I hate credentialism like this. It's anti-intellectual.
There are many smart people on HN who are perfectly capable of reading and understanding research literature in an area that is NOT their primary expertise, and who can make half-decent arguments and deductions.
Reply to the points he raises (if you can). Don't belittle him.
> There are many smart people on HN who are perfectly capable of reading and understanding research literature in an area that is NOT their primary expertise
Most of those people know that reading papers is far from enough to understand the issues at hand and making such stark statements. I've seen this issue a lot in HN: people who have no experience at all in a field, who think they are very smart, reading something on a paper or Wikipedia and then making these broad statements, specially without any sources. Intellectualism implies being humble and recognizing when one is not an expert in the matter, in order to avoid spreading misinformation and doubt.
>Most of those people know that reading papers is far from enough to understand the issues at hand and making such stark statements
Some of us people know that grad school literally consists of learning potentially exclusively from papers. These papers represent the cutting edge of human knowledge and for someone familiar with scientific literature they are not that hard to parse. Doctors and scientists are humans like you and me.
I'm not making policy here, I'm posting on a forum where unusually intelligent people from all disciplines gather and casually speculate on any number of topics. Though I guarantee at this point that I understand the problem at least as well as some 90% of our politicians...with or without sources.
> Some of us people know that grad school literally consists of learning potentially exclusively from papers. These papers represent the cutting edge of human knowledge and for someone familiar with scientific literature they are not that hard to parse. Doctors and scientists are humans like you and me.
Doctors and scientists are humans that, like you and me, spend years and years of study, practice and research on their own specific topic, with knowledgeable people by their side to guide, question them and answer their doubts. Parsing a paper is very different from actually understanding it, specially if you don't have the context they have.
> I'm not making policy here
You're right, but we have already enough disinformation as it is. We all have a certain responsibility towards not sharing unfounded statements in this situation.
> a forum where unusually intelligent people from all disciplines gather
HN is an echo chamber of programmers, mostly from the start-up world. A constant issue in this forum is programmers thinking that they can solve any problem (I still remember some discussions on sheet music that were extremely out of touch with actual music practice).
Also, "unusually intelligent"? Really?
> Though I guarantee at this point that I understand the problem at least as well as some 90% of our politicians...with or without sources.
Credentialism aside, there were no sources provided.
I do find the point about no SARS vaccine compelling, however. Is that true?
I know the flu vaccine gets updated every year or so with new viruses that are prevalent, so if there is no SARS vaccine then I wonder if that is true and if it is true it couldn't be developed.
Again, multiple things thrown out there without sources which are not common knowledge, followed by "you can look this up yourself".
I've only quickly searched, and to be honest it's a bit thick with foreign terminology for me to quickly glean much from it, but seems like this might be the paper the top-level commenter referred to:
Please don't tell people to do a search. If you're going to provide information provide sources, otherwise you're no better than Russian disinformation bots. This is exactly the kind of tactic they use. If you have a claim, provide a source. You are currently spreading unhelpful misinformation.
by your admission the person you are referring to is an engineer in a highly specialized field and is likely to be skilled at probabilistic inference as well as dissecting (certain types of) scientific publications.
I do have some special knowledge in biological research, from seven years of post-doc work at the Max Planck Institute for Computational Biology. We're the group that created the standard of care software for HIV treatment, and for a time we hosted the global influenza database. In other words, viral disease was a specialty of ours.
In my view, as someone with some high level expertise in the field, is that the original poster is correct in their assessment.
Oh my God yes I'm glad to see someone understand. Especially if you have a generalists background. ML is incredibly powerful because deep neural networks, as we're discovering, are extremely powerful, extremely general function approximators. We're inching toward a new type of mathematics.
I've been reaching out to prominent people and I'm getting interest in some kind of collaboration with funding. People are onto the tech and it's myriad applications. There's a reason ML engineer was ranked #1 be demand growth (300%) on indeed last year!
You don't understand. I am making this assertion as a former physicist. We have stumbled upon extremely general function approximators.
Effectively the same way that we can use equations to make inferences about reality - but because of the nature of mathematical notation and the limitations of human ability, though powerful, math has limitations. E.g. you can write out the idealized equations for heat propagation in a conductive medium, but solving them for a real object requires empirical simulation.
Deep neural nets are the next step. Now you can essentially train these neural networks to infer not just the general idealized behaviors, but specific details, discrete values for X and Y on a fine grid that are beyond the practical limits of applied math.
But this is much bigger. It turns out that, much in the way that idealized equations apply to many problems (e.g. exponential growth arising from diff EQ), neural nets generalize to all manner of real world problems, provided the training data is appropriately curated.
These neural networks excel at learning human like heuristics, with machine level precision. You can make inferences for both continuous and discrete probabilistic systems. This is a major development and it's just starting. We've finally assembled the pieces in the last few years.
Could you please stop posting unsubstantive comments to Hacker News? I'm afraid you've been doing it a lot, and we ban accounts that post that way. We're trying, if possible, for something a bit better than internet median here.
I don't understand why LOC was even brought up. You can put all your code on a single line in most languages. The Github code linked even has 26 lines of Haskell which makes this even more nonsense.
Google is now totally useless for any error message from popular software. I was getting an esoteric error from nginx; pasting the long error in quotes to google gave a full page of irrelevant ads with no organic results. DDG and Bing both had a forum post from Igor himself explaining the error as the first hit.
This sort of highly targeted search used to work so well in GOOG and now it doesn’t work at all and hasn’t for a while. I moved to DDG a year ago and never even seem to use !g.
