3y ago
What are you on about? There was no stealing and there was no plagiarism. They made a PR that was built on top of another PR. The authorship information was preserved in the git history, and there was no attempt at deception. They also supp
814.
by hobofan 3y ago
No need for a media narrative here. One just has to take a look at what he shouts about on his precious town square and it's plain as day to see.
815.
by hobofan 3y ago
That's completely fine in kernels and low-level libraries, but if I find that in a library as high-level as opendal, I'll definitely mark it down as a code smell.
816.
by hobofan 3y ago
Yeah, but even if you'd take this on as your responsibility (while it should really be the CPU vendor fixing it), you would like to resolve it much lower in the stack, like the Rust compiler/standard library or LLVM, and not indiv
817.
by hobofan 3y ago
I don't think you will find a lot of other CEOs of car companies publicly promoting antisemitism and most recently QAnon-type conspiracies.
818.
by hobofan 3y ago
If that's a bug that only happens with AMD CPUs, I think that's totally fair. If we start adding in exceptions at the top of the software stack for individuals failures of specific CPUs/vendors, that seems like a strong regre
819.
by hobofan 3y ago
> Blender and Krita are both the new kids on the block. They were started from scratch with the ability to have a clean UI. > Blender What? Blender's codebase is about as old as GIMP's is. Up until recently (IIRC the 2.80 re
820.
by hobofan 3y ago
The bottleneck for "easy integration" into games and applications right now is as much the RAM usage as is the slowness. This would probably bring the speed to an acceptable level but you would still have to hold the whole model i
821.
by hobofan 3y ago
I'm not 100% sure, but those seem mostly mutually exclusive (or redundant), with the decision tree in maddness taking on a similar function as the binary tree in FFF that decides which neurons to activate.
822.
by hobofan 3y ago
Most inference builds on top of BLAS libraries, which in their implementation take advantage of SIMD.
823.
by hobofan 3y ago
I think with that magnitude of a speed improvement it should become feasible to do just-in-time embedding creation for semantic search for much larger documents.
824.
by hobofan 3y ago
ollama doesn't come packaged with an easy to invoke UI though.
825.
by hobofan 3y ago
> of their own stated purpose & values You mean the official stated purpose of OpenAI. The stated purpose that is constantly contradicted by many of their actions, and I think nobody took seriously anymore for years. From everything
826.
by hobofan 3y ago
I would assume that the revenue story they are pitching to VCs is licensing the model to AWS, which has pre-existing infrastructure for distribution.
827.
by hobofan 3y ago
I haven't looked to deeply into the code, but it looks to me like it just uses OCI/Docker images as a distribution format. You can find some evidence of that in the model directory (mentioned in the FAQ[0]), where you can find man
828.
by hobofan 3y ago
Yes, I think that is the general trend. Have one model tuned for reasoning that decides a plan, based on which you invoke other models as tools (see e.g. the ReWOO paper[0]). If I had to guess, an approach like this is what powers the recen
829.
by hobofan 3y ago
Mistrals main strength is that it is a good allrounder model with a 7B size, so it runs on a lot of consumer hardware. If you are looking for specific tasks to run locally (with decent RAM and no GPU), the realistic best option in most cate
830.
by hobofan 3y ago
Ollama doesn't use Docker by default. While it is one of the install/download options, the default mode of operation is installing it as a native application (where they provide prebuilt binaries for the main platforms/archit
831.
by hobofan 3y ago
Thanks for clarifying!
832.
by hobofan 3y ago
Is there any data to back up the "faster to start and more resource-efficient", especially if you compare it to the native non-Python solutions that most people are using to run LLMs on local machines? I'm as big of a fan of
833.
by hobofan 3y ago
Just because that's the goal they have written on the tin doesn't mean that that is/was their actual goal. Especially in the early days where the largest donor to OpenAI was Musk who was leading Tesla, a company way behind in
834.
by hobofan 3y ago
For coding tasks (without API access), especially in a conversational setting, Phind has been by far the best one for me. I sometimes still compare it to ChatGPT with GPT-4, but it almost always comes out on top (not missing the point of th
835.
by hobofan 3y ago
<deleted>
836.
by hobofan 3y ago
In the "grand scheme of things", no, it's probably not a big deal. I think in the short term, I think it has the potential to set back the space a few months, as a lot of the ecosystem is still oriented around OpenAI (as they
837.
by hobofan 3y ago
Thanks for posting! I think there are in general quite a few party-like games which could incorporate the current generation of AI in fun ways. E.g. I'd love to see a Garticphone[0] game mode where instead of drawing yourself a DALL-E&
838.
by hobofan 3y ago
I think there is would be another great way to take advantage of AI here, following inspiration from the Jackbox games. In the Jackbox game I played the most, there were intermediate games where there was a chance that you would lose your f
839.
by hobofan 3y ago
Yeah, you can just materialize required items "out of thin air" and it almost always just allows that to happen. I would guess that overall not a lot of effort went into tuning the prompt, which is reasonable as that can still be
840.
by hobofan 3y ago
Because regulators want to tie phone numbers to identities. It helps curb illegal activity, but of course also makes surveillance a lot easier.

More ›