4 ms·

We constantly hear about how the most mundane criticism on social media is inflicting harm, and words are violence and stochastic terrorism, and ChatGPT needs b

by memish 4y ago

We constantly hear about how the most mundane criticism on social media is inflicting harm, and words are violence and stochastic terrorism, and ChatGPT needs baby guardrails, but when there are literal death threats (https://twitter.com/dataracer117/status/1519167734910619649 https://twitter.com/dataracer117/status/1519167734910619649) and physical proximity to kids (https://twitter.com/elonmusk/status/1603190155107794944 https://twitter.com/elonmusk/status/1603190155107794944) the very same people are suddenly dismissing any possible harm. What's going on?

greenthrow 4y ago
This is so disingenuous. The location of Elon's Private Jet is publicly available info just like any aircraft operating in the US. It also has nothing to do with the alleged stalker attack on his car.
Sebb767 4y ago
> Legal action is being taken against Sweeney & organizations who supported harm to my family. Is he going to sue the FAA? His live might not be easy, but if he really wants Twitter to be a bastion of free speech, reposting very easy to find public information is probably not where the line is crossed. Also, I'm pretty sure that the people who think of the most mundane things being said by ChatGPT are harmful are not the same ones siding with elonjet/sweeny here.
- 8note 4y ago
  What are the harmful things chatgpt is saying? I thought it only said wrong things with great confidence
  - Sebb767 4y ago
    OpenAI is explicitly trying to filter ChatGPT from saying things that may insult people, describe criminal activity (like how to hotwire a car or how to make methamphetamine), incite self-harm and the likes. Instead, you will get a generic "I'm a language model and can't answer that" response. I did not see much people complain about ChatGPT in particular (most likely because the filter works quite well in non-adverserial scenarios), but previous language models have famously started insulting people and have gone as far as to try to convince people to commit suicide. In assuming best interest, though, the grandparent most likely referred to worries about it expressing non-politically correct views.
tlear 4y ago

by memish 4y ago

greenthrow 4y ago

This is so disingenuous. The location of Elon's Private Jet is publicly available info just like any aircraft operating in the US. It also has nothing to do with the alleged stalker attack on his car.

Sebb767 4y ago

> Legal action is being taken against Sweeney & organizations who supported harm to my family. Is he going to sue the FAA? His live might not be easy, but if he really wants Twitter to be a bastion of free speech, reposting very easy to find public information is probably not where the line is crossed. Also, I'm pretty sure that the people who think of the most mundane things being said by ChatGPT are harmful are not the same ones siding with elonjet/sweeny here.

8note 4y ago
What are the harmful things chatgpt is saying? I thought it only said wrong things with great confidence
- Sebb767 4y ago
  OpenAI is explicitly trying to filter ChatGPT from saying things that may insult people, describe criminal activity (like how to hotwire a car or how to make methamphetamine), incite self-harm and the likes. Instead, you will get a generic "I'm a language model and can't answer that" response. I did not see much people complain about ChatGPT in particular (most likely because the filter works quite well in non-adverserial scenarios), but previous language models have famously started insulting people and have gone as far as to try to convince people to commit suicide. In assuming best interest, though, the grandparent most likely referred to worries about it expressing non-politically correct views.

tlear 4y ago