As someone that is familiar with using nvidia-smi to track util, what are some commands people use to track the SM efficiency? The end of the article had some references, but no examples of what to use explicitly.

Unfortunately, SM efficiency is not accessible via nvidia-smi. The best methods to track it would be to:

1. Profile your model with Pytorch Profiler 2. Export metrics with Nvidia DCGM

At what point is google liable for the increased rate of accidents from features like this? Surely they have some data that can estimate users getting in accidents while using gmaps.

You also had to go in store to buy cigarettes, so the application of regulation would work a little differently in the case of social media.

Hey Jeremy, it seems like you could calculate exactly how much a model learns in a single step by calculating the loss for a batch a second time (with no_grad) after the loss is calculated the first time and gradients are updated. This seems like it could produce interesting outputs when graphing the difference of first and second losses at the batch or observation/question level.

In spez's AMA, he mentioned reddit was not profitable. I don't see an advantage for the CEO to say that publicly and it not be true.

There are many ways of determining profitability (see https://en.wikipedia.org/wiki/Hollywood_accounting)

The CEO absolutely has an incentive to make it seem like this is the case, so they can use it to justify more user-hostile changes in the name of profits

I certainly don’t see an advantage in believing anything that any CEO says publicly.

I wanted to do this but never figured out where I could get access to the data. I think with setting up the inputs correctly to handle 0-5 heroes chosen per team it could work. Once you have a model you just need to rank remaining heroes by the expected win probability if they are included with the team. If you have a way to get data I would be interested.

There's a free API endpoint, you can get an API Key based on your Steam logon.

The problem is that it has been highly throttled, throwing 429 errors after just a few dozen calls. When I looked at it before it was "soft" throttled and would return data at a pretty decent rate. If I remember correctly I got something like 80 million game results downloaded in about a week.

You can get 100 matches at a time[1] via this API: https://wiki.teamfortress.com/wiki/WebAPI/GetMatchHistoryByS...

The "ID" is the game ID, which is 570 for Dota 2. Hence the actual API endpoint is:

    GET https://api.steampowered.com/IDOTA2Match_570/GetMatchHistoryBySequenceNum/v1

[1] It would be ever so nice if Steam provided daily batches in gzip files. That would be thousands of times cheaper for them to host, and much more useful for AI researchers.

You should be able to get the data pretty easily from opendota: https://www.opendota.com

Use stratz or opendota. I used statz myself a few times

Does bitwarden work well with autofill? Lastpass was awful with that and very finicky.

They haven’t had the feature for a huge amount of time but it works alright, with some minor mistakes here and there. As with most managers, it struggles when there are more than two fields to populate

For the sites I normally visit, it works well.

GitHub and some famous banks/credit card providers.

Mfa code doesn't autofill, but it's just a click to copy.

Pretty sure you can set it to auto-copy the TOTP, so you technically don’t need to click anything

Do you have experience with it working on android? It seems Lastpass is only semi reliable with how android allows auotfilling fields in.

The formula still works for scales of 5 or 10, you just have to divide by the max rating first and then multiply by it again at the end.

For example a 3/5 stars turns into 0.6 positive and 0.4 negative observation. Following the formula from there will give a lower bound estimation between 0 and 1, so then you just multiple by 5 again to get it between 0 and 5.

Not passed, but bought since the underlying price will have changed.