kettleballroll

What gave the impression that I stay with lots releases? New kernels,.distro releases, different amdgpu ppas.... I've tried a lot of stuff, but these errors don't go away

Fwiw, i have a Thinkpad x395 with AMD Apu. It's mostly nice, but there are GPU driver issues that have been unsolved for almost 3 years now with that generation of ThinkPads (0). Let's hope such issues are sorted out for newer models

(0) https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1870971

It's the other way around, actually

Don't we have these exact types of sanctions already in place i. NK for exactly the reasons you mention?

Innovation can come from anywhere, but the authors of this paper chose to submit to a venue that is outside the focus of ML research. From what I understand, this paper would've been a great fit for COLT (or even a general/broad ML conference like ICML) and since it wasn't submitted (or wasn't accepted), i think it makes sense to assume that it was written by authors who either chose to circumvent the community, or were rejected by it. I don't find your comparison to ivy league fitting, because while there are still a significant number of great candidates from non ivy leagues, the number of great ML papers that have not been published at top tier ML venues is in the single digits -- most researchers I know would rather just upload their paper to arxiv and move on, rather than publishing in 2nd tier venues. Which should tell you how much of a bad signal the wrong venue can be in ML research.

Granted, it could also be that this was authored by people outside of the field who.sont know the rules, but given how much research gets published each day, it's impossible not to rely on heuristics, especially since this paper was.posted without context.

I'm a researcher in the field, and the amount of new research coming out each day is just way too big to handle. Which means I have to rely on heuristics like this to judge what is worth my time. Otherwise I'd be spending 25 hours/day reading papers, and waste brain capacity digesting worthless findings. Not everything that passes peer review is read-worthy, far from it. It just means there aren't any gaping holes in it. Don't get me wrong, the heuristic is far from perfect. But a random paper on HN (which is not known for its abundance of researchers) that was posted without any context explaining why I should care is not a good signal.

PNAS is not exactly the venue of first choice for publishing AI research, so heuristically speaking, this article is likely not worth the reading time. Furthermore neither the abstract nor the first few words of the introduction give me any reason to read on. I'm going to assume this is irrelevant, until someone here can convince me it's not.

What kind of spam detection do you mean, isn't this a messaging service?

It is huge in Central Europe at least. Pretty much everyone is on it, you would be left out of a lot of functions if you didn't have access to it (eg my kids childcare has its parents' chat groups there, my building block has a group there. In online dating, asking for a number to move the discussion to Whatsapp is pretty common... Heck, last time I brought my bike into the repair shop, they asked me if it was okay if they sent me the status report via Whatsapp).

To be fair though, while Whatsapp is the default messenger for pretty much everyone, both signal and telegram also have their users, especially with younger people. Eg the volunteers for helping the Ukrainian refugees in my city organize via telegram, and the above-mentioned bikeshop ended up messaging me via Signal.

As far as I understood, the first order effect is that excluding Russia from swift disables the usual way for Germany to pay for Russian gas and oil, which means that unless they physically move a truckload of euros to Moscow (and get Moscow to agree to that) they are cutting themselves off from one of their major energy supplier. Plus Russia is a big trade partner, but any outstanding bills will then likely go unpaid, and most likely a lot of contracts will be voided.

Most SWEs at my job (FAANG) run Linux on their workstations (not sure if there is even a other option), and a lot of them run it on their company laptops too (the other popular option is using a Mac laptop)

> I will say it turned me into a huge proponent of massive amounts of protein and fiber in my diet (but kinda turned me off Thai food :/ )

Why?

Speaking as an academic for a minute, that this works was already known since "hopfield networks is all you need". That paper even has some theoretical work on how much each layer can store, IIRC. It's weird that they didn't cite that paper though it's a clear extension of that.

> We're not even close to our limits on research capacity.

Maybe not at your stage of research, but it's my understanding that further down the pipeline (eg clinical trials) you're mich closer to the limits, simply due to financial constraints: given that a trial for something you can sell to millions isn't that mich more expensive than one you can sell to dozens,and given that you need to get payed for your job, i don't see how it doesn't make sense to prioritize the lower hanging fruit. Am i missing something?

"Comorbid" implies that it's lethal, doesn't it? Wouldn't "correlated" be better here?

