3 ms·

> Hacker News still seems to have a deeply skeptical culture with regard to machine learning Is... that a bad thing? Skepticism is good. When it's about someth

> Hacker News still seems to have a deeply skeptical culture with regard to machine learning Is... that a bad thing? Skepticism is good. When it's about something as hyped as "deep learning", even more so.

dkarras 5y ago
>Is... that a bad thing? Yes, when it is there for no valid reason, or ridiculous reasons. Skepticism is not a default position you can take like a toddler refusing to eat their vegetables. You need some informed (and non-fallacious) intelligent reasoning behind that. "I'm skeptic about this thing using X because X is so hyped these days" is not such reasoning.
- akiselev 5y ago
  "I'm skeptical about this thing using X to do Y because the burden of proof is on people claiming X does Y and historically they have failed to meet that burden" I don't know what skepticism has to do with ridiculous toddlers - they are almost universally incapable of grasping the nuances of epistemology.
- andrepd 5y ago
  Well, it kind of is. Blockchain has been hyped by charlatans as the cure to all world's ills. That means when you read something about blockchain you should be especially suspicious. Similarly, I've read too many people hyping up glorified chatbots as one step below AGI (see the :o reactions to GPT3), so I'm now extra skeptical about claims about machine learning.
bastawhiz 5y ago
Skepticism is good when it targets bold claims with vague proof. This is not a bold claim (it's a video demo showing the process) and its proof is not vague (you can inspect the source). Skepticism over something like GPT-2 without more than sample output is good. Skepticism over GPT-2 with a workable demo and source is unhelpful.
- andrepd 5y ago
  Funny you mention GPT-2/3, which is by all accounts a glorified chatbot, but which has nevertheless been hyped as one step below AGI by many people.
  - bastawhiz 5y ago
    Has anyone at OpenAI made that claim?
godelski 5y ago
> Skepticism is good There's skepticism and then there's being a non-expert in a field and talking with high confidence. How do you differentiate these? Conspiracy theorists use the same logic. You're right that skepticism is good, but it is easy to go overboard.
- ekianjo 5y ago
  And then there are so called experts who are charlatans as well. Dont ever forget that possibility.
  - godelski 5y ago
    Sure, but skepticism should decrease if there are a community of experts are saying the same thing. As an example, anti-vaxxers often claim skepticism and that they have done their own research. The reason we don't trust them is because we think doctors have a greater expertise in the subject than them (it is, either way, trusting someone). Unless you're a virologist you probably don't actually have the expertise to actually verify vaccine claims. So sure, you are right, but in the context of this discussion you're implying that the vast majority of ML researchers (myself included) are charlatans. I'm not sure what the meaningful difference here is. We're publishing results, people are actively reproducing them, and then some person on the internet that doesn't understand the subject comes along and says "you're full of shit." We can even disprove the claims being made (e.g. I've explained why the network can't be memorizing the game in another comment). That is literally happening in this thread (GAN Theft Auto is in fact a replication/extension effort). Is that meaningfully different from the anti-vaxxers?
jcims 5y ago
>Skepticism is good. It can be, but its certainly not an unmitigated good. Especially when it leads to aspersions of fraud and conspiratorial thinking (e.g. rasz's comment thread below).
roystonvassey 5y ago
I think it’s a problem when it turns to - being skeptical for the sake of it. Not been too long on HN but the top comments on most threads are a contrarian one (and one which I truly appreciate because it provides a different POv) but sadly because it is encouraged through the high upvotes, the crowd tendency is to regress towards this approach, even if sometimes the rigour of the critique is lacking

dkarras 5y ago

>Is... that a bad thing? Yes, when it is there for no valid reason, or ridiculous reasons. Skepticism is not a default position you can take like a toddler refusing to eat their vegetables. You need some informed (and non-fallacious) intelligent reasoning behind that. "I'm skeptic about this thing using X because X is so hyped these days" is not such reasoning.

akiselev 5y ago
"I'm skeptical about this thing using X to do Y because the burden of proof is on people claiming X does Y and historically they have failed to meet that burden" I don't know what skepticism has to do with ridiculous toddlers - they are almost universally incapable of grasping the nuances of epistemology.
andrepd 5y ago
Well, it kind of is. Blockchain has been hyped by charlatans as the cure to all world's ills. That means when you read something about blockchain you should be especially suspicious. Similarly, I've read too many people hyping up glorified chatbots as one step below AGI (see the :o reactions to GPT3), so I'm now extra skeptical about claims about machine learning.

bastawhiz 5y ago

Skepticism is good when it targets bold claims with vague proof. This is not a bold claim (it's a video demo showing the process) and its proof is not vague (you can inspect the source). Skepticism over something like GPT-2 without more than sample output is good. Skepticism over GPT-2 with a workable demo and source is unhelpful.

andrepd 5y ago
Funny you mention GPT-2/3, which is by all accounts a glorified chatbot, but which has nevertheless been hyped as one step below AGI by many people.
- bastawhiz 5y ago
  Has anyone at OpenAI made that claim?

godelski 5y ago

> Skepticism is good There's skepticism and then there's being a non-expert in a field and talking with high confidence. How do you differentiate these? Conspiracy theorists use the same logic. You're right that skepticism is good, but it is easy to go overboard.

ekianjo 5y ago
And then there are so called experts who are charlatans as well. Dont ever forget that possibility.
- godelski 5y ago
  Sure, but skepticism should decrease if there are a community of experts are saying the same thing. As an example, anti-vaxxers often claim skepticism and that they have done their own research. The reason we don't trust them is because we think doctors have a greater expertise in the subject than them (it is, either way, trusting someone). Unless you're a virologist you probably don't actually have the expertise to actually verify vaccine claims. So sure, you are right, but in the context of this discussion you're implying that the vast majority of ML researchers (myself included) are charlatans. I'm not sure what the meaningful difference here is. We're publishing results, people are actively reproducing them, and then some person on the internet that doesn't understand the subject comes along and says "you're full of shit." We can even disprove the claims being made (e.g. I've explained why the network can't be memorizing the game in another comment). That is literally happening in this thread (GAN Theft Auto is in fact a replication/extension effort). Is that meaningfully different from the anti-vaxxers?

jcims 5y ago

>Skepticism is good. It can be, but its certainly not an unmitigated good. Especially when it leads to aspersions of fraud and conspiratorial thinking (e.g. rasz's comment thread below).

roystonvassey 5y ago

I think it’s a problem when it turns to - being skeptical for the sake of it. Not been too long on HN but the top comments on most threads are a contrarian one (and one which I truly appreciate because it provides a different POv) but sadly because it is encouraged through the high upvotes, the crowd tendency is to regress towards this approach, even if sometimes the rigour of the critique is lacking