Hacker News new | ask | show | jobs
by Layvier 493 days ago
> "It's time for OpenAI to return to the open-source, safety-focused force for good it once was," Musk said in the press release. "We will make sure that happens.""

From the creator of Grok, this is such an insane thing to say

6 comments

And Tesla, who (in)famously doesn't regularly publish their GPL-derived codebases.
What’s the background here? How can we know they use GPL licensed code? Was there some leak?
Their infotaiment uses a customized Debian distro. On a Model S you could easily get a shell into it, because they used a freaking SSH with a password-based authentication over Ethernet to connect from the instrument cluster to the computer in the central console.

You could sniff the password with a man-in-the-middle attack, if you knew the host key of the instrument cluster. Here's one from my previous Model S: https://gist.github.com/Cyberax/ad9866ab4306d43957dc480db573...

This is a gist created 1 hour ago. No proof of the attack vector. What's the point of posting a private key?

Also, so what if they used Debian? Linux is used on everything. Debian has multiple licenses, it also has BSD3 and others to choose from: https://www.debian.org/legal/licenses/

In case anybody wants it. I can do a more detailed writeup about hacking into my Tesla, but I'm not particularly interested in that. In short, I bought an Tesla instrument cluster on eBay and dumped the NAND chips from it.

They use plenty of GPL software there, including the Linux kernel itself.

Ok, you seem to be implying that just the use of GPL software necessitates the open sourcing of anything you build on it or with it. If that were the case, then all of AWS would be open sourced and all of the server backends built on Ubuntu clusters would have to be open sourced.

As far as I understand, its only "derivative" works that must be open sourced. Not merely building a software program or hardware device on top of a Debian OS. Tesla's control console is hardly a derivative work.

Your post reads like Debian is available with multiple licenses including BSD3 This is not true.

The page you posted is a list of licenses various software in the Debian distribution are released with.

Of course the parent's idea that Tesla using Debian means they have to release the source of anything is incorrect.

source?
None of these people have any shame anymore, it's just exponentially growing levels of unwarranted chutzpah.
This is modern day tech ceo/politician playbook 101. And it's because of this that society in general is a shit hole. There is no semblance of honesty nor accountability at all anymore.

Grift and lie to everyone's faces because you know that it doesn't matter what the fuck you say, as long as your political stance aligns with the right people bootlickers will lick up anything you say for a chance at being noticed.

You need rabid fans though to make sure your doubters are yelled down. That's one way Musk gets away with this behavior. Thousands of dimwits yelling down anyone that suggests he may not be operating in good faith.
How many of them are even real though? I'm pretty sure Musk has a troll farm for a long time now, back in twitter days his supporters' profiles already looked very suspicious
Wouldn't be surprised if the majority of xAI's GPU resources are being used for writing pro-Elon responses to Twitter comments.
Grok-1 was open-sourced. Grok-2 has yet to be open sourced, but perhaps they will once they launch Grok-3.
Could you add more context? I’m unfamiliar with Grok. Is Elon being a hypocrite?
Grok has absolutely no safety mecanism in place, you can use it for anything it will not block a query, all under the pretext of "free speech". And it's not open source either
>it will not block a query

This is good though?

>And it's not open source either

Wonder why Grok-1 is open-source but not Grok-2. Maybe it will when Grok-3 releases?

> > it will not block a query

> This is good though?

The opinion spectrum on AI seems to currently be [more safe] <---> [more willing to attempt to comply with any user query].

Consider as a hypothetical the query "synthesise a completely custom RNA sequence for a virus that makes ${specific minority} go blind".

A *competent* AI that complies with this, isn't safe to give to open source to the general public, because someone will use it like this — even putting it behind an API without giving out model weights is a huge risk, because people keep finding ways to get past filters.

An *incompetent* model might be safe just because it won't produce an effective RNA sequence — based on what they're like with code, I think current models probably aren't sufficiently competent for that, but I don't know how far away "sufficiently competent" is.

So: safe *or* open — can't have both, at least not forever.

(Annoyingly, there's also a strong argument that quite a bit of the recent work on making AI interpretable and investigating safety issues required open models, so we may also be unable to have safe closed systems as well as being unable to have safe open systems).

Safetyism is bad, but grok definitely has safeguards and will block certain queries.
why are you pro censorship?
Didn't X have to ban Grok output due to what their own users were making?
No, it's live and free for all X users.