hobofan

48 ms·

721.
by hobofan 3y ago
It's under /apps in the repo: https://github.com/cucapra/filament/tree/f5da227a059f181b996...
722.
by hobofan 3y ago
One thing I tried using it for was for a summarization/reformulation tasks, where it did RAG of ~3-4 smallish (~single sentence) documents per instance where each should be in the end form a coherent sentence. There, batching either ca
723.
by hobofan 3y ago
While there are (commonly used) alternative serialization formats, RDF is married to XML data model, as all core datatypes of RDF are defined in terms of XSD (XML Schema Definition) datatypes and very closely mimics it's data model. If
724.
by hobofan 3y ago
> we are used that everything around us have some kind of hierarchy I think the problem is more that we are used to the illusion/delusion that everything is hierarchical. The problem that we then encouter is with graph drawing is th
725.
by hobofan 3y ago
There is a good repo that collects many of the issues with RDF[0], which I would have loved to have known about earlier in my journey with RDF. [0]: https://github.com/w3c/EasierRDF
726.
by hobofan 3y ago
I think the post mostly answers the questions "why are graph _algorithms_ not better supported in programming languages", with a focus that is much more on "big data" graph processing than graph support in general. I thi
727.
by hobofan 3y ago
I attempted similar mechanics multiple times in the past, but always ditched them, as there was always a non-negligable amount of cross-contamination happening between the individual instances you are batching. That caused so much of a head
728.
by hobofan 3y ago
I think you interpreted that wrong. Less refusals than "previous generations of models" presumably means that is has less refusals than _their_ previous generations of models (= Claude 2), which was notorious for being the worst i
729.
by hobofan 3y ago
Or you could go to the primary source (= the article this discussion is about): > Sonnet is also available today through Amazon Bedrock and in private preview on Google Cloud’s Vertex AI Model Garden—with Opus and Haiku coming soon to bo
730.
by hobofan 3y ago
I think that's not the updated list, but a different list. https://www.anthropic.com/supported-countries lists all the countries for API access, where they presumably offload a lot more liability to the customers to en
731.
by hobofan 3y ago
How so? Because they have taken large investments from Amazon and Google? Or would you also characterize OpenAI as "Microsoft in a trench coat"?
732.
by hobofan 3y ago
I think "bare" Kubernetes is still a quite nice tool that allows for learning transferable skills across clouds (similar to Terraform). E.g. even if I have to spin up my own nginx-ingress to be able to handle ingress resources, af
733.
by hobofan 3y ago
> because the one you picked in AWS doesn't support bare metal That's just because AWS's Kubernetes offering is laughably bad. There is huge difference in your experience whether you use Kubernetes via GKE (Autopilot) or a
734.
by hobofan 3y ago
Yeah, with plain Kubernetes I'd also see the practical limit around ~0.5s. If you are on GKE Autopilot where you also have little control over node startup there is likely also a lot more unpredictability. Something like Knative can al
735.
by hobofan 3y ago
> Having read the article, isn't the point of the article kubernetes in general and what the author prescribes you sign up for/avoid? Yeah, and I think that provides a good basis to discussion, where people can critique/di
736.
by hobofan 3y ago
In my experience, if you use first-party charts (= published by the same people that publish the packaged software) that are likely also provided to enterprise customers you'll have a good time (or at least a good starting point). For
737.
by hobofan 3y ago
OT: Can something be done about HN commenting culture so that the comments stay more on topic? Some technologies (like Kubernetes) tend to attract discussions where half of the commenters completely ignore the original article, so we end up
738.
by hobofan 3y ago
Yes, in a country filled with badly trained police that open deadly fire upon hearing a firework[0] or an acorn dropping on a car hood, more police presence is a bad thing. [0]: https://news.ycombinator.com/item?id=39577403
739.
by hobofan 3y ago
Yeah, it's his second/third try. OpenAI was already his way to "commoditize your complement" so that Tesla's AI division could catch up to DeepMind etc.. Now that this accidentally created something even more powerf
740.
by hobofan 3y ago
Something being behind a paywall and copyright being used to prevent its redistribution is one of the textbook definitions of "proprietary". In almost all cases of standards you can implement those standards without reading the do
741.
by hobofan 3y ago
There is next to nothing "open" (if we go with "open" = "open access") about most of these organizations. As a fun experiment, I went through the first 10 entries of the Wikipedia list. Only one of them[0] prod
742.
by hobofan 3y ago
In the purest of the pure internet, sure (depending on what level of the stack you look at). There you of course have other problems(?) like the fact that most browser-related standards are essentially steered by Google (via its dominance i
743.
by hobofan 3y ago
> Aren't we able to create and use open standards? In theory, sure. In practice you'll have to construct a financially sustainable organization that is able to motivate all interested parties to chip in and is also able to cert
744.
by hobofan 3y ago
> wrap your build environment with docker (or your favorite packager) so that your tooling and dependencies become both explicit and reproducable If you want explicitness and reproducibility please don't reach for Docker. Unless you
745.
by hobofan 3y ago
I would assume that the advantage (for Mistal) here is Microsoft paying them money to be the exclusive model hosting partner, so that everyone has to go to Azure to get top-tier hosted models.
746.
by hobofan 3y ago
Is it actually in a usable state right now? The last 3 times I tried to use it the ~5 projects you had to run/put together to get any higher-level usable application were always in some way incompatible with each other, as they always
747.
by hobofan 3y ago
They have a bit about their technical foundation here[0]. Given that Readyset was co-founded by Jon Gjengset (but has apparently since departed the company), who authored the paper on Noria[1], I would assume that Readyset is the continuati
748.
by hobofan 3y ago
I don't think it's fair to hold the number of open issues and pull requests against them. Looking through them for a minute, it looks like 95%+ are from their own team members, with a good chunk of the issues being "low prior
749.
by hobofan 3y ago
> You have all the problems of a cache with fewer benefits No you don't. You have next to none of the problems of a cache (especially as you directly have to opt-in individual queries to it) like cache invalidation, etc., with all t
750.
by hobofan 3y ago
You are reading from "fresh" (though only eventually consistent) data. For many public facing queries that's enough. After an initial query seen by their proxy, you can configure all future queries of the same kind to be pre-

More ›