Interesting to hear, i was not aware of that! In my subfield (Machine Learning) publishing in Scientific Reports would be considered 3rd tier at best (frankly, i don't know anyone who's ever done it, because the community would just ignore it)... What field are you in?

still, I wasn't aware they had an impact factor of 4.x, that's way higher than i thought.

It's worth pointing out that this is "published" in Nature Scientific Reports, which has (by design!) A very low barrier for publishing/peer review. It's just a small step up from arxiv. This also means we shouldn't take this report too seriously until it's published in a serious journal

This will be Phase 1 clinical trials, ie they're mostly about side effects (and potentially dosage): ie they'll mostly check whether this will kill you / which dosage van be considered mostly safe. But at later trial stages, this will be one method to do this. Another and mich simpler one: you check whether the vaccinated people develop the antibodies you're looking for.

Very, very, very far away. This paper is about target selection and hit identification, the very first steps of the pipeline. What follows is Backbone optimization (tweaking the chemistry, and figuring out how to actually produce it), pre-clinical trials (in cells and later in animals), and should those prove successful (fairly unlikely in general), then we'll have clinical phases 1-3. Each of those will take several years. Most candidate drugs that even manage to come to clinical trials will fail there (i forgot the exact numbers, but far over 95%). In general: this research is at the stage of someone coming to a software engineer with the words "I have an idea about an app". The actual work hasn't really started yet, and most likely this won't work at all.

Should this molecule be successful (Wich is highly unlikely, as most hit targets aren't) it will probably take an decade until this becomes an experimental treatment, and likely two or more until it becomes a standard treatment.

All in all, this isn't really big news at all, this is a very unexciting thing, and only gets views because it has AlphaFold in the title

(Caveat: it's been a few years since I worked in a remotely related area)

I use the browser for all social media sites, so not installing Apps isn't a solution for people like me

It's very weird, but i can't shake the feeling that this article always uses Facebook examples for positive things (well working algorithms, putting Zuckerberg right up to newton), but very negative towards Google (biased search results, misleading algorithms). It might not be on purpose, or it might really just be in my head, but I found it super weird.

I'm not considering that at all, no. From talks with all my peers, it's pretty obvious to me that this is the very normal aging process, and TRT has too many known side effects (and unknown 2nd order side effects) to be worth it

What's TRT?

I think you're severely underestimating what 20 years mean for your body. I'm a very fit 37 year old, and I very much notice the continued wear and tear. Recovery takes a lot longer, and the body just isn't as limber as it used to be. Partying for days on end, or some extreme sports, and maybe even some intellectual persuits are just WAAAAAY harder then they were 20 or 15 years ago. I can still do them all, but someone who thinks it's a good idea to trade in their 20s for more free time in their 40s is in for a rough surprise.

Yo! I'm mostly doing back-end development as a hobby and to learn some new tech (ai researcher by trade), and I'm looking to join forces with a front end person to work on larger projects. Shoot me a message if you're up to it, we can explore this space together. (Mind you, I'm doing this as a hobby, so time commitment varies a lot.)

That isn't the ViT paper, this one is https://paperswithcode.com/paper/an-image-is-worth-16x16-wor...

The ViT paper doesn't make your list?

It's been some time since I've worked in biology, but: telomere shortening is mostly a protection against aberant growth (ie, cancer): it sets an upper limit to how often a cell can multiply. If we'd live long enough that even healthy cells run up against that limit, we'd have to deactivate/reverse that mechanism. But then we'd lose one very effective defense against cancer.

> So, this is one of those areas I know way too much about and many of the comments seem wrong-headed. Makes me wonder how much faith I should put in comments on topics where I don't know as much.

Experiencing the Gell-Mann amnesia firsthand is always weird. In an expert in deep learning, a research topic that arguably is mich closer to HN's heart. Yet HN comments on the topic are usually a bit off the mark. They feel more like something people would say after taking a 101 undergrad class in the topic, rather than a truly informed discussion. A lot of my PhD work revolved around computational biology. And while I don't work in the area anymore, i know enough about it to see that in sich a topic that's further from HNs core, comments are often not just "a bit" uninformed,but utterly so

The key difference is that RNNs store the already seen parts of the input to use it when processing future inputs, while self attention allows you to focus on all parts in parallel: information from the last input token can be used on the first token, which in RNNs is not possible (NB: that's one of many differences, but likely the one most relevant in this context)