I don’t think any modern planes have calibrated/approved mechanisms for weighing themselves. I think some older freight models and conversions had load cells but they were advisory tools used to better balance the loaded cargo, not primary sources for weight and balance. I think it would be quite a regulatory and reliability feat to get automated weight approved in such a way that didn’t require doing the same manual calculations as today as a fallback. And those manual calculations are hard to decentralize because they rely on many disparate sources (catering, freight, baggage / ramp, gate crew, pilots, fuelers) that have to aggregate somewhere.
I’m not aware of any approved automated weight and balance system, and I’m not aware of a modern deployment of landing gear load cells.
So, that. It’s centralized because to do a decent job at guessing the weight of the plane you need all planning data (catering reloads, freight load out, bag counts) in addition to fuel load and passenger locations plus overhead bag counts. So a centralized system is a fairly logical place to fuse the data.
It was an option on the Airbus A300 back in the 80's, AFAIK nobody bought it because it was a maintenance issue.
It again became an option on the Boeing 777F [1], I also believe nobody (or very few) bought it.
There's benefit in the freight side of things because the density of stuff they deal with (plastic vs metal bricks vs some unwieldy turbine shaft) varies widely, so just eyeing it can't help detect mistakes.
>So, that. It’s centralized because to do a decent job at guessing the weight of the plane you need all planning data (catering reloads, freight load out, bag counts) in addition to fuel load and passenger locations plus overhead bag counts.
Strange how the managed to do it for decades before those systems then.
Likewise in Denver. They said the operators for the manual fallback are in Atlanta though, which might be some of the news confusion. It was also unclear if it was the actual weight and balance calculation or some upstream data source automation that was broken.
The process model you’re describing is also what’s old is new again; a lot of the web frontend on a database RAD tool systems like 4D did this in the late 90s, as did the IBM HTTP server on AS/400 (basically the whole family of “stuff that originally spawned a process for a terminal user session and got ported to web.”). I think some of the more session oriented frameworks like Wt did this too, although a lot of others preferred to use multiple session handler continuations in a single OS process since things were so expensive back then.
> were the UK users of ADP informed when ADP was pulled? Or was their data just suddenly up for grabs?
This wouldn't really be possible with the design of the system, and existing users retained their ADP; only the option to enable it was removed.
Apple did a _good_ job here IMO; they did not secretly backdoor anyone's data or remove protections they had promised, they just (after fighting, even!) removed the ability to enable the feature that was banned.
Apple has privileged access to most Internet connected iOS devices, via software updates. Once a user unlocks their device, a patched OS can just upload decrypted data or keys from RAM. If secrecy is not required, they can do this broadly. When secrecy is required, they can do this undetected for a single phone. They could prevent themselves from doing this with a cryptographically chained public log of binaries.
> They could prevent themselves from doing this with a cryptographically chained public log of binaries.
I do find it a bit annoying that they don’t publish one of these for system updates, since they do for cloud enclaves - it seems like a cheap feel good policy.
But, if you’re going to this level of threat modeling, I’m not sure it really matters - how would you perform both an audit of the published software and an attestation of the system in a way which assured you that your device was running what was published in the chain? You can audit iOS today by unpacking updates, and many researchers do (which also makes an existing broad backdoor of this kind somewhat unlikely). If you’re worried about personalized attacks, I believe you can also skip OTAs and only install using IPSWs, where afaik the disk images aren’t device personalized (although the tickets are). But at the end of the day any attestation feature on the device is going to run on the device, so you have to trust something.
I agree with this, and I think most of the posters here aren’t following.
LLMs are very good at RTL (all I can consider is that the parallel posters arguing otherwise haven’t tried a modern one against significant RTL?). It’s easy to verify, the optimization constraints are relatively straightforward, and there’s adequate simulation to build a good gym for it.
LLMs are not good at layout and floor planning and because there is near zero public data available, I’m not sure how well they’ll iterate. For PCB design they’re improving because while it’s a similarly thorny problem, the physical constraints (trace properties on a given substrate) are highly public and can be simulated and turned into an RL gym. For ASIC, any public PDK is 10-20 years old so making an RL gym becomes outrageously difficult on top of the fundamental problem (highly multivariate/layered spatial optimization) being hard to solve.
Well, you're not alone in that thought. We've had a steady stream of papers claiming RL models that can macro place and floorplan since at least 2023. [1] And I won't even mention that Google paper...
There have been some minor victories, but it's still mostly done the traditional way in every group I'm aware of.
The “normal” Ghidra MCP is exceptionally bad but straight scripting it headless or having an LLM build a better MCP delivers similar or better results to IDA. Ghidra emulation is much more robust which is helpful. The official IDA MCP is very well made also though and it’s definitely an easier win if you want to spend the money.
I wonder if Radare would work well with LLMs. I never had the patience to learn how to use it before good LLMs appeared. Ghidra's decompiler, Hopper's mach-O support, and IDA Pro's Windows symbols support were always good enough for me.
Keys in European and newer American cars reliably have Some Kind of real-ish crypto scheme applied, as this is fairly regulated in Europe. It’s an anti theft measure driven by the real stolen auto trafficking issue faced especially in Europe. Most of the aftermarket tools vendors use some kind of exploit (code exec, arbitrary read, update signature bypass, etc.) to extract the crypto material from a control module participating in the immobilizer - it’s neither a weird conspiracy nor as simple as it seems to bypass.
Old news - for 20+ years in Europe, 5-10 years in the US, and now in Korean and Japanese cars too, ECM updates have been cryptographically signed for warranty, emissions, and now EU cybersecurity regulatory compliance. It turns out there are still bugs in things with cryptography too though.
There are explicit data retention contracts for Workspace (including Gemini transcription) and Google claim this data is never used for training or human review, so this is fairly baseless.
This gist is specifically about the consumer version of Gemini used in Gemini Apps. Basically if you connect your consumer Gemini (which always sends stuff to the training pipeline) to your personal Google Apps account, it gets tools that it can invoke to read your Google Apps data, which then might appear back in your consumer Gemini transcripts that are getting sent to the training pipeline.
Doing this is a stupid product move from Google (per the usual) because it confuses their enterprise customers into thinking Really Bad Stuff like this is happening.
reply