Welcome to the second AI One Lecture of this academic year.

Let's start this off the right way.

Are there any questions about the stuff from last week?

Ah, no, from Tuesday.

Okay

so we were talking about general admin setup and those kind of things.

We were in one of the resources.

We were talking about homework.

Homework will not give you any bonus points.

I would like to incentivize them better,

but there's just no way without going to jail.

So I'm not going to do it.

But let me tell you

and you can verify that by looking at the old exams

the homework questions aren't that different from the exam questions.

And exams are well incentivized

so kind of by proxy.

You might want to care about the homework problems.

Right.

Tutorials.

We have five dedicated TAs.

They hold tutorials.

That's a safe space for asking questions.

Some of you don't want to ask questions in front of all of the others

and the scary professor.

And maybe the TA who's just essentially one year older than you might be a safer option.

Take advantage of it.

Right.

Collaboration is fine.

In almost everywhere.

And the exam collaboration is not okay.

And I think collaboration is something you need to manage.

Generally

collaboration is one of the motors of scientific progress.

And it's essential.

You'll see that in AI

there's almost no single authored paper

which means it's much more fun to do research and develop stuff.

And by the way, learn together.

It's more effective.

You should get used to that.

Only in some cases where we want to certify your competence

like in the exam

we can't have collaboration.

Another thing in a collaboration is in science and engineering

of course

your ideas

science and engineering is on is a market economy.

Only the main driving force is not money

but reputation.