mrktf

Imagine hypothetical scenario where model take care life support system for spaceship - then single neuron flip due bit flip and decides to vent air out.

Ah, then my blunder. I saw comments about 64/128 sample buffers automatically tough that most packets will be in 128/256 bytes range where difference not that big.

gigabit have throughput, but not latency. For small packets and on low load link there will be almost no difference (i had better source: https://serverfault.com/questions/276651/network-latency-100... , ethernet cnc controllers have same problems). It needs to go 10gbps or even 25gpbs to see lower latency, where electrical signal uses much higher electrical signaling bandwidth.

I totally agree with number 1: allow transaction processing fees include in recipient as additional fee and make it illegal for payment processor to require match prices (currently merchants cannot differentiate price depending on what card and method used).

Why that drastic, checking system state (telemetry) can be done using 'data diode' style networking - one network just broadcasting sensor data, other network allows changing parameters and 3rd one allows allows software upgrades and so on.

You can be very defensive and design any remote sensing controller to act as two systems - one management cpu only does data routing (no other connection than administrative tasks), sensor cpu works only with sensors. As bonus you can have management cpu act as active firewall.

Main problem it is necessary to have in house expertise (hw, fw and process knowing) which in making company lean are optimized first and outsourcing custom solutions suddenly too expensive.

currently my wish for energy storage if the could manufacture cells with <1% self-discharge rate. Imagine possibility to charge your battery system in summer and live though winter (sadly I need something of 1MW cells)

I would go with something like:

  seed=`date +%s%N`; ( echo "secretknock-20260811|$seed"|sha512sum  ; echo $seed ) | xargs nc -u 192.168.1.1

It is not secure as hmac and it can be 'trivially' brute forced, but don't require extra tools (probably nc not always readily available).

On other hand if threat vector includes network monitor with ability to replay i would use wireguard to wrap ssh traffic.

> the main difference between myself and my peers is that I chose to live with my parents and save for a house instead of moving out

That is 'small' cheat code if parents tolerable. I have multiple peers which got much better life and places to live because lived with parents in their 20s.

It just doing napkin math and even if you live 5 years with parents and doing minimal savings (saving rent and part of food) you already poses good chunk of deposit when comes to buying property.

Silly author didn't take time to measure on X11, instead 'KDE drops it' author could use older version of KDE or SonicDE fork (rebranded KDE and only X11 support). And while on X11 i'm intresting how XLibre would fare in this metric, because git master of this project feels very snappy.

I will add this: if you treat systemd as one trick pony and use for few use-cases which developers envisioned - it run flawlessly, but moment do something not in this path prepare for problems and inferior experience (example of randomly picked tool: timedatectl - no force update date like ntpdate command, you cannot quickly insert ethernet cable update date and disconnect... need to wait for synchronization)

I did napkin math some time ago for my eastern european region: these number works with two caveats you need to get hefty government subsidy and electricity price doesn't rise.

My points are:

- one cold prolonged winter and any savings gone.

- you need to get lucky unit which doesn't need to much maintenance

- if electricity gone pray it will be back in 48 hours (where I live - to get subsidy from government you need demolish furnace)

- using solar panels to compensate when electricity lost or without taking extra electricity: minimum 15Kwh battery storage and at least 20KW panels for generating 8KW heat energy. I based on winter solstice day.

Or in other words: LLM it is optimizing function which is generated by same LLM, think you have random variable y, where generator sin(x+r) and your optimizer trying to fit function sin(x+unkown1) + unknown2 ("unknown" function) - it is obvious that will find best fit.

> but cell tower data isn't going to be as accurate as GPS

My knowledge in this topic is not deep, but cell precision should be pretty accurate, because modern cell tower areas are much smaller, then to have well tuned beamformer it need to have relative precise angles between antennas and know signal travel time (distance). I think it should achieve something 30 or 15 meter precision (doing assumption that distance is accurate in 50/100ns order)

I agree and sadly I wouldn't hold hopes to see actual meaningful changes (granted - last time had windows was win 7),

My reasoning is from bitter experience. I saw too many these honest talks/commitments - it always this pattern when product/company starts to decline. Suddenly somebody with technical background shows up talks about past mistakes and what need to fix. Even sometimes holds discussion, which is usually very reasonable. But as time goes there only cosmetic changes with excuses like lack of resources, market wind changed this time, too hard make changes due politics and etc.

or in other words: people in power are corrupt, but they usually are corrupt equally for everybody.

It can crank proof of work schemes to maximum, something like you need to burn 15-20 minutes 16 core cpu to post a single comment. It will be infuriating for users, but not cheap for bots

Little misconception here (beware i'm using xlibre and causal user). On X11 you can find two mechanisms which can be called compositor :

1st: "enable display compositing" option - this one increases latency as every window draw need go though compositor application (in nutshell it exchanging opengl textures - only synchronization messages goes over "wire")

2nd: the Xserver rendering pipeline compositor, this one goes with modesetting (intel, amdgpu) driver TearFree option - almost everything inside X11 server in OpenGL textures and compositor perform direct blending to screen (including direct scanout).

What I want to tell, on modern X (there are merge requests for Xorg server to modesetting driver, amdgpu have this code) with TearFree enabled you by default optimal hardware acceleration - there comes lower latency

You shouldn't notice lag. On modern Xorg the only round-trip is context switches between server and compositor, because the only thing what is shared is texture dma-bufs (there is inefficiency in mesa code for GLX_EXT_texture_from_pixmap extension, but it is other story). And if dma-bufs is working (Xorg needs to test and pull one MR) you have buffer direct scanouts as in wayland.

With today requirements for accounting, somebody with economics background could tell what would be wrong with following solution?:

If you house owned by commercial entity - taxes are payed from full value, but the valuation to any collateral/derivative goes by something like (0.75x)^l, where l how many levels deep (counting ownership levels). For example it house is in some sort collateral/derivate/indirect ownership mix with 4 levels deep, it can only valuated as 0.31x value (you can only account as it is worth 1/3). In my mind it should reduce attractiveness for speculative buying.

Yes, i couldn't agree more with this. The problem these "bad employees" earns wage by getting results and not entering multiple times mfa codes during day or repeating same logins. And talking from experience: these secure practices starting to approach at least hour of productive time everyday, which is literally robbing time

I think parent means: there no network protocol which can propagate blocking in sane manner between providers (something like bgp for firewalls)

edit: yes, you can you bgp to blockhole subnet traffic - the standard doesn't play well if you want blackhole unrelated subnets from upstream network

Digging deeper there are mechanisms for long time on internal X side (see https://www.x.org/releases/X11R7.6/doc/xorg-docs/specs/Xserv... ) - granted never seen it practically implemented.

And going to rabbit hole there are even proof of concept security implementation named Xnamespace for Xorg fork (needs polishing and much more patches but looks doable. see wip documentation: https://raw.githubusercontent.com/X11Libre/xserver/d2b60a3d6... )

It depends on organization size, just my anecdotal example, I would say the moment IT department becomes own island (for example: can totally ignore requests, with excuses staff overbooked/we need extra planning/6 months extra meetings. Or even worse - process request,but up to point where it can show for upper management and blame you for wasting resources) - you can go full cloud, at least there it is possible get something working in reasonable time.

More on technical side I'm interesting what is plausible explanation for this type "glitches"?: it inconsistent backend router state between processing nodes, processing application restart and screw up in shared memory segment (i can imagine to decrease load times - use "persistent" shared memory block for outstanding data), or just plain hash table collision and lack of empty slots (i mean: https://en.wikipedia.org/wiki/Hash_collision).

My speculation: 1st one - it just DNS fails and you can repeat later. second one - you need working DNS to update your DNS servers with new configuration endpoints where DynamoDB fetches its config (classical case of circular dependencies - i even managed get similar problem with two small dns servers...)

As for occasional follower, my opinion is that: Kent overdid with bending rules until Linus & co got fed up.

It is called recuperator. If it is possible to retrofit i suggest to use mounted in attic (it will be silent). Otherwise, if you want higher quality not and these `breathing` types the box will have similar size like air conditioner (search: mitsubishi vl-100eu5-e)

As long as only TMSC is only top performance chip producer and it is possible to reserve all it manufacturing capacity for one two clients the NVIDIA will hold without problem...

My opinion, the problems for NVIDIA will start when China ramp up internal chip manufacturing performance enough to be in same order of magnitude as TMSC.

Yes, sad part it will be implemented and I betting even in worse form than it is proposed... And worst part of it "safety" it for current governing party to destroy any opposition.

My wild guess it will voted for with overwhelming majority using "times changed" argument.

I believe law environment need to change to make possible digital custom car's ROM. Now everything can be closed in same of safety, security, user convience...