I mean this sincerely, do public apologies ever help in situations like this? Anecdotally, I always seem to see them backfire as insincere or insufficient.
What makes a huge difference for me is having a progression plan and tracking my progress. With both running and lifting, it’s easy to set short term and long term goals and see yourself reach them week after week. That helps with building the habit, and the habit is all you need.
Agree with this and the following FTA:
'Norm was a pure comic. He once wrote that ‘a joke should catch someone by surprise, it should never pander.’ He certainly never pandered. Norm will be missed terribly.”'
I think this is due to the self-reported nature of Glassdoor, and it exposes how this is a flawed methodology. The original point of salary disparity still stands, but any single datapoint here is likely not reliable.
This is my experience, either the recruiter will want to pay you directly, thus the company can avoid its own HR red-tape, or their deal with the recruiter is deliberately hidden from you, but you still join FTE. Of course, there's a million different ways to do this, but an organization like Triplebyte is upfront about taking 30% of your salary for two years.
This is a great tip. Earlier in my career I was stuck at a mid-point where I didn't have quite enough experience to get hired by a household-name tech company, and working with third-party recruiters made finding a job easy. Knowing what I know now, I never would have even made it past screening at a lot of companies. They will take a cut of your salary, but they'll also work hard to sell you, as that's the only way they get paid. You are to them what a house is to a realtor, and the market is hot right now.
I tried the example prompt and I got a copypasta filled with racial slurs. Is this like the tay.ai situation where user input is being added to training data?
In regression and curve fitting, you can add additional, often spurious parameters, and get a closer fit. A better fit does not mean you have a better model, it usually means the opposite, that you've overfitted the model, and that it's completely useless for forecasting and predictive purposes. This is why in machine learning, researchers are careful to separate training and testing data. This is just physicists roasting each other over the same issue.
Where I work, we almost pulled Facebook sign in for the same reason. I'm currently heading the Sign In with Apple integration project, and it was a godsend to get that deadline extended
Have you been able to put any of these techniques into practice? It's been a few years since I tried reading it, but I wasn't able to push through the fluff to make it to the practical advice.