Llm systems are not really build for adhering to a grammar (other than "a string og tokens").
It is also not clear whether the llm adhering to a grammar is necessary for intelligent agents.
Certainly,a harness can easily correct for it.
2,360 karma · joined April 10, 2024
It is also not clear whether the llm adhering to a grammar is necessary for intelligent agents.
Certainly,a harness can easily correct for it.
Pointing out that regressive Ai politics likely are not going to help researchers crisis of purpose is not unethical.
You also assume in your comment that we need new large generations of researchers, and that knowledge transfer through academic institutions is important.
I get the fear. But entertaining ideas of impact to institutions and potentially new post ai institutions is not unethical.
I think this is an aspect of academic math that a lot of people whish to see crash and burn - the attention and accreditation economy.
> it's probably easy to miss if you have never engaged with research in maths
I don't think anybody are missing anything, in particular not here.
The argument is not far from the senio developer who knows the ins and outs of a code base. Now AI comes along and they complain that they will loose grip of the code base.
At first that is correct. Secondly you accept that the grip might not be that important after all. At least not for a commercial project where you are a cog in a machine.
The question is whether it is different for mathematics.
That's the open question.
A but like whenever the first sprinter hits a new world record other runners follow along.
Knowing that something is possible tends to strengthen our ability to work with it.
We will potentially see the same with math.
We will definitely see a large group of people needing therapy, but suggesting that it is worse than people loosing what little they have is poposterous.
You can take a leisurely drive on your Mc even when self driving taxis can take you from a to b.
It must necessarily reduce to a fear of reduced funding to math fields.
Which is congruent to the taxi analogy.
> problems in many fields of mathematics
Developing these different fields moves complexity from the field itself to the interactions of these fields.
Getting too preoccupied with the established terminology risks us a local minima.
Anf because the field overall has become so complex that we need to decompose into subfields, there will be a good chance that we will not, as individuals, have the capacity to truly see progress.
The map has become so big that we need better tools to work with it.
Out sourcing construction jobs was great for the economy while leaving entire cities in rubbles.
But as soon as it hits the privileged class there is a call to "provide a specific replacement mechanism".
Just keep it in mind when you are stuck in your 12-lane city streets.
It is a weird position though. Public transport is good for car drivers as it reduces load on the roads. Regardless of that, car drivers loathe people in public transit.
That is a very far stretch.
I definitely want to make transport easier. But a 50.000 dollar ticket to entering the previlige of transport is simply not fair.
> "Let's make car owners more miserable"
I proposed to let people pay for the services they use.
That is not an attack on redistribution to unhealthy older people living outside of the city.
In the contrary. By reducing the load on the transport infrastructure you make it easier for them to use it.
And for your wish to redistribute: That can easily be done without the city making the use of the land for roads free. You could literally just give subsidies to people who absolutely need cars.
Then people would likely not have a issue with public transit.
What small scale author receives license payments for anything they have authored from bug tech?
To me it appears that individuals and large companies benefit from copyright asymmetrically, which is the core of the argument.
These types of schemes ring very poorly in my head for charity.
If someone has the ability to give 2$ for every 1$ I give, why are they only giving on contingence?
Just give the money!
And granted, I don't know the exact details about Lean. It might be that they don't have an incredibly simple core - as has elsewise been the norm.
I am not entirely sure about lean, but the core algebras for systems like lean are in the 100s of lines of code.
You can likely convince yourself it is correct in a weekend or less - especially with an Ai to help you understand it.
Especially if cognitive technologies mean that we need to travel less (eg communiting to work, or ineffecient supply chains).
Anyways, there is nobody saying that you can not stay in a big city.
The core of the claim is that people on aggregate will not pay a city premium when it does not come with a job premium.
To your benefit: cities will likely become much cheaper! (and more filthy and crime ridden, but that's another talk).
However, what I think we will see is hyper agglomoration. We will likely see 2 or 3 mega cities where the last interesting jobs will remain along with basically all wealth.
These cities will be incredibly exclusive.
Likely one city in the US and one city in Asia somewhere.
The EU will likely be the big looser.
This is likely one of the absolute worst features of the US dining system.
So many times I have been sitting in deep conversation with friends or colleagues just for a waiter to claim all attention to ask if we want to have refilled water.
That they can even push themselves to such rude behavior is entirely insane to me.
This comes directly after they wrote "statements we can't verify".
Why is it that we can not verify the statements, but we can believe the reasons that people will not touch the code base to be "good"?
Latently they say that it is a quality of itself to do the work.
That is a reasonable statement, however, I am sure they also extend it to harvesting, baking etc. As to not use mechanical help.
Heck, if they actually believed their own statement then they'd not even be commenting here, as that require immense amounts of mechanised help.
You see: normally one reads for recreation or enlightenment.
If that same one would approach their job as a recreational activity, or solely for enlightenment, you might not have that job for very long.
This appears entirely unreasonable.
Normally the reason is not good. The reason is that unit testing is missing or that downstream effects are not entirely mapped out.
Exactly activities that traditional software developers are loathing because they are boring and mentally straining.
Science,at its core, does not care about the credentials or institutions. It cares about the results and to what extend they can be falsified.
This feel a bit like "we know all about physics, we can only get more precise" - moment
I get that there is a cultural benefit to keeping it alive. Just like we ideally want the languages represented at the universities.
But keeping humans in the loop does not appear to be necessary in order to call it science, and certainly not in order to have progress or dessiminate that progress.
I don't have a problem with people doing math. As long that we don't idiomatically hold on to that way of doing things.
I do, however, find it hard to belive that individual humans will play a big role from here and forward, in any scientific desciplines.