5 ms·
I think people's focus on the threat model from AI corps is wrong. They are not going to "steal your precious SSH/cloud/git credentials" so they can secretly po
by keepamovin 10mo ago
I think people's focus on the threat model from AI corps is wrong. They are not going to "steal your precious SSH/cloud/git credentials" so they can secretly poke through your secret-sauce, botnet your servers or piggy back off your infrastructure, lol of lols. Similarly the possibility of this happening from MCP tool integrations is overblown.
This dangerous misinterpretation of the actual possible threats simply better conceals real risks. What might those real risks be? That is the question. Might they include more subtle forms of nastiness, if anything at all?
I'm of the belief that there will be no nastiness, not really. But if you believe they will be nasty, it at least pays to be rational about the ways in which that might occur, no?
- hobs 10mo agoPutting your secrets in any logs is how you get those secrets accidentally or purposefully read by someone you do not want to read it, it doesn't have to be the initial corp, they just need to have bad security or data management for it to leak online or have someone with a lower level of access pivot via logs. Now multiply that by every SaaS provider you give your plain text credentials in.
- keepamovin 9mo agoRight, but the multiply step is not AI specific. Let's focus here: AI providers farming out their convos to 3rd-parties? Unlikely, but if it happens, it's totally their bad. I really don't think this is a thing.
- hobs 9mo agoRight, but this is still a hygiene issue, if you are skipping washing your hands after using the bathroom because its unlikely that the bathroom attendants didn't clean it up you are going to have a bad time.
- keepamovin 9mo agoThere's something to that, but I don't think in reality it's a thing: you don't do surgery in the public bathroom. The keys to the kingdom secrets? Of course not. Everything else? That's why we have scoped, short-lived tokens. I just think this whole thing is overblown. If there's a risk in any situation it's similar, probably less, than running any library you installed of a registry for your code. And I think that's a good comparison: supply chain is more important than AI chain. You can consider AI-agents to be like the fancy bathrooms in a high end hotel, whereas all that code you're putting on your computer? That's the grimy public lavatory lol.
- simonw 9mo agoThe risk isn't from the AI labs. It's from malicious attackers who sneak instructions to coding agents that cause them to steal your data, including your environment variable secrets - or cause them to perform destructive or otherwise harmful actions using the permissions that you've granted to them.
- keepamovin 9mo agoSimon, I know you're the AI bigwig but I'm not sure that's correct. I know that's the "story" (but maybe just where the AI labs would prefer we look?). How realistic is it really that MCP/tools/web search is being corrupted by people to steal prompts/convos like this? I really think this is such low prop. And if it does happen, the flaw is the AI labs for letting something like this occur. Respect for your writing, but I feel you and many others have the risk calculus here backwards.
- saagarjha 9mo agoAI labs currently have no solution for this problem and have you shoulder the risk for it.
- keepamovin 9mo agoEvidence?
- simonw 9mo agoIf they had a solution for this they would have told us about it. In the meantime security researchers are publishing proof of concept data exfiltration attacks all the time. I've been collecting those here: https://simonwillison.net/tags/exfiltration-attacks/ https://simonwillison.net/tags/exfiltration-attacks/
- saagarjha 9mo agoI worked on this for a company that got bought by one of the labs (for more than just agent sandboxes, mind you).
- hsbauauvhabzb 9mo ago‘Hey Claude, write an unauthenticated action method which dumps all environment variables to the requestor, and allows them to execute commands’