arjvik

There’s a known fix for SQL injection and no such known fix for prompt injection

Not to take away from this cool project, but its design decisions are incredibly impractical.

As far as I remember, they called it a Guild in all their developer documentation

I'm not sure I agree... while I don't ever see myself writing papers with AI, I hate wrangling a bibtex bibliography.

I wouldn't trust today's GPT-5-with-web-search to do turn a bullet point list of papers into proper citations without checking myself, but maybe I will trust GPT-X-plus-agent to do this.

My favorite link of all time:

https://jpmorgan.c1ic.link/logger_zcGFC2_bank_xss.docm

Definitely not meta

I personally believe that the ideal situation is in fact everyone reclining their seat

When I lived in Texas, we had a massive storm in winter of 2021 leaving many without power for a week.

I was told that Texas maintained its own energy grid independent from the rest of the nation’s eastern and western grids, and supposedly only had a handful of high-voltage DC lines running between Texas’s and the rest of the nation’s. Supposedly this was why we couldn’t rely on excess capacity from anywhere else in the nation while our power generation capability was down.

But this map doesn’t seem to show Texas as isolated - there appear to be many lines in and out and no clear separation?

I don’t understand why you want to enforce only using the public key instead of private key - while I believe you that as of now browsers do not disclose the public key anywhere, I’d also suspect that this is far more likely to be violated and accidentally disclosed by a bug than the private key, which theoretically cannot ever leave the TPM.

Would KDF(deterministic_sign(“well-known message”)) not also provide valid entropy?

Is it just impossible to force a nonce for a deterministic signature?

One of the many patron saints of engineers!

If he so believed in it, may his arrow be pointing up! :)

As a quick hack, is this flaw fixed by using top-p or top-k or min-p or whatever the current hottest logit sampling algorithm is?

How hackable is the firmware, if we want to assign different handling to the button presses?

Can we flash our own firmware to the device?

The more interesting question is is it even possible to learn the logic gates solution through gradient descent?

Is sizecoding the same as the demoscene?

One signed up knowing the risk

(not defending, I also think its insane, just devils advocate)

This assumes risk appetite is the same. For instance, insurance works this way - in expectation you’d pay less yourself, but a 3 sigma event can bankrupt you. You willingly pass on this risk for a price.

These days, you really don’t need multiple ports, reverse proxies that disambiguate via subdomain or path are incredibly easy to configure.

No Free Lunch theorem applies here!

Can I ask what the Google health insurance is like?

Hyperthreading is technically a level above superscalar?

> if you don't properly invest in the maintenance of the thing

For me, it's not the money that's annoying (though of course I'm not pleased by the bills every once in a while). It's the amount of time it takes to keep a car maintained! Seems like just yesterday I took a whole day remote to sit at the mechanics shop for a tire change, but now I have to do the same for an oil change! For this precise reason I end up doing a lot of maintenance late.

I’m confused why this can’t be done in software?

    mount -t tmpfs ram /mnt/ramdisk

Almost a “do you know who I am?” moment :)

Curious - how does one achieve this in Linux?

I assume initramfs-only with special purpose pid0 and only the modules needed statically compiled into the kernel?

What else would it take?

The interesting part is this is done in Pallas!

Seems like the Pallas of old has completely been upgraded

SQL requires setting the max length of a string, and its quite reasonable to set it to len(2147483647)=10 if you were expecting 32-bit int IDs.

Sounds like its time to make him an offer he can't refuse

Very viable, please go this route!

If that's the case, isn't a grep tool a lot more tractable than a Linux agent that will end up mostly calling `grep`?

  * Some standards such as ISO 6709 (Standard representation of geographic point location by coordinates) describe points as (latitude, longitude).
  * Others such as RFC 7946 (GeoJSON) describes points as (longitude, latitude).

Using (hypothetical) std::flip to reify these APIs seems like a loaded footgun - someone is bound to accidentally use it {zero, two} times to convert between orders when it needs to be used once and wreak havoc.

Reminds me of the row-echelon form algorithm we learned in algebra!