Do you say that as a supermarket manager, or a random person on an Internet forum? Because if it's the first, I'm very interested in hearing more on how you actually manage all the devices in your store and which ones require remote automation and which ones don't.
They are all being sold on the idea of granting external access to a cloud application so that a third-party contractor can know when the system goes into disarray and pay for one less maintenance guy's salary.
Could the same thing be done with SMTP way more securely than opening up your firewall to some random AWS EC2 instance? Yeah no doubt.
Do you think grocery retail execs actually understand computers past Microsoft Word and Excel? I report to one, they can barely listen to me say the word WAN failover.
Like all things in IT, experience is the best teacher. Organizations will fail to invest in their technology, security included, and only then will the execs start to trust that the wizards in the backroom are telling them truths instead of asking for new toys - if they're lucky enough to still be the same exec that heard they should take security seriously before the Bad Thing happened.
In that case it's just a tradeoff between convenience and security, it isn't because it would be too technically difficult to isolate the systems. Whoever installed them could isolate them if there was no reason for them to be connected to the internet. I doubt most grocery store managers understand the precise engineering behind their freezers either, but they still buy them.
Very likely they were sold on these internet-connected systems without being told about the drawbacks, which is sad.
The people selling the systems often don't fully know the downsides themselves either and in a sense they're incentivized to not know as it makes it easier to sell the lie that there are no drawbacks to having your HVAC system exposed to the wider web. But you are also overestimating the average skill of an IT person in these environments and the capacity for their IT or MSP team to step back and evaluate the security of all the systems they have to manage. There's only the ability to evaluate convenience vs security if you actually have the time to formulate the question - if you're drowning in ticket submissions you will likely fail to properly catch every such issue. Certainly for these larger EU corps they should have the capacity to make those determinations, but who knows exactly how the management of these systems were delegated. My bet would be it was the responsibility of an under-resourced team with too much to do already.
I'm catching issues as I can as the head of IT at an org replacing our Danfoss systems with something else, but it wasn't until about 1.5 years since the initial decision to replace them that I'm finding out from a sub-contractor of the contractor that there's a cloud app our HVAC network will need to talk to - which is not-so-funnily enough a requirement by the state for us to receive associated grant funding for a portion of the new system cost. This was never mentioned to me at all as the head of IT. It's not even a convenience vs security issue, it's a misaligned corporate communications and approval channels issue. In other words, the capacity to secure your technical systems is downstream of political organization.
> They are all being sold on the idea of granting external access to a cloud application so that a third-party contractor can know when the system goes into disarray and pay for one less maintenance guy's salary.
Even in that case there is zero reason for the devices to be reachable from outside the network rather than the devices reaching out to known hosts only.
I think the poster got it completely wrong. I suspect the suggested command was designed to help new users, who don't always understand that they should do something. An interface like chat, where you are given an empty box, has a daunting learning curve (remember DOS prompt?). So I see this feature as onboarding, to get users from novice to intermediate.
surely going to max out at slightly less good (we probably can’t get perfect data) than the best human
With that argument, the basic calculator would have been impossible.
It's pretty clear that machines can be better than the best humans at memorizing things, computing relationships between concepts quickly, etc. So while you can argue that LLM are not the path forward, I have no problem imagining that using all of humanity's knowledge with a machine could beat the best human at a field.
In particular, I found this very misleading or irrelevant:
a typical CPU project anecdotally has about three times as many specification and validation engineers as design engineers and a 5:1 ratio is not unheard of
The reason silicon design has such verification to design ratio is because the cost of one bug is many, many orders of magnitude higher than software. Both in dollar cost and in schedule cost (it takes months to fab a chip, and if you messed up and need to spin a fix, it costs tens of millions of dollars, not counting any design engineering cost).
I don't think you can extrapolate these very industry-specific facts to judging LLMs.
> The reason ... is because the cost of one bug is many, many orders of magnitude higher than software. Both in dollar cost and in schedule cost (it takes months ... and if you messed up and need to spin a fix, it costs tens of millions of dollars, not counting any design engineering cost).
Aren't you just describing waterfall? That's still very prevalent in software engineering, and pretty much any other type of engineering – civil, chemical, building, architecture, drug discovery.
It's typically true that software can fail faster and cheaper, but it's also true that the costs are still vastly higher to fix later in the process.
No. Silicon is on another level. Which is why the EDA verification is an industry on its own.
Sure, there are some software that have similar "can't have bugs" requirements. I imagine the computers on Moon missions also had that kind of high bar. I wouldn't use NASA requirements as a proof for how LLMs should be used.
I wonder what kind of new AI law would be useful right now. Maybe this one:
if an AI agent does something, you (the prompter) are responsible by default, unless you can show that your the agent itself behaved in an unexpected way and that you in no way prompted or hinted at the bad behavior, in which case the model provider is liable
The idea is that by making it clear who is responsible, corporations and others start paying more attention because they become financially liable.
On the other hand, I wonder if we'll end up with another variation of the cookie law, where every AI user or vendor just adds "don't do anything illegal" as part of their prompt to defend against that law. Thoughts?
Why is it a threat and not an opportunity? Imagine if in ancient times there was their Oracle that could produce mathematical proofs of any question you asked it, would you burn it or try to understand how it works and use it to ask questions you are stuck on?
I would burn it because it's too much power for any one person to handle. Or, I would take it for myself by force and use it to dominate everyone else.
Yes as I said in the comment, I know they do design them - based on arm.
But come on - if they had to use intel’s fab - their chips would be terrible. The reason the m6 is faster is largely because TSMC invented a smaller process node.
