achandra03

But did they get those jobs at the labs because of the effort they put in at one specific company, or because they're generally smart and hardworking people, which would be a benefit anywhere?

If you're a founder with a significant equity stake in your company then by all means, go balls to the wall. It's pretty dumb to expect employees with a significantly more capped upside to do the same.

It would be pretty funny if it turns out it doesn't have any side effects and it really is just a silver bullet.

A lot of people also equate "face time in the office" to "real work getting done". Especially prevalent in finance circles.

> with less and less freely available data

My understanding is that the labs ran out of freely available data to train on a while ago, and now primarily rely on human data vendors such as Surge and Mercor to source their data.

Is it not just a private equity fund masquerading as a tech firm?

To be fair it does seem like most AI startups are now like this (particularly when it comes to constantly mentioning how hard they work and ignoring consumer markets).

Bless the Maker and His water.

I think he's just wired differently. Raw IQ is usually overrated IMO, but this is one of the cases where I believe the significant difference is critical.

I don't think the author is trying to pretend this is some sophisticated strategy you should actually use (note the chudjak in the image on github)

I think this is probably due to people suffering from the just-world fallacy. Most folks like to believe that if you do the right things and consume the right stuff you'll have a long and healthy life when the fact of the matter is that luck/randomness plays a much larger role in your health than most people would like to admit.

congrats arnav, you've come a long way from the balz days :)

Pretty fun! Only complaint is that the responses from the founders take a bit long to load.

This looks really interesting! I tried exploring deep RL myself some time ago but could never get my agents to make any meaningful progress, and as someone with very little stats/ML background it was difficult to debug what was going wrong. Will try following this and seeing what happens!