I think I might have missed something, having tried to recreate this in my own Notion, this searches the URL but doesn't actually send data to that URL.. right? Where's the exfil? (Apart from to the search service)
I just tested Notion's AI bot by asking it to make me a new page with the contents of a URL, then confirmed from my server logs that Notion accessed that URL.
It used user-agent Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 and connected from an IPv6 address of 2600:1f14:1c1:bf05:50ec::13
I think the idea was to trigger a request to the specified URL by passing it as the query string. But the search tool doesn't appear to work that way. Or maybe it does and they just forgot to show server logs with the exfiltrated data to demonstrate that the attack succeeded.
That depends. If you put the catfood in a robot friendly container that is a lot easier than getting the robot to scope if out of the bag. Can you train the delivery company (driver or robots) to place packages in a specific location on the porch - hunting a package down is hard, unless the package is a standard package that is somehow designed to be easy to locate (there are a lot of options here, but you and the packager must agree on using it).
In large part what makes this crazy ambitious is not that it can't be done it is that you need to program all the details and minor variations in environment mean the robot can't do anything. It is easy to program a robot to move 5cm, it is hard to program it to identify random items that are placed in random locations. Things are getting better, but this is a hard problem.
Automated cat feeders already exist, we have a water fountain, a kibble feeder than needs monthly refill and a wet food feeder that is stocked once per day.
I've heard this before but I don't think I believe it. I spend many hours every week 1. picking up clutter and putting it away 2. sorting clean clothes into drawers (I have a family of 5) and 3. shuttling dishes to/from the dishwasher.
I would pay quite a bit of money to stop doing those things. Especially #1. Is there a simple non-humanoid automation I'm missing?
In India you can pay people to do this for less than $1/day. My coworkers there generally are paying 2 people that price even though there isn't enough work to keep even one busy - that way if one servant quits they don't have to do the work themselves.
For me I'd have to pay something like $30/hour after you account for taxes, their take-home would be in the $15-20/hour range. I'd also have to learn a foreign language because almost nobody who speaks English (or even Spanish) will accept a part time job making that little part time. I'd need an accountant to figure out the exact price of course so more cost but lets just work with $30/hour. It would cost me nearly $10,000 per year to have such a servant in the US, and the only reason it is less is I can expect servants to quit often enough that I never have one for the full year as they find other jobs that are better (30*365 = 10950 if they work every day)
That is why so many Americans (and Europeans) want robots - there are a lot of tedious things that we are doing ourselves that we would love to have someone else do but labor is so expensive we can't afford it. Even if you live below the poverty line in these countries you realistically have a rich lifestyle in many ways - HVAC, lights, smart phones... modern life as provided many luxuries that kings of the past couldn't get (don't get me wrong, life for the poor is hard despite those luxuries)
That depends on what this can do. 6k would be a price I pay if it can do enough. I already have a $500 robot vacuum - which cannot do stairs and I have to carry it if I want it to do a different floor, so already $1000 is reasonable if all it is is a robot vacuum that can do stairs. If it can do things like fold and put away laundry, or load/unload the dishwasher (or wash dishes?). Can it put away the toys my kids leave all over (and also save me the bother of putting my toys away when I'm done)?
Related, non-causal event: BGP origin hijack of 1.1.1.0/24 exposed by withdrawal of routes from Cloudflare. This was not a cause of the service failure, but an unrelated issue that was suddenly visible as that prefix was withdrawn by Cloudflare.
I'm a bit uneducated here - why was the other 1.1.1.0/24 announcement previously suppressed? Did it just express a high enough cost that no one took it on compared to the CF announcement?
CF had their route covered by RPKI, which at a high level uses certs to formalize delegation of IP address space.
What caused this specific behavior is the dilemma of backwards comparability when it comes to BGP security. We area long ways off from all routes being covered by rpki, (just 56% of v4 routes according to https://rpki-monitor.antd.nist.gov/ROV ) so invalid routes tend to be treated as less preferred, not rejected by BGP speakers that support RPKI.
Looks really cool! But I have to able to bring my own LLM. In a space moving this fast by the time my enterprise has finished vetting a tool that wants to be it's own LLM/process my data itself, it's already out of date.
I'd love to be able to connect Azure AI and Vertex to this (for the full range of models it uses)
reply