Its really hard to say just how clever AI is getting IMO (as a non-expert in the field).
On one hand people say transformer models are just sophisticated autocomplete engines. You look at how they work, and yes this seems to be true.
But then when you give a LLM a completely new problem, not similar to anything they have been trained on - For example, give it a snippet of code and ask it to find the bug.
And they can do this. They can explain what the bug is, and give you a solution. They give all appearances of completely understanding the problem you have given them, and they can pick apart the problem, explain it and solve it. I have done this when stuck on various things with great success.
It really does make me wonder about the nature of our own intelligence, if a program can emulate so much of it but with such curious limitations - such as the difficulty a LLM has telling the difference between a correct answer, and in incorrect answer - Nearly all answers are given with 100% confidence.
>Its really hard to say just how clever AI is getting IMO (as a non-expert in the field).
>But then when you give a LLM a completely new problem, not similar to anything they have been trained on - For example, give it a snippet of code and ask it to find the bug. And they can do this. [...] I have done this when stuck on various things with great success.
I'm afraid you follow the same way of thinking about AI as used by the authors of the article: you accept the anthropomorphization of AI programs. Plus you use an unconfirmed assumption in your anecdotal example ("completely new problem, not similar to anything they have been trained on") to support your unjustified delight in AI capabilities.
Both are - in my opinion - bad for AI developments as they support misunderstanding and false image of LLMs and their application in the real world just like "I, Robot" did to create a false understanding of robotics (and AI...).
I made the mistake of arguing with a somebody on facebook about Gardasil once. In the end I asked her flat out would she rather her daughter got cervical cancer, or had the vaccine. She said she chose cancer.
Some people you just can't reach. I stopped arguing at that point.
Do you think part of their resistance to the vaccine is that HPV is a sexually transmitted disease? The parent probably wants to believe that their daughter can avoid STIs ... because magically thinking.
If it is a pure STI it can be avoided by celibacy. This is not a protocol most humans are comfortable with though. Unfortunately most disease are just mainly spread through sexual contact and have additional vectors so vaccination might still be preferable
And the (simplistic) answer is because many of those 50% vote Republican, because the Republicans say they will fix things and yet always make things worse for the bottom 50%
My understanding is, the danger with Beta emitters is if they are broken up into dust, and you breathe the dust in (or eat/drink it) then you are toast, as your lungs and internal organs get a small, but continuous bombardment of Beta particles which will eventually give you cancer.
If you are contaminated internally, with radioactive dust then there is no way to fix that.
Yeah, the opposition party that is proposing building nuclear plants justifies their plan by pooh-poohing the detailed costings of the current government's plan to move to mainly renewable sources, they claim (without evidence) that the ACTUAL costs will be 10x higher.
And they also have no detailed costings of their plan, but they claim that Australia, who currently has no nuclear industry, can build nuclear power plants faster and cheaper than any other country has before.
The LNP nuclear plan is pure fantasy. The actual plan is just to disrupt and delay the shift to renewables, and extend the use of fossil fuel generators.
>In Australia for example rural areas are all being wired for fibre.
What? Thats simply not true at all. I'm guessing you don't live in Australia, or you'd realize how much rural area there is, and how ridiculous that sounds.
What is actually happening is some areas previously served by FTTN (fiber to the node) or FTTC (fiber to the curb) or other, older technologies such as coax are being upgraded to FTTP (fiber to the premise), but there are around 500,000 Australian homes and businesses served by fixed wireless (4g) and around 400,000 additional homes and businesses only served by satellite.
I am willing to be corrected, but I have not heard of any plans for the NBN to upgrade anybody currently using a wireless/statellite connection to fiber.
The intention has always been to upgrade everyone to fibre but there is obviously a question of budget and priorities. So in the interim they will continue to align the technologies on speed as best as they can.
But you can skip the queue so to speak if you're willing to pay for it:
The issue is orbital velocity is an absolutely bonkers high velocity. Nothing can even come close to that velocity in the atmosphere. The SR-71, fastest air-breathing aircraft ever flow topped out at around 1/9th of orbital velocity.
Yes scramjets could certainly beat that speed, once they are developed but even if that doubles the speed of a SR-71 (which already was pushing the limits of heating) then you are still only doing 1/4th of the speed you need for orbit.
So you have a bit of a catch-22 situation - You can have plenty of oxygen for your engines all around you, or you can have the speed you need for orbit - But not both at the same time. Yes, a scramjet can reduce the amount of oxidizer required (by a lot) but to do this you need the extra weight of wings and everything else you need for proper aerodynamic flight.
A booster stage solves exactly that. I noticed in the launch yesterday that the booster was a hog - it was bigger, but it also had a lot of engines and a lot of work to do, so it burned through its fuel faster. This particular configuration is set up for the separation to happen pretty high, but one can imagine different configurations as well - a bigger ship and a smaller, fuel only booster that works only as long as there's atmosphere.
Or even better - keep the current setup but have a ring of air breathing engines in the booster that work as long as they can.
There's a lot of stuff some extra engineering can do, given enough time and resources. For now, SpaceX is going for the biggest bang for the buck, and they have a very healthy aversion of complicated solutions. But in time, adding a few air breathing engines may become simple enough to be worth it.
Right but the SR-71 had a pretty significant fuselage cross-section that is basically pure drag. If going for max efficiency, the entire nosecone of the rocket should probably be the air intake. Basically take one of the SR-71's engines put it vertical and mount the payload inside of it under the shock cone. TWR needs to be high enough to eliminate the need for wings.
Some of the newer missiles like the BrahMos are designed like this and have a ramjet powered second stage to save on oxidizer. It doesn't reach anywhere near orbital speed, but with this design, the shock cone takes the majority of the heating vs. the SR-71 where there were many exposed parts and materials.
Twitter have apparently geoblocked the requested videos. But according to the law, the e-safety commissioner can require a social media company to remove the content in question. For everyone, globally.
I'm not often in agreement with Musk, but I think I am in this one. Yes, the video (a priest getting stabbed during a live-stream) is horrific and I can understand it getting blocked in Australia as it was deemed to be a terrorist attack.
Musk shouldn't be playing childish games and throwing insults out there about communist censoring with this one, he should be making the very reasonable argument that any one country should not be able to effectively restrict content viewable in a different country. If twitter complies with this and removes the video, then this sets the precedent that China can require twitter (or youtube, facebook, or anyone) deletes a video criticizing Xi Jinping that they don't like.
Unfortunately the media here in Australia aren't showing this nuanced viewpoint either, they are just reporting Musk's name-calling remarks, and the government's thunderous objections.
Its really hard to say just how clever AI is getting IMO (as a non-expert in the field).
On one hand people say transformer models are just sophisticated autocomplete engines. You look at how they work, and yes this seems to be true.
But then when you give a LLM a completely new problem, not similar to anything they have been trained on - For example, give it a snippet of code and ask it to find the bug.
And they can do this. They can explain what the bug is, and give you a solution. They give all appearances of completely understanding the problem you have given them, and they can pick apart the problem, explain it and solve it. I have done this when stuck on various things with great success.
It really does make me wonder about the nature of our own intelligence, if a program can emulate so much of it but with such curious limitations - such as the difficulty a LLM has telling the difference between a correct answer, and in incorrect answer - Nearly all answers are given with 100% confidence.