At the end of the day it's just a big weighted graph traversal. Its output is a result of many combined probabilities. It's not deterministic and even if it was the input range is so massive that it would be impossible to comprehensively test.
You cannot possibly know an LLM will do what you command it to. It's impossible by design. LLMs are inherently unpredictable. They can still be useful, but that unpredictability needs to be accounted for to use them safely.
If the tool is inherently unpredictable AI companies should either be held accountable for any mistakes or should not sell/market their services as if they were infallible.
> They are WAY more complex to plan, install and maintain than traditional heating.
I'm curious what about them would be more difficult to plan, install, and maintain. Obviously there are many things to consider when retrofitting a building with a central gas furnace... but otherwise why would they be much more complicated than an air conditioning system?
I've had a lot of mold problems with mine. Because they have to be strong enough to handle the coldest winter days, which makes them way overpowered when running air-conditioning in the summer, which means that when you run them in energy efficient mode, they are actively cooling only a small fraction of the time and all of the condensed water just sits there growing mold all day long. It also leaves the home far more humid than usual because it's not removing nearly as much humidity from the air as a less powerful unit running constantly would.
This isn't a problem with regular air-conditioning that is provisioned correctly for the size of your home, because it winds up actively running a lot of the time so the water is draining as new humidity condenses.
Sounds like you've got a single stage/speed heat pump, the good ones nowadays are variable speed, with pretty significant turndown ratios, so oversizing is less of an issue. I've been idly hacking on this site for comparing heatpump stats, if you're curious to learn more: https://www.heatpushers.com/
I got a ton of value out of mine... but I graduated in 2011, when smartphones were only just taking off and relatively few people had them yet.
But my wife is also a high school teacher and one of the most consistent problems I hear about from her is smartphones being a distraction. If she lets a kid use their smartphone as a calculator, odds are they'll soon be scrolling content feeds, playing games, or chatting with others. If her school required students to have a graphing calculator with limited functionality, it would probably be a benefit to her classroom.
> Why the group is targeting London-based Canonical remains unclear and no reason was given via its Telegram channel. It is presumably because Ubuntu is one of the most popular Linux distros.
Okay... so? I do not understand the connection between Linux and the US/Israel. You'd think Iran would be very pro-Linux since Windows is a very obvious liability for them.
Is there any reason to believe this attack even has anything to do with Iran? They could simply want money and they just happen to also be pro-Iran.
Systemd was NOT Canonical’s “fault”. They pushed upstart until Debian chose systemd, at that point it made no sense to resist assimilation like all other distros.
It's not Linux, it's Ubuntu. Which is developed by Canonical. Which is a military contractor that has a permanent DoD team and works with the USAF. Which is bombing their country.
> I do not understand the connection between Linux and the US/Israel. You'd think Iran would be very pro-Linux since Windows is a very obvious liability for them
Canonical is a British company and the employees are westerners. That makes them targets in the eyes of Iran.
People are forgetting that the Iranian government detests Israel and the entirety of the West. The core principal of the revolution is reversing westoxification (Gharbzadegi) and returning to the norms of the Imam Husayn (Velayat-e Faghih). That's the whole crux of the Islamic Revolution and why the Islamic Revolutionary Guard Corp (IRGC) exists.
Open source and anti-war westerners are viewed opportunistically but with disdain.
Depends on what you mean by "free market". If you think "free market" means market participants should just be free to do whatever they want then yeah, it's weird. But if you think "free market" means that products and prices should be determined by supply/demand and a competitive race to the bottom between market participants, then it's not weird at all.
A lot of people do outsource their thinking to AI, so it's not that weird to bring up. That's effectively how many AI companies are marketing the technology.
But it's definitely possible to use AI without letting it think for you. OP should at least acknowledge that.
Those who dogmatically refuse AI outright may be disadvantaged for some things in the future. But it's also probably hyperbolic to say they will be "left behind".
Humans understand rules to be commands with risks and consequences. They conceously evaluate the benefits of breaking rules against the risks and consequences. They also have their own needs, self-interests, and instincts for preservation and community.
LLMs don't do or have any of this. To them "rules" (just like all prompts) are just weights on a graph traversal used to output text.
At the end of the day it's just a big weighted graph traversal. Its output is a result of many combined probabilities. It's not deterministic and even if it was the input range is so massive that it would be impossible to comprehensively test.
You cannot possibly know an LLM will do what you command it to. It's impossible by design. LLMs are inherently unpredictable. They can still be useful, but that unpredictability needs to be accounted for to use them safely.
reply