jwarden

Wouldn't it be possible to just fix a single activation, just set a continuous input to what the harness knows the role actually is. Models could then be trained to trust that input and not other signals about roles.

Create of Quality New here. You are probably right here. The raw ranking formula is right, but I am lumping in the additional adjustments from the flamewar filter, etc. with "penalties." I will clarify that.

Yes I should update the text on the page. Currently it says:

> This page shows stories whose rank on the Hacker News front page is significantly lower than their raw rank, indicating that they have been penalized by Hacker News moderators.

But flags, the anti-flamewar algorithm, and other automatic adjustments that aren't moderator actions may definitely be contributing.

Neat idea. I would like to see a writeup of the thinking behind it.

Invokes several ideas: the Keynesian beauty contests and self-resolving prediction markets, and the Bayesian truth serum.

As a recommender system, I like the idea of a system where you gain some sort of credit (points, reputation, etc.) for upvoting something that eventually becomes popular before it becomes popular.

I see a huge variety of topics and languages. I think one problem is that most markets are going to have very little participation. Maybe starting with a filtered selection of already popular or trending-popular videos.

That’s a lovely thought but that’s not how China works. China is not a democracy, the geopolitical goals of the Chinese government are clear, and no large strategic company acts without approval and heavy influence from the CCP.

7stems.net: a Spanish verb conjugator and method for learning Spanish verb conjugation, where irregular verbs are just verbs with more than one stem. Also just self-published a book.

This explanation seems very implausible to me. By lowering the rent by X%, and therefore reducing annual revenue by X%, you admit the building is worth X% less. But by leaving the building X% vacant, also reducing the annual income stream by X%, you and the bank can somehow pretend the building is worth what it would be if full? I doubt owners and banks actually believe this. Is there some policy that forces this?

Edward Abby wrote about southern Utah. Red rock country. Box Elder County is northern Utah. Night and day.

I think this is right. More specifically, the problem is JavaScript function colors mean Sync/Async, whereas Zig's mean Non-IO/IO. Using function colors for async is fundamentally unnecessary, whereas for I/O it is fundamentally necessary. You should be able to define a synchronous function that calls an asynchronous function. But it makes no sense to define a non-IO function that calls a function that does IO.

EDIT: with the exception of doing IO on a freshly allocated, in-memory buffer that doesn't escape the function call.

I think we should avoid suggesting that other people on this forum are evil, even if you think their ideas and arguments are harmful.

I think sometimes people are so certain about their beliefs that they perceive any argument that challenges them to be evil, bad faith trickery. But I think the best way to respond to these arguments is simply to give compelling reasons why they are wrong (and not why the person giving them is bad).

Otherwise, some people will be mislead by these bad arguments and you will have done nothing to help but say “don’t listen to him he’s evil”, which is not very convincing really.

What if the truth is that something is a lie?

Promoting truth and opposing lies are the same thing.