rogers18445

an example:

    nvme fw-log /dev/nvme0
    nvme id-ctrl /dev/nvme0 -H | grep Firmware
    nvme fw-download -f firmware.ebin /dev/nvme0
    nvme fw-commit /dev/nvme0 -s 2 -a 3
    nvme fw-log /dev/nvme0

In an unlikely event, may need to change the slot (-s)

The approach to feed text through a language model and then obtain some kind of vector/hash from it will be leveraged to fingerprint writing and track people.

SCD30 also has autocalibration with outside air though?

That's interesting. For regular use however this may not be a big deal, auto calibration has a secondary mechanism by which it will recalibrate in just a few minutes if ppm < 400.

So after 8 days plugged in, the sensor is sure to be calibrated incorrectly to inside air, then you take it outside and it recalibrates again to correct values within a few minutes?

How would that work? Having one sensor sealed with reference gas does nothing to help the second sensor. Is there actually a dual sensor which shuttles the gas between them? Seems complicated.

This is indeed the absolute best value co2 consumer 24/7 monitor you can buy right now.

You don't have control over calibration however, it's set to auto. So you need to expose it to outside air every 7 days or it will drift. The good news that a drift should be automatically fixed in just a few minutes after you take it outside.

If you live in a high co2 area, the ppm values will not be accurate. But relative measurements will be. And you can just add the difference if you know it to any number you see.

Sensirion CO2 Gadget is good with one of the best [small, reasonably priced] sensors (SCD41). Has BLE interface (with android/ios app). Only shows co2. https://sensirion.com/products/catalog/SCD4x-CO2-Gadget/

It does use auto calibration, which means you need to bring it outside for a few minutes every once in a while (7 days IIRC). Not doing it will not break it, but the readings will drift downward (sensor assumes that the lowest reading it gets in a period of time must be 400 ppm) - and will be fixed in a few minutes outside (few minutes of <400 ppm sets lowest reading to 400).

You can record first, and calibrate/train later. Then analyze recorded data.

There are scenarios where you can record data first, and then train with later access to the space. Then process recorded data.

> This is wrong, regulation can make a difference.

Regulation is per country or bloc. With ML the value of defecting is so high that any regulation you impose on it which restricts its utility will amount to self-sabotage.

It's surprising that so many people on this site side with the Luddites. The back pressure ML is generating is at this point too strong for anything to make any difference. Anyone who attempts to stop it will just be practicing self-sabotage.

An answer to security theater is a performance of your own it seems.

It was inevitable, what people generally want is to be able to interact with another node on the network in a visual way. The fact that browsers started as simple HTML is only due to technology limitations.

If the web was being made form scratch today, it would probably start with a repurposed game engine.

With containers the entire surface area of the kernel is available to attack (syscalls). With a VM the surface is restricted to the VMM and KVM.

This is an oversimplification, there may be other protocols that are passed through or utilized, they would add to the surface.

This is the way things should be done by any competent developer.

Generally a VM jail is preferable - firecracker or cloud-hypervisor(virtiofs & gpu passthrough) recommended.

A proper namespace jail (eg. bwrap) is sufficient for 99.9% of cases. To break out of a properly configured namespace jail you would need to sacrifice a 0day.

A bee colony is to a bee what you are to one of your cells. The author's attempts to apply group selection, kin selection or even natural selection to individual bees is a misunderstanding of what is happening.

There is no selfish calculus possible for individuals within a colony, they are all parts of the whole. Even the queen can get killed by the other bees if there is a younger more fit queen.

Male bees leave the hive every day to fly to a congregation area to mate with potential queens. They are quite literally analogous to sperm cells.

It feels like a tautology. Of course people with power are going to concentrate it and nudge things to ever greater heights in their favor. To attempt to stop it you need other people with power to do something, at which point they will want to do the same.

The only 'solution' is for everyone to be equally poor, because wealth will never be distributed evenly. Those that generate wealth will not stand to lose it, and they have the options to make sure it doesn't happen.

Yes. This is essentially a binary with the browser acting as the OS. And if it's hosted on a server it's a binary which may change at any moment. There isn't even a mechanism by which to freeze its hash and alert you if it changes - a browser would have to provide that functionality.

To be fair, the exact same problem afflicts something like Protonmail, so calling it a toy may be too harsh.

Predictable large inflation each year with no taxes may make theoretical sense in some models. But it seems like it would not be compatible with human psychology.

'with communication' does a lot of work here. tit for tat is a strategy where you start with the positive-for-both-parties strategy and then just copy whatever the other party did last time.

Prisoner's dilemma has a Nash equilibrium where defection is the best move regardless of anything else that happens. If you change the game to be a repetitive 'prisoner's dilemma' with the same actors the best strategy would be tit for tat.

I really like how wise.com made themselves #2 in that chart instead of #1. Very clever.

Technically, the keys are in the processor's state. You are just trusting that it won't divulge the keys without a correct PIN. You are also trusting the processor is properly secured. And you are trusting that no one would go through the effort to extract the keys physically with scanning probe microscopy or something.

They apparently tried to combat NSFW generation by filtering the training dataset not to include any.

Captured traffic is now safe from future quantum computers. They are using classic McEliece ( https://en.wikipedia.org/wiki/McEliece_cryptosystem ) which is many decades old. One of the finalists for the PQC NIST competition - and may still be standardized, key sizes are too large they said.

https://github.com/mullvad/mullvadvpn-app/blob/master/talpid...

https://github.com/mullvad/mullvadvpn-app/blob/master/talpid...

My understanding is that it's not quite that simple. GANs have stability problems (and as a result somewhat out of favor atm) and if the fake detection mechanism isn't a differentiable function itself no training can happen.

> It's an endless cat-and-mouse game.

This is often stated but I think it has to be obviously wrong. This isn't a traditional interactive game such as malware & anti-malware.

You have existing sensors which operate under the constraint of [ real world -> theoretic pixel space -> optics & aberrations & sensor noise -> compression ]. And a single adversary which attempts to fake this chain.

The detection of fakes isn't even an adversary in this game, it's merely a detection of deviation of the faking process.

At some point, probably soon, the faking process will reach a point where any deviation will be drowned out by the noise aspect of optics & sensors & compression.

Usually, you would see RS used in a setting where any error is going to be erasure (non erasure error -> failed checksum -> erasure error).

Then you use something like LDPC codes for in-packet ECC. RS for multi-packet ECC.

If the article writer is concerned about some helicopter footage, he would have an aneurysm over this: https://www.youtube.com/watch?v=mJLr0KMsRAA

AFAIK this was done with commercially mass produced silicon, and it was 10+ years ago. Nowadays they can probably do 10x+ better.

Yes. If you look you will find published literature on the subject. The plots plotting SAT/ACT and IQ will have an obvious trend but be very noisy. Hence why they are poor IQ tests, you can have a decent guess at someone's IQ based on their test results but it's not ideal.