oersted

39 ms·

631.
by oersted 3y ago
It has a more descriptive and clickbaity title too, which I appreciate, and probably why it has fewer upvotes.
632.
by oersted 3y ago
People sometimes forget that we had 2-3 years before ChatGPT of fairly high-quality LLMs (from Microsoft, Meta, Google), but they would have to take them down after a few weeks of being released because they became too controversial. Galact
633.
by oersted 3y ago
> i think much of this has to do with stuff like Q, and it makes you wonder how "grassroots" (astroturfing) that movement really was, because of how conveniently it has destroyed the reputation of anyone asking non state approv
634.
by oersted 3y ago
From Warhammer 40K (I am sure it was done retroactively for the joke). Land Raiders and Land Speeders are named after Arkhan Land :) https://youtu.be/LhspXkE8wVQ?feature=shared
635.
by oersted 3y ago
The Spanish voice has an interesting accent: 85% Castillian (from Spain) pronunciation, with a few unexpected Latin American tonalities and phonemes (especially "s") sprinkled in. I guess it's what you'd expect from aver
636.
by oersted 3y ago
> The hardware cost is an argument that always makes me chuckle because there’s always cost in tooling involved. Whether you buy a hammer, a windows laptop or a Mac. At that point why stop there, why not throw in the cost of electricity
637.
by oersted 3y ago
Slightly irrelevant fun fact: the Rust compiler was implemented in OCaml for a long time during the R&D phase of the project.
638.
by oersted 3y ago
The studies discussed in the article state that practice accounts for just 26% (violin playing) and 18% (athletic performance) of what counts for being elite. So no, the whole point is that a swimmer with the wrong build wouldn't get a
639.
by oersted 3y ago
> Practice still mattered: It accounted for 26 percent of the difference between good violinists and the less accomplished students. But the original study claimed that practice accounted for 48 percent of the difference. This is very di
640.
by oersted 3y ago
Could someone recommend a good maths research book that covers the whole messy discovery process, rather than just the end result? I don't mind in which field, and it's fine if it's quite formal (not looking for pop science).
641.
by oersted 3y ago
I was thinking in terms of augmenting our capabilities. Absolutely there is a lot of value in this tech for dealing with medical issues and for monitoring.
642.
by oersted 3y ago
You are right if you define “me” as body and brain, what they can do. What I am more focused on is: what impact “me” as an individual person (an independent human agent) can have on my environment. We build technology mostly to optimize for
643.
by oersted 3y ago
I've always thought that it is a bit close-minded to focus on wearable technology when it comes to human augmentation. Vehicles and computers are, in practice, much more powerful and feasible augmentations than literal sci-fi robot lim
644.
by oersted 3y ago
The weaker model is just a stand-in for human supervisors. They are experimenting on GPT-2 supervising GPT-4 as an analogue to humans supervising the superhuman AGI.
645.
by oersted 3y ago
Functionally it is similar, but trains are very critical civil infrastructure. In the case of John Deere such fraud can also have a serious impact, but it does not affect the public in the same way. If they want to explicitly claim exclusiv
646.
by oersted 3y ago
Yes that makes sense. I am not familiar enough with llama.cpp, but from what I see they have mostly copy-pasted it into WasmEdge for the WASI-NN implementation. Surely a simple compiled binary of llama.cpp is better than Rust compiled to WA
647.
by oersted 3y ago
I'm all for Rust and WASM, but if you look at the code it's just 150 lines of a basic Rust command-line script. All the heavy lifting is done by a single line passing the model to the WASI-NN backend, which in this case is provide
648.
by oersted 3y ago
Alright let me go over all the points. But you clearly care about finding a tool to help you with your work and are quite critical of any half-assed solutions. I would love to work together to literally make a custom tool just for you that
649.
by oersted 3y ago
These are valid remarks, we'll see about including them in the landing page. Before going to the answers, just to note: if you are interested, we are looking to work closely with more researchers, to make sure it saves time for you in
650.
by oersted 3y ago
Did the emails have different domains? We use the domain as a simple filter for the free version, but there might be a bug around to that. It would be helpful if you could help reproduce it.
651.
by oersted 3y ago
Never rely on the AI's own "memory". GPT-4 can remember a lot of deep technical things correctly from its training, but it can also write non-factual statements that look perfectly reasonable. We only use LLMs to retrieve and
652.
by oersted 3y ago
You can sign-up for free and see the options. There's a free version if you register with an institutional email (university or company email domain), we want to support scientists (both academia and private). For full access it's
653.
by oersted 3y ago
Just to plug that we have built MirrorThink.ai: an AI assistant specifically designed to help professional researchers in their day-to-day work. (I am aware that self-promotion is allowed but somewhat frowned-upon in HN. I believe this is r
654.
by oersted 3y ago
I finally got around to watching The West Wing. There's a lighthearted small subplot of exactly this in S01E19. By the way, it is so refreshing nowadays to watch a show about extremely competent imperfect people that legitimately striv
655.
by oersted 3y ago
From Wikipedia: > Publius is thought to derive from the same root as populus and publicus, meaning "the people" or "of the people".
656.
by oersted 3y ago
The provided link does not lead to the post, I think the three dots at the end have been omitted. https://playstationcouch.com/post?this=Lies-of-P-is-a-well-m... ...
657.
Show HN: MirrorThink – Agent for Professional Science and Engineering(mirrorthink.ai)
2 points by oersted 3y ago|0 comments
658.
by oersted 3y ago
I haven't worked on GNNs specifically, but I’ve done academic research on other architectures, and I used MirrorThink.ai to perform a comprehensive literature review. Hope it gives some useful context. --- The paper “Graph Neural Netwo
659.
by oersted 3y ago
That is interesting. Do you mean that the judges like hired mediators? Was this only for civil disputes or in general? Who held the authority to enforce the judge's decision?
660.
by oersted 3y ago
She uploaded her weekly news video yesterday. Seems to remain skeptical but not openly so, simply highlighted that most replication attempts are inconclusive or failed, and commented on some of the new papers (thr DFT theoretical one, 100K

More ›