... are still idiots and assholes, just with skills upgraded from "none" to "somewhat there, but not too impressive".
The problem, obviously, is not "idiots using LLMs in security space", but instead "motivated, skilled hackers using LLMs".
There is a lot of reserach on this topic laltely,
AI is shrinking the time from finding a bug to exploiting it.
with the obvious one that the universe started off tiny and empty space filled in around it -- the big bang.
No, there was no "empty space" around it. The space itself expanded in the cosmic inflation, per current theory.
"model persistence over long task horizons." That's the whole point of these models, that you can give them a task and they keep working on it for a long time!
No, the models themselves are Request-Response -machines. In order for them to be able to work on a task for a long time, you need to create a loop, where you store the context of the previous interaction and keep feeding it to the model, over and over again. And THIS is where there should be more stricter constraints to actually monitor, wtf is going on. The one who is running the model has ALL the control they need, should they choose to do so.
I'm assuming that this works because when the AP is communicating with endpoints in fixed locations, the AP can tell when signal quality changes as attenuators move through the space. This might work better with modern 'beam forming' APs using modern 802.11 standards, but theoretically one could constantly measure the signal strength between a stationary endpoint and the AP and observe when that signal is weaker than it otherwise should be.
If there are insufficient stationary endpoints it might not work so well unless it can somehow measure reflectivity with enough precision to know that something moved through.
Probably, yes. There was a paper on this, just a week ago: using it to identify people
Another issue is that many Europeans either live in apartment buildings or don't have parking in front of their houses so they can't charge off their home grid. Or if they can, they're limited to something like 6.5 kWh which is a real pain if you have a 100 kW car battery (i.e. 15 hours to charge the car).
Typically, when charging at home, you do not have an empty battery when starting charging. I.e., for short daily commute even 6,5kWh is enough. On longer trips one can charge along the highway with more powerful chargers.
What the article ' forgot ' to mention is the fact we have a Super El Nino weather event sitting off the West Coast of the United States which has a tendency to disrupt normal climate patterns. No, they're not new due to ' climate change '.
I'll just leave this chart here for you to study: El Nino trend
Domestic panels will get more expensive, too, as they are made from polysilicon.
Yes, and this is exactly what this is about. As with paying companies to NOT build offshore wind farms to create cheap electricity, the Orange is also trying to limit solar electricity, so that his friends in the Oil & Gas industry would profit more. https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fhardware.slashdot.org%2F...
Kind of does not make the kernel much more secure in general
Ah yes. gweihir. The resident AI denialist.
Actually, I'm starting to think that this "gweihir" is actually somebody's bot tasked to argue against LLMs. Bot running one of the cheaper models, as his (its?) arguments on oftentimes quite lacking in thought and argumentation...
I think it is becoming more and more obvious that LLMs are just another, relatively small improvement of things. Like happened with all the other AI hypes before. Useful? Yes. Complete game-changer? No. And it is not only that running these models is far too expensive and that is several years in.
It is funny how you have had to move the goalposts all the time in your trolling against AI. At least you are now admitting that LLMs are useful, which was hard for you earlier.
(One also needs to realize the scale and size of the US - the complete European continent is about 10% larger by area than the US - and I'm sure your systems don't work seamlessly from the UK through Russia and Turkieye and the Stans.
Stans are not part of the Europe, but for the other countries, money transfers with IBAN should work just fine across borders. The system covers 89 countries, as of now.
The way most US online bill pay systems work is that they put both mechanisms (electronic and mail-a-check) behind the same user interface. You tell them who you want to pay, they look it up in their database. If it's someone they've got electronic info for then they just ask for the account number and it goes through that way.
How would they identify the recipient based on the name alone? There must be multiple John Does (or whichever name you come up with) in the US?
The way this is done in the rest of the world is that you use the recipients bank account number directly. And with IBAN as the standard, it can be applied worldwide, across any and all financial institutions, so I do not understand what is this other "electronic info" that you are referring to.
Oh wait, the U.S. is not one of the 89 developed countries using this system....
I haven't seen a cheque in 35 years. Our banks don't accept them as a form of payment, just as we can longer take a horse and cart to the mall.
I love how you assume your entire country must work the way YOU do. You assume that because YOU haven't written a cheque that banks must not accept them.
Newsflash. He/she may not be from the same country as you. The situation described in the post IS the reality in most of the developed countries.
Read e.g. about Finland here: https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.expat-finland.com%2F...
"Top Tip: Forget about using cheques / checks in Finland! Checks have not been in use for over 25 years."
Memory crypto stripped - does this spell doom to other Cryptos as well? Trumpcoin, Bitcoin, everything crashing ?
JK. Actually happy that someone is still using "crypto" to refer to cryptography
Now, if you apply a Transformer to the task of language prediction it is, at scale, highly capable, but at the end of the day it is just a mashup-generator recombining language patterns it was trained on into "novel" outputs.
What strikes me odd and surprising is that LLMs actually seem to use written language as an "inner voice" for their thinking. As they do not have that internal ability for thiking processes (as you described) they need to output their "thoughts" and continue reasoning from there. When using Agentic AI for development, this pattern is clear, especially when the AI is trying to debug an error.
God made the integers; all else is the work of Man. -- Kronecker
