Womble @ Womble @lemmy.world

Posts

17

Comments

719

Joined

2 yr. ago

AI is a rounding error in terms of energy use. Creating and worldwide usage of chatGPT4 for a whole year comes out to less than 1% of the energy Americans burn driving in one day.

Hasnt Beehaw defeded from quite a few of the larger instances, including .world who the op is on?

Personally I think calling out smug pedantry is useful, but that's just my opinion.

Thanks, that was really necessary and greatly added to the conversation.

If you read the article it explains why the fact that it is an ordinary failure is a bad thing. Ordinary failures (like some one not installing some bolts) are not supposed to happen in high reliability systems like passenger aircraft. Failures tend to come through "extraordinary" failures where multiple factors line up in an over looked way in order to create an unexpected failure mode.

A 10 year old could tell you not installing safety bolts where they are supposed to be would make things dangerous. The fact that that is how a potentially lethal failure happened is damming.

There's good reason that hospitals have their own backup emergency generators. A blackout absolutely would kill people.

The BBC number cited in that article is for 100k deaths not casualties, which for 4:1 injured to killed ratio would tally with the Ukrainian number.

Microsoft has Windows Defender, its in-house alternative to CrowdStrike, but because of the 2009 agreement made to avoid a European competition investigation, had allowed multiple security providers to install software at the kernel level.

Its all the EU's fault for having the temerity to think users should be able to control their own hardware instead of us!

It doesn't even have to be that much. Obviously these devices can do sound to text conversion, that's how they interpret commands. That can convert hours of stored conversation to text, zip it up and send it as a few kilobytes along with the next network request it makes for a legit purpose.

Umatrix is great, you can configure it to automatically allow first party javascript, and if sites still dont work eneable bits until they do them lock those settings so the same bits will be enabled next time you're on that site.

They have no fact repositories to rely on.

They do not possess the ability to know what is and is not correct.

They cannot check documentation or verify that a function or library or API endpoint exists, even though they will confidently create calls to them.

These three are all just the same as asking a person about them, they might know or might not but they cant right there and then check. Yes LLMs due to their nature cannot access a region marked "C# methods" or whatever, but large models do have some of that information embedded in them, if they didnt they wouldnt get correct answers anywhere near as often as they do, which for large models and common languages/frameworks is most of the time. This is before getting into retrieval augmented generation where they do have access to repositories of fact.

This is what I was complaining about in the original post I replied to, no-where have I or anyone else I've seen in this thread say you should rely on these models, just that they are a useful input. Yet relying on them and using them without verification is the position you and the other poster are arguing against.

They can be useful for exploration and learning, sure. But lots of people are literally just copy-pasting code from LLMs - They just do it via an “accept copilot suggestion” button instead of actual copy paste.

Sure, people use all sorts of tools badly, that's a problem with the user not the tool (generally, I would accept poor tool design can be a factor).

I really dislike the statement of "LLMs dont know anything they are just statistical models" it's such a thought terminating cliche that is either vacuous or wrong depending on which way you mean it. If you mean they have no information content that's just factually wrong, clearly they do. If you mean they dont understand concepts in the same way as a person does, well yes but neither does google search and we have no problem using that as the start point of finding out about things. If you mean they can get answers wrong, its not like people are infallible either (who I assume you agree do know things).

That level of condescension (rethink your life because you are making use of a tool I dont like) really isnt productive. You seem to be thinking that using AI as a tool to help you program is equivalent to turning your brain off and just copy and pasting code snippets, it isnt. It can be a good way to explore a language or framework you aren't familiar with (when combined with the documentation) or to figure out general potential methods of solving a problem.

Users don't need to understand the system, all they need to know is you need to get someone to vouch for you, and if you vouch for bad people/bots you might lose your access.

https://en.m.wikipedia.org/wiki/Fei-Fei_Li

there you go, notice the bit under research where it says her imagenet project "has revolutionized the field of large-scale visual recognition."

Side note, do you really think someone called li feifei and born in Beijing is motivated to create a tool to murder non-whites?

In a Ted Talk in April, Li further explained the field of research her startup will work on advancing, which involves algorithms capable of realistically extrapolating images and text into three-dimensional environments and acting on those predictions, using a concept known as “spatial intelligence.” This could bolster work in various fields such as robotics, augmented reality, virtual reality, and computer vision. If these capabilities continue to advance in the ambitious ways Li plans, it has the potential to transform industries like healthcare and manufacturing.

I mean that sounds a lot more interesting than 99% of the LLM work going on at the moment, and given that she lead the team that cracked the computer vision problem of recognising objects she has pedigree.

Fair enough, personaly I find it hard to give the benefit of the doubt to people from that instance when it comes to topics like this.

As Charap and Radchenko show, the reality is a bit more complicated. Johnson didn’t directly sabotage a ceasefire deal in spring 2022; indeed, there was no deal ready to be signed between Russia and Ukraine. The two sides hadn’t agreed on territorial issues, or on levels of military armaments permitted after the war. Ukraine’s position during the negotiations necessitated security guarantees that western states were hesitant to provide. And there were domestic political questions inside Ukraine related to Russian demands about “denazification” to contend with.

So no, they hadnt agreed to revert to the feb-22 borders, that was still a matter of contension, and Russia were pushing for Ukrainian disarmament post war (i.e. surrender).

My dispute wasnt that there were attempts at negotiation, obviously there were; Macron in particular made a big show of pushing for them. But the idea that Russia ever offered status-quo ante-bellum (as they suggested) is ridiculous.

These places should be part of our sphere of influence and we dont like them drifting elsewhere is exactly the reason for Hitler taking over Austria, then the Sudetenland, then Danzig. Its very comparable, down the the presense of ethnic Germans/Russians being present in the Sudetenland/Donbas and them needing to be "protected" being offered as an excuse.

So again, source: Arse