Oh c'mon. Here's what Sid's done to help others in the process
* Started 10 companies to enable access to others
* Detailed policy proposals to make this easier process for others
* Open sourced the entire process and all of the associated data (25TB)
and probably other things I missed.
But nah, what would reaaally help is an acknowledgement.
make sure you opt-out anyway before deleting your account. they'll probably train on some archived version if it sees your profile didn't opt-out at some point.
honest question: is there any realistic mechanism that will make them accountable if let's say they just train on 100% of repos without regards to opt-ins? I operate under the premise these tech companies can do whatever they want and there's very little oversight.
No, there isn't. You're relying on PMs or engineers who still care will see it happening and call it out. Other than that, it can become evidence in some future lawsuit. But the damage to you would've already been don.
Fuck Geohot for lending his hand to Musk during the Twitter takeover. He is obviously "sorted" and successful. But his recent blog posts suggests to me that he has started to realize, despite all his success, that if/when the system collapses, he'll be queuing up in the breadlines just like the rest of us.
The sooner the other techbros get the same realization the better.
It was actively good that Elon Musk took over Twitter. Twitter itself is exactly as free a social media platform under Musk as it was under Parag Agrawal (which is to say, it was a privately-owned platform that made arbitrary moderation decisions and engaged in de-facto user lock-in both before and after the acquisition); and the political distaste that a lot of the most active users of Twitter had for Musk actually got them to move off of Twitter and onto to alternate social media platforms, typically Mastodon in the ActivityPub ecosystem or BlueSky in the ATProto ecosystem. Both of these protocols have issues with not being decentralized enough to really mitigate censorship from the system operators, but the status quo now is certainly better than it was before the Musk acquisition.
I didn't know that Geohot had anything to do with the acquisition, but insofar as he did, I'm glad it happened. There's a bunch of different and mutually-incompatible ways "the system" might collapse in a way leading to breadlines, and I have no reason to think your theory that it will be a result of Musk buying Twitter is any better than any other random person's theory about why the world is going to decline in terms of material prosperity in the near-future.
Why? I find Polymarket and these gambling sites just as bad. Is AI with it's fake porn is any better? And just yesterday, Meta and Google was found responsible for addiction[1].
How is this worse? I guess I forgot to say that a site like that would be simply be a poke in the eye of how stupid the internet has become. It's ridiculous to see folks not liking this idea yet finding the idea of betting on wars to be ok.
i'd say it's more like intentionally choosing to use naive string interpolation for SQL queries than a trusted library's parameter substitution. Both work.
There is no "parameter substitution" equivalent possible. Prompt injection isn't like SQL injection, it has no technical solution (that isn't AGI-complete).
Prompt injection is "social engineering" but applied to LLMs. It's not a bug, it's fundamentally just a facet of its (LLM/human) general nature. Mitigations can be placed, at the cost of generality/utility of the system.
MCP itself is just an API. Unless the MCP server had a hidden LLM for some reason, it's still piece of regular, deterministic software.
The security risk here is the LLM, not the MCP, and you cannot secure the LLM in such system any more you can secure user - unless you put that LLM there and own it, at which point it becomes a question of whether it should've been there in the first place (and the answer might very well be "yes").
reply