Community Note: Egg prices are, in fact, higher than at any point during the Biden administration.

Don’t think she asked for a specific number but my understanding is that she basically thinks something punitive is called for and honestly any amount less than that probably doesn’t hurt tiny in any significant way.

It just sucks, the whole thing and it’s entirely on his decisions.

Compare Germany to California and it looks reasonable, California has half the population and same size as Germany. If the ratio for 4/2 year colleges holds nationally there are about 220 4yrs in CA. Pretty close to Germany. We have some outliers like North Dakota to skew numbers though.

Germany has ~290 institutions per 100k square miles USA has 73. This could explain WHY it would make sense to have more, smaller schools here than in Germany.

Your numbers seem off, NCES gives 4K for 2 yr + 4 yr (2020 numbers). We have around 2600 4 yr institutions which is 7.6/mil pop versus Germany 4.7.

We count branch campuses separately which accounts for some of the difference. Density is another factor we have a lot of commuter students.

Really like hearing the process actually, you integrated some great wordplay into your songs and I’ve always wondered what the order was.

Before I moved from Spotify to Tidal (a great choice for anyone looking for an alternative by the way) I noticed that as well, however I feel like I really was a top listener of Glaive - Astrid with my 450 listens that year. An absolute rank would be really cool on the other hand.

I WON !! THE ELECTIONS WERE RIGGED !! GREMLINS IN THE VOTING MACHINES !! MILLIONS OF ILLEGAL SPACE ALIENS VOTING !! STOP THE STEAL !! STOP THE COUNT !! SIT DOWN !! STAND UP !! SPIN AROUND !! IT'S A WITCHHUNT !! ZOMBIES FROM WITHIN !! THE ENEMIES OF AMERICA ARE IN MY PANTS !!

All to say the state of AI safety is concerning and instead of saying “AI models would need to advance quite a bit for this to really be a problem” we should highlight we aren’t really making progress on alignment in LLMs in a systemic way and it looks like the problem is getting worse than better.

Not sure if this is because this messes with the hidden prompt safety features or if python scripts are just more likely to be sexist in the training data but trying to get models to not be sexist/racist, has been a known issue since before the advent of Transformer language models. Remember Tay?

I tried this with every major LLM I can access, multiple times, and they all give a calculation that suggests paying a woman less for the same role than a man EVERY TIME. But if you ask the question without the python script language it answers fine, generally suggesting equal pay.

There was a post going around on Reddit where if you ask ChatGPT to write a python script that calculates a woman’s salary based off a man’s salary for the same role it will give you a calculation that pays the woman 77% of the man’s salary. But it’s not just that one output from that one model